Date: Wed, 20 Oct 2010 09:20:34 +0200 From: Bernhard Schmidt <bschmidt@freebsd.org> To: "Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r214070 - head/sys/dev/if_ndis Message-ID: <201010200920.34413.bschmidt@freebsd.org> In-Reply-To: <20101019220455.D46881@maildrop.int.zabbadoz.net> References: <201010191911.o9JJBaQe024210@svn.freebsd.org> <20101019220455.D46881@maildrop.int.zabbadoz.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday, October 20, 2010 00:06:09 Bjoern A. Zeeb wrote: > On Tue, 19 Oct 2010, Bernhard Schmidt wrote: > > Author: bschmidt > > Date: Tue Oct 19 19:11:36 2010 > > New Revision: 214070 > > URL: http://svn.freebsd.org/changeset/base/214070 > > > > Log: > > There is no reason to call rt_ifmsg(), remove it. > > While it's indeed very uncommon to have that in a driver, I don't know > how much black magic if_ndis(4) is. Have you verified that you still > get the events from elsewhere on the routing socket when > loading/unloading if_ndif.ko? I followed the function calls, and the affected ndis_linksts_done() function is only a wrapper around others. Those functions do all call if_link_state_change() which itself calls rt_ifmsg(). It seems the call is just a leftover from the the transition to if_link_state_change(). -- Bernhard
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201010200920.34413.bschmidt>