Re: A lockd bug in linux 2.1.127

H.J. Lu (hjl@lucon.org)
Mon, 9 Nov 1998 08:05:53 -0800 (PST)

>
>
>
> On Sun, 8 Nov 1998, H.J. Lu wrote:
> >
> > schedule_timeout: wrong timeout value xxxxx from xxxxx
> >
> > It looks like timeout here can be negative in lockd. Is there a typo?
>
> No, the problem is that "nlmsvc_retry_blocked()" sometimes returns 0, and
> sometimes returns a jiffy. Making "nlmsvc_retry_blocked()" always return a
> timeout should fix the problem. Does something like this fix it for you?
>
> Linus

Thanks. It seems to work.

>
> -----
> diff -u --recursive v2.1.127/linux/fs/lockd/svc.c linux/fs/lockd/svc.c
> --- v2.1.127/linux/fs/lockd/svc.c Sun Nov 8 14:03:06 1998
> +++ linux/fs/lockd/svc.c Sun Nov 8 13:58:53 1998
> @@ -135,7 +135,7 @@
> * during grace period).
> */
> if (!nlmsvc_grace_period) {
> - timeout = nlmsvc_retry_blocked() - jiffies;
> + timeout = nlmsvc_retry_blocked();
> } else if (nlmsvc_grace_period < jiffies)
> nlmsvc_grace_period = 0;
>
> diff -u --recursive v2.1.127/linux/fs/lockd/svclock.c linux/fs/lockd/svclock.c
> --- v2.1.127/linux/fs/lockd/svclock.c Wed Aug 26 11:37:41 1998
> +++ linux/fs/lockd/svclock.c Sun Nov 8 14:02:19 1998
> @@ -615,6 +615,7 @@
> }
>
> if ((block = nlm_blocked) && block->b_when != NLM_NEVER)
> - return block->b_when;
> - return 0;
> + return (block->b_when - jiffies);
> +
> + return MAX_SCHEDULE_TIMEOUT;
> }
>
>
> 

-- 
H.J. Lu (hjl@gnu.org)


-
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/