Date: Mon, 25 May 1998 06:59:01 +0900 (JST) From: Michael Hancock <michaelh@cet.co.jp> To: Terry Lambert <tlambert@primenet.com> Cc: scrappy@hub.org, freebsd-current@FreeBSD.ORG Subject: Re: May 17th UP machine 'panic' Message-ID: <Pine.SV4.3.95.980525064410.4856A-100000@parkplace.cet.co.jp> In-Reply-To: <Pine.SV4.3.95.980524140422.3345B-100000@parkplace.cet.co.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 24 May 1998, Michael Hancock wrote:
> > The problem is in the vput in quotaoff for the group quota file
> > when doing a quotaoff on the user quota file.
> >
> > Basically, the locking semantics of vget/vput pairings have apparently
> > become assymetric?
>
> vget() is confusing, because the lock flag means don't lock when 0, but
> lock if LK_{lock flags} are passed. vput() will always unlock. But I
> don't think this is the problem.
Tor or Julian might have a patch for this soon.
Tor might have identified the problem. vput() doesn't take a process
argument and always uses curproc. I suggested splitting up the offending
vput() into vrele() and VOP_UNLOCK().
So I guess you are right, vput() is assymetric in that it doesn't take a
proc arg, but this would be too much work to fix. I think separating the
call to vput() into calls to vrele() and VOP_UNLOCK() only in cases where
it's needed would be sufficient.
Regards,
Mike Hancock
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.SV4.3.95.980525064410.4856A-100000>