Re: Still oopsing in ip_queue_xmit (2.2.0-pre7)?

root (root@mail.securities.com)
Fri, 29 Jan 1999 03:08:34 -0500 (EST)

Thanks for the responses.
I've checked this function (ANK was first to ask about this) and amazingly
it is exactly as pointed by him and you, and still I can reproduce the
oops (it take on that box anything between few hours and 3 days - this was
max. uptime I had with pre7).

I've upgraded that machine to 2.2.1 but for now it's too early for me to
say anything about stability (I mean servers in my company are being
really hammered by traffic and load). I bet that if that bug is still
there (2.2.1) I can trigger it on either that box, or another server I
have here (it's running at this time 2.1.125, but will be upgraded next
week with both software - 2.2.1 and hardware - quad xeon + 1gig of ram).

Thanks again!

-- Tomek,

On Thu, 28 Jan 1999, David S. Miller wrote:

>
> Look at linux/net/ipv4/tcp_timer.c in the kernel source, the function
> tcp_delack_timer should read exactly as:
>
> void tcp_delack_timer(unsigned long data)
> {
> struct sock *sk = (struct sock*)data;
>
> if(!sk->zapped &&
> sk->tp_pinfo.af_tcp.delayed_acks &&
> sk->state != TCP_CLOSE) {
> /* If socket is currently locked, defer the ACK. */
> if (!atomic_read(&sk->sock_readers))
> tcp_send_ack(sk);
> else
> tcp_send_delayed_ack(&(sk->tp_pinfo.af_tcp), HZ/10);
> }
> }
>
> The important part is the check of sk->sock_readers, if you do not
> have this check, upgrade. pre7 has several tcp bugs, you might as
> well squash them all at once by using 2.2.1 if you can.
>
> Later,
> David S. Miller
> davem@dm.cobaltmicro.com
>

-
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/