Date: Sat, 24 May 2003 10:14:57 -0400 (EDT) From: Robert Watson <rwatson@freebsd.org> To: Poul-Henning Kamp <phk@phk.freebsd.dk> Cc: CURRENT <freebsd-current@freebsd.org> Subject: Re: /dev/null panic still alive Message-ID: <Pine.NEB.3.96L.1030524101053.73987A-100000@fledge.watson.org> In-Reply-To: <20810.1053728868@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 24 May 2003, Poul-Henning Kamp wrote:
> >> Kirk's on vacation and I havn't heard back from rwatson what he think
> >> we should do.
> >
> >He did mention he thought trustedbsd had a patch for this.
>
> Yes, unfortunately, that patch which is also in -current does not work.
>
> It works for VFIFO's but not for VCHR
Yeah, the mixup in the buffer cache is going to cause us problems, and
doesn't look like something we want to attempt for 5.2-RELEASE. I think
there are two work-arounds we should put in place though:
(1) Return EOPNOTSUPP if you try an EA operation on a UFS specfs node.
(2) Suggest firmly that people use devfs; symlink from the Linux /dev to
the FreeBSD/dev, or mount a new devfs instance for Linux.
While there might not be time before 5.2-RELEASE, the Linux compat stuff
needs to recognize a devfs system and not use hard-coded devices -- that
will break more and more as we move away from device numbers.
I'm tesitng a patch for (1) and will send it out this evening for review.
Robert N M Watson FreeBSD Core Team, TrustedBSD Projects
robert@fledge.watson.org Network Associates Laboratories
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1030524101053.73987A-100000>