On Fri, Dec 09, 2016 at 07:22:06AM -0800, Eric Dumazet wrote: > On Fri, 2016-12-09 at 06:24 -0800, Eric Dumazet wrote: > > > It looks that you want a seqcount, even on 64bit arches, > > so that CPU 2 can restart its loop, and more importantly you need > > to not accumulate the values you read, because they might be old/invalid. > > Untested patch to give general idea. I can polish it a bit later today.
I'm preparing a patch now, so you can review it. Eric, thanks a lot for reviewing and proposing a working approach!
