On 02/01/17(Mon) 21:51, Mike Belopuhov wrote: > On Fri, Dec 30, 2016 at 18:57 +0100, Mike Belopuhov wrote: > > On Thu, Dec 29, 2016 at 09:30 +0100, Martin Pieuchot wrote: > > > On 29/12/16(Thu) 01:15, Alexander Bluhm wrote: > > > > On Fri, Dec 23, 2016 at 12:09:32AM +0100, Martin Pieuchot wrote: > > > > > On 22/12/16(Thu) 20:45, Mike Belopuhov wrote: > > > > > > I think this is what is required here. Works here, but YMMV. > > > > > > > > > > splnet() in a pseudo-driver seems completely wrong, you could get rid > > > > > of > > > > > it. > > > > > > > > Yes, but that is another issue. Can we get the netlock splasserts > > > > fixed first? This diff looks good to me. > > > > > > Sure I'm ok with the diff. > > > > > > > I agree with Martin and have cooked a diff but couldn't test it yet. > > This is it for the reference. > > > > I got to test the diff and I had to make another adjustment: > vxlan_if_change is setup as a detach hook, however dohooks is > called very early in if_detach before we remove IP addresses > from the interface. It makes vxlan_config find these IP > addresses just fine and re-add its own detach hook again.
Why not fix vxlan_if_change()? > This > repeats ad infinitum hogging the machine. I couldn't think of > anything better than deferring an operation via a task. Seems > to do the trick. That's ugly. Why would you re-add anything in a detach hook? This is obviously broken.