Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 26 Jul 2007 06:53:58 -0400
From:      Kris Kennaway <kris@obsecurity.org>
To:        Anders Nordby <anders@FreeBSD.org>
Cc:        arch@FreeBSD.org, Robert Watson <rwatson@FreeBSD.org>
Subject:   Re: Removing NET_NEEDS_GIANT: first patch
Message-ID:  <20070726105358.GA43979@rot26.obsecurity.org>
In-Reply-To: <20070726102328.GA12293@fupp.net>
References:  <20070724110908.T83919@fledge.watson.org> <20070726102328.GA12293@fupp.net>

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

--9amGYk9869ThD9tj
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Jul 26, 2007 at 12:23:28PM +0200, Anders Nordby wrote:
> Hi,
>=20
> On Tue, Jul 24, 2007 at 11:17:45AM +0100, Robert Watson wrote:
> > Attached is the first of a series of patches to start removing the=20
> > NET_NEEDS_GIANT and debug.mpsafenet scaffolding.  This source code=20
> > declaration was used by optionally compiled components to declare a str=
ict=20
> > requirement for Giant, and forced Giant over the entire network stack. =
=20
> > debug.mpsafenet could also be set by users in loader.conf in order to=
=20
> > similar force Giant over the network stack, and existed for two reasons=
: to=20
> > allow Giant to be put back over the network stack for debugging purpose=
s,=20
> > and to support these recently removed or fixed unsafe components.  As s=
uch,=20
> > this patch removes the following:
> >=20
> > - NET_NEEDS_GIANT() macro
> > - debug.mpsafenet tunable/sysctl and associated debug_mpsafenet variabl=
e, as
> >   well as functions supporting these.
> > - Use of this variable to control acqusition of Giant in network-related
> >   interrupt handlers and various other paths.
>=20
> I've used and still use debug_mpsafenet to get rid of watchdog timeout
> problems on a lot of HP Proliant servers, particularly with the bge
> driver:
>=20
> Dec 21 06:42:51 videovm1 kernel: bge0: watchdog timeout -- resetting
> Dec 21 06:42:51 videovm1 kernel: bge0: link state changed to DOWN
> Dec 21 06:42:54 videovm1 kernel: bge0: link state changed to UP
>=20
> This problem goes away with debug.mpsafenet=3D"0", for me.
>=20
> I can try to turn off this setting, and see how it goes. I remember
> there was something one could do, to get more information about the
> watchdog error, but can't remember what.

Please do.  There is no sense in crippling your network for the sake
of an unre{solved,ported} driver bug.

> Another thing, bde did some watchdog fixes for bge in december 2006. I
> wonder why those didn't make it into 6.2, which was released in january
> 2007.

Too late, probably.

Kris

--9amGYk9869ThD9tj
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4 (FreeBSD)

iD8DBQFGqH1GWry0BWjoQKURAg+7AJ4rURsR7odNKC26Wvmr75aQc1EqjgCg9YA+
VATS3w3kORYCe9/5N5KTCqw=
=q9To
-----END PGP SIGNATURE-----

--9amGYk9869ThD9tj--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070726105358.GA43979>