Re: [BUG][2.4] /proc/kcore is a random generator ?

[Richard B. Johnson]

On Thu, 26 Feb 2004, Willy Tarreau wrote:

> Hi all,
>
> I'm encountering a rather strange behaviour here with 2.4.25 on a P4 HT,
> even when I boot it with "nosmp" :
>   - du /proc/kcore says '525255' (I have 512 MB RAM)
>   - if I do it again many several times, sometimes it says '0'
>   - after a moment (or several tests, I don't know), it stabilizes to some
>     value (either 0 or 525255).
>   - then, just starting vmstat, or logging into the system is enough to
>     make it switch to the other value.
>
> At first, I had the feeling that it displayed '0' when I have an even number
> of processes, and 525255 when I have an odd number, but it's not even the case.
> Once it doesn't move by itself, a few fork/exec are enough to switch the value.
>
> If I boot in HT mode, sometimes I can reliably make it display 0 and 525255
> alternatively, just as if the pid parity or CPU number was involved in the
> result.
>
> Update: same results with 2.4.26-pre1. du from coreutils 4.5.4 and 5.0.
>
> This is amazing. Did anyone notice this ? A friend just told me that plain
> 2.4.22 on his notebook does the same !
>
> Regards,
> Willy
>

Well with 'older' utilities on 2.4.24, I always get '0' with `du`!

Script started on Thu Feb 26 11:48:23 2004
# du /proc/kcore
0	/proc/kcore
# ls -la /proc/kcore
-r--------   1 root     root     335536128 Feb 26 11:48 /proc/kcore
# exit
exit
Script done on Thu Feb 26 11:48:43 2004

Probably moon-phase.


Cheers,
Dick Johnson
Penguin : Linux version 2.4.24 on an i686 machine (797.90 BogoMips).
            Note 96.31% of all statistics are fiction.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/