Date: Fri, 13 Nov 2015 23:26:38 +0000 From: Steven Hartland <steven@multiplay.co.uk> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r290450 - in head: sbin/ifconfig share/man/man4 sys/net Message-ID: <564671AE.1030405@freebsd.org> In-Reply-To: <20151113223627.GC73031@FreeBSD.org> References: <201511061533.tA6FXSmq054394@repo.freebsd.org> <20151113223627.GC73031@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 13/11/2015 22:36, Gleb Smirnoff wrote: > On Fri, Nov 06, 2015 at 03:33:28PM +0000, Steven Hartland wrote: > S> Author: smh > S> Date: Fri Nov 6 15:33:27 2015 > S> New Revision: 290450 > S> URL: https://svnweb.freebsd.org/changeset/base/290450 > S> > S> Log: > S> Add sysctl to control LACP strict compliance default > S> > S> Add net.link.lagg.lacp.default_strict_mode which defines > S> the default value for LACP strict compliance for created > S> lagg devices. > S> > S> Also: > S> * Add lacp_strict option to ifconfig(8). > S> * Fix lagg(4) creation examples. > S> * Minor style(9) fix. > > AFAIU, you are adding the sysctl to close a very short period > of time between cloning a lagg(4) and configuring it. > > This is not a pretty solution to the problem. But the network > stack has an API for supplying extra parameters at the cloning > time. The lagg_clone_create() accepts the params, that can > be supplied by ifconfig. If you put there the same arguments you > supply with SIOCGLAGGOPTS, you can create and configure a lagg > in one syscall. > In our case strict mode totally breaks for us on some old switches, the only option to restore functionality was to disable it so we use this loader.conf to ensure sysadmins don't mess up when creating new lagg's. So yes we could do that as well but it wouldn't fix the issue we're using the sysctl for. Regards Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?564671AE.1030405>