On Tue, Jan 06, 2015 at 05:39:31PM -0800, Greg Kroah-Hartman wrote: > On Tue, Jan 06, 2015 at 05:00:20PM -0800, Linus Torvalds wrote: > > So this one isn't technically needed. It turns out the memory stomp can't > > actually happen, because while the code is ugly and wrong, the hlist_del_init > > will actually fix things up almost by accident. Although it might even have > > been on purpose by Al originally. > > > > So I'm not sure it needs back porting. It's *probably* a good idea, but the bug > > wasn't as bad as originally thought. > > > > Al added to cc just to keep him in the loop. It might be worth back porting > > just to keep stable similar to head wrt this very confusing code. Al, comments? > > I think it's probably worth keeping in sync just to keep the confusion > to a minimum, unless Al objects. No objections, except that I'd prefer more accurate commit message... FWIW, I still think we need to add __hlist_splice(), but that'll go as a followup to this one. I'll post it tonight or tomorrow morning. TBH, originally it was just a quick and dirty solution - back then I had convinced myself that everything worked (and it does), and left it as-is until later, then forgot about that ugliness until Eric stumbled across it and posted his patch. -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html