Markus Koskimies (
Sat, 16 Nov 1996 21:35:13 +0200 (EET)

I was asked to send this message to this list. So here it is:

>> 5. There was probably bugs in TCP/IP-handling in kernel version 2.0.0.
>> My server program, which uses heavily 'cloned' threads and sockets,
>> print out kernel oops, if it's stopped while there are clients
>> connected (it should be possible; if server machine faults, clients
>> starts to try reconnection). After that I'm starting to get 'invalid
>> wait queue' -messages and after some time for example init segfaults.
>> After that I'm no longer able to use system. Is this behaviour known
>> and is it corrected in these later versions (2.0.24/25)? This behaviour
>> was fully reproduceable.
>Ouch. Please report it to the kernel developer mailing list.
>The problem was predicted, but never observed. Perhaps you could
>send an exploit program to the kernel developers.
>The kernel mailing list is You can
>subscribe by sending mail to with a
>body containing "subscribe linux-kernel".
>Linux 2.1.10 should not suffer from the problem.

Please, contact by email, if you want more information. I will test this
program out using 2.0.24, when I'm ready to boot. I'm afraid that situation
is just as bad as previously (high traffic) and that's why I'm going to