On Wed, Jun 03, 2015 at 12:03:53PM -0700, Scott Feldman wrote: > On Wed, Jun 3, 2015 at 11:27 AM, Andy Gospodarek > <go...@cumulusnetworks.com> wrote: > > On Wed, Jun 03, 2015 at 11:15:55AM -0700, Scott Feldman wrote: > >> On Tue, Jun 2, 2015 at 8:07 PM, Andy Gospodarek > >> <go...@cumulusnetworks.com> wrote: > >> > This feature is only enabled with the new sysctl set (default is off): > >> > net.core.kill_routes_on_linkdown = 1 > >> > >> One more thing, sorry. This feature is typically implemented today in > >> user-space on a per-interface basis. The example I'm thinking of is > >> Quagga's "link-detect" directive which goes on an interface. Should > >> this be a bool on each interface in systcl? That would let user not > >> enable on selected interfaces. > > That would not be my preference. > > > > I'm willing to investigate the per-namespace support if Hannes would > > like and add switchdev support for v2, but would prefer this not become > > that granular. > > > Is there a reason? Seems you moving a feature from user space to the > kernel, so I'm curious why you wouldn't want to mimic functionality > that's available today. To me it seemed like a feature that is more globally useful so I wanted to just have it that way to start. I'm not opposed to per-device granularity in the future, but I did not want to differentiate right out of the gate when there are enough other moving parts.
> I like the patch/feature, BTW. Thanks for doing this. Glad to hear it! :) -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
