Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 24 Jul 2014 09:36:52 -0400
From:      Ryan Stone <rysto32@gmail.com>
To:        Bryan Drewery <bdrewery@freebsd.org>
Cc:        "freebsd-fs@freebsd.org" <fs@freebsd.org>
Subject:   Re: NFS DEADLKRES panic with getfsstat() on unresponsive mount
Message-ID:  <CAFMmRNzOBpM-x8z=jY9LXb2fgmq-MqJqL9uoQLDscsu_GF5Avw@mail.gmail.com>
In-Reply-To: <53CED455.1030701@FreeBSD.org>
References:  <53CED455.1030701@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Jul 22, 2014 at 5:15 PM, Bryan Drewery <bdrewery@freebsd.org> wrote:
> I had an NFS server go offline for about an hour and had a client panic in
> deadlkres after a 'zfs list -o mounted' ran which called getfsstat(2). It
> seems reasonable that this could be a false-positive. Is deadlkres known to
> have acceptable false-positives? If so, is there a way to mark this one to
> not cause a panic?
>
> This was r268621.

There is a list of "blessed" sx lock names in sys/kern/kern_clock.c.
deadlkres will ignore any threads sleeping with one of those wmesgs.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAFMmRNzOBpM-x8z=jY9LXb2fgmq-MqJqL9uoQLDscsu_GF5Avw>