[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20061213.162235.21925707.davem@davemloft.net>
Date: Wed, 13 Dec 2006 16:22:35 -0800 (PST)
From: David Miller <davem@...emloft.net>
To: leigh@...inno.co.uk
Cc: netdev@...r.kernel.org
Subject: Re: kernel panic with the newly merged md5 signature support
From: "Leigh Brown" <leigh@...inno.co.uk>
Date: Wed, 13 Dec 2006 23:43:50 -0000 (GMT)
> I just had to kill the parameter:-
>
> --- net/ipv4/tcp.c.orig 2006-12-10 21:46:26.000000000 +0000
> +++ net/ipv4/tcp.c 2006-12-13 17:38:29.000000000 +0000
> @@ -2362,7 +2362,7 @@ struct tcp_md5sig_pool *__tcp_get_md5sig
> EXPORT_SYMBOL(__tcp_get_md5sig_pool);
>
> void __tcp_put_md5sig_pool(void) {
> - __tcp_free_md5sig_pool(tcp_md5sig_pool);
> + tcp_free_md5sig_pool();
> }
>
> EXPORT_SYMBOL(__tcp_put_md5sig_pool);
>
> It still seems to me that it shouldn't be necessary to alloc and free
> the pool from tcp_(get|put)_md5sig_pool, as you'll always have alloc'd
> the pool before calculating the hash, but I'm happy enough as at least
> its working now.
Once a socket adds an MD5 signature, there will be at least a
single reference, so it won't get release until all sockets with
MD5 signatures are closed and released.
The quick hold/release cases are for transitory references to the
per-cpu MD5 pool, such as when sending resets or SYN+ACKs
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists