Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 27 Jul 2012 10:55:21 -0400
From:      Andrew Boyer <aboyer@averesystems.com>
To:        attilio@FreeBSD.org
Cc:        FreeBSD Stable Mailing List <freebsd-stable@freebsd.org>, John Baldwin <jhb@freebsd.org>, Andriy Gapon <avg@freebsd.org>
Subject:   Re: IPMI hardware watchdogs Re: dell r420/r320 stable/9
Message-ID:  <4ECC422A-F7A8-4F6C-9E9D-01080927C36D@averesystems.com>
In-Reply-To: <CAJ-FndCsSZmKTif7Mn%2BZSMX1rbkgkjytRRY5viGimATDMm5RXQ@mail.gmail.com>
References:  <1343350238.12294.10.camel@powernoodle.corp.yahoo.com> <23294764-F30B-4732-8C41-3F0ECA5F273C@averesystems.com> <CAJ-FndCsSZmKTif7Mn%2BZSMX1rbkgkjytRRY5viGimATDMm5RXQ@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help

On Jul 27, 2012, at 10:42 AM, Attilio Rao wrote:

> On Fri, Jul 27, 2012 at 3:33 PM, Andrew Boyer =
<aboyer@averesystems.com> wrote:
>>=20
>> On Jul 26, 2012, at 8:50 PM, Sean Bruno wrote:
>>=20
>>> For the time being I had to revert the following from my stable/9 =
tree.
>>> Otherwise I would get a kernel panic on shutdown from ipmi(4).
>>>=20
>>> http://svnweb.freebsd.org/base?view=3Drevision&revision=3D237839
>>> http://svnweb.freebsd.org/base?view=3Drevision&revision=3D221121
>>>=20
>>=20
>> On a somewhat related note: We noticed recently that you can't pet or =
disable the IPMI hardware watchdog once SCHEDULER_STOPPED() is true.  =
This means it can fire unexpectedly while you're dumping core or =
rebooting, depending on how long the timeout was on the pet before the =
panic.  The ipmi driver will need to process the command differently if =
the scheduler is stopped.  I haven't had time to look at a fix yet.
>=20
> I recall I fixed that internally for SV, but the key here is that we
> need to find an unified (or a default policy).
> More specifically, do we want the watchdog also covers the kernel dump
> part (because of possible deadlocks when dumping). If the answer is
> yes, we likely need pat the watchdog from within the dumping cycle
> itself. If the answer is no, then we can just disable it when entering
> the panic path. But anyway, we need to identify a default policy that
> makes sense first.
>=20
> Attilio
>=20

For our use case, we need the system to reset if the dump hangs.

As the code stands now, you can't disable the HW watchdog from the panic =
path.  Prior to stopping the scheduler early in panic(), you don't know =
the lock state, so you can't safely initiate the IPMI command.  (It hung =
the first time I tried it.)  After stopping the scheduler, you can't pet =
it to turn it off.

-Andrew

--------------------------------------------------
Andrew Boyer	aboyer@averesystems.com







Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4ECC422A-F7A8-4F6C-9E9D-01080927C36D>