Re: 2.2.0 Bug summary

kuznet@ms2.inr.ac.ru
Wed, 30 Dec 1998 14:30:17 +0300


In article <199812290146.BAA12687@terrorserver.swansea.linux.org.uk> you wrote:

: o Run two processes that keep rejoining multicast groups on an
: SMP box - crash

Actually, this bug sits in hundred of places over networking.
I was obliged to repair it earlier, when it was less painful :-(
The patch is set of one-liners.

: o bootp autobooting stuff corrupts other hosts arp stuff it seems

Martin Mares made correct ipconfig.c.

: o Tulip driver/fast routing stuff needs to be resolved. If they cant
: be merged the default tulip should be a current one.

They can be merged and they have been merged, but I am still nervous
about driver correctness.

: o fchmod on AF_UNIX sockets doesnt work like BSD

It works like linux-2.0, right?

Seems, fixing it requires to attach i_sb to socket inodes
to hook notify_change. It may have some bad side effects for VFS.

Is it really necessary? I never saw someone used this feature.
BSD AF_UNIX sockets are intrinsically racy, so that fchmod() has
no real advantages vs. chmod().

: o IPv6 calls set_multicast_list in the wrong context

Actually IPv6 calls it in correct context.
set_multicast_list must be bh protected, otherwise it is not
serialized wrt transmission. It is all another calls that
are made from wrong context now. I have patch.

Alexey

-
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/