Date: Sat, 12 Apr 2003 21:08:53 -0600 (MDT) From: "M. Warner Losh" <imp@bsdimp.com> To: hsu@freebsd.org Cc: harti@freebsd.org Subject: Re: LOR in if_detach Message-ID: <20030412.210853.23179827.imp@bsdimp.com> In-Reply-To: <0HD7006Z9O9T0N@mta5.snfc21.pbi.net> References: <20030411175926.U1137@beagle.fokus.fraunhofer.de> <0HD7006Z9O9T0N@mta5.snfc21.pbi.net>
next in thread | previous in thread | raw e-mail | index | archive | help
In message: <0HD7006Z9O9T0N@mta5.snfc21.pbi.net>
Jeffrey Hsu <hsu@freebsd.org> writes:
: > Do you think of something like the following?
:
: > static int
: > foo_detach(device_t dev)
: > {
: > ...
: >
: > LOCK(softc);
:
: The other big race here is between detach and the next attach.
: This is better solved with a lock at the interface layer rather
: than on the softc.
I'm working on locking at the newbus level, which should solve that
problem. Since all changes to the dev tree are going to be under a
global lock, this race is eliminated.
My big problem is LOR when I enable this code in my tree :-(
Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030412.210853.23179827.imp>