> Date: Wed, 15 May 1996 14:30:00 +0300 (EET DST)
> From: Linus Torvalds <torvalds@cs.Helsinki.FI>
> To: Dan Merillat <harik@ao.net>
> Cc: Linux Kernel List <linux-kernel@vger.rutgers.edu>
> Subject: Re: A few questions and notes....
>
>
>
> On Wed, 15 May 1996, Dan Merillat wrote:
> >
> > Ok, I'm still running 1.3.99 (I know, obselete...:-) but I'm still
> > getting double-lock on socket at 00149b33 (tcp_sendmsg)
> >
> > Grr. Well, a judicious #if 0 in the source nuked that one for now. It
> > happens when I have a telnet session open that I havn't used in a while.
> > (Virtual consoles do that... :-) As soon as I type, I get a few of
> > those, then it stops.
>
> The "double lock on socket" message is harmless - I'll take it out for
> 2.0. It happens when there is swapping activity and the same socket is
> used concurrently by two programs - this was why I rewrote the socket
> locking code in the first place. That's why you see it on telnet sessions
> that have been idle - they've been swapped out.
Ok. I figured it was harmless.
> The reason the message is there at all is that I wanted to make sure I
> hadn't forgotten to unlock the socket anywhere. So the kernel prints it
> out even when it's valid, just in case... I think we can sleep safely
> knowing that the locking works correctly (it ahsn't been causing any
> probles for a _long_ time).
>
> So don't worry if you get this one - it's going away.
>
> (Note: the "double lock on device queue" message that sounds similar _is_
> a real bug, so if people see that one then tell the world)
How about "tcp_send_skb: attempt to queue a bogon"? Got those scrolling
so fast it wiped the machine. Rarely, though. And not recently (1.3.75
I believe was the last time.)
--Dan