On Tue, Jul 10, 2018 at 09:44:33AM +1000, Benjamin Herrenschmidt wrote: > On Sat, 2018-07-07 at 18:48 +0200, Greg Kroah-Hartman wrote: > > No, kobject_get() should never happen on a 0 refcount object. That > > being said, the code does allow it, so if things are messed up, it will > > happen. I think that change happened when the switch to refcount_t > > occured, before then we would WARN_ON() if that ever happened. I should > > go fix that up, and restore that old behavior, so that syzbot starts > > complaining loudly when stuff like that hits. > > > > So I hate using kobject_get_unless_zero(), and resisted ever adding it > > to the tree as it shows a bad locking/tree situation as you point out > > here. But for some reason, the block developers seemed to insist they > > needed it, and so it is in the tree for them. I don't want it to spread > > if at all possible, which makes me want to reject this patch as this > > should be "a case that can never be hit". > > Except it can in that situation... at least unless you get my patch 2/2 > (or the newer one I'm about to send that avoids adding a child counter > and uses the one in kernfs instead).
To follow up on this. I've applied the 2/2 patch for this series, so this 1/2 "should" not be needed. Ben, if you still see this trigger with that, I guess I can take this, but it still feels wrong to me :) thanks, greg k-h