On 08/17/2016 12:36 PM, Dragan Stancevic wrote: > Acked-by: dragan.stancevic@canonical.com > > > I agree with your patch, I have been tracking an issue where multipathd > dumps core on the exit path just past the treads being canceled. Your > patch is very similar to mine (minus nuking the depth) that I was going > to send out to a user to test with. The checker thread accesses a valid > pointer with garbage values.... Hello Dragan, When I prepared my patch I overlooked that multipathd creates most threads as detached threads. So I started testing the three attached patches on my setup. It would be appreciated if you could have a look at these patches. Thanks, Bart.