On 10/09/2017 05:46 PM, Paul E. McKenney wrote: > On Mon, Oct 09, 2017 at 04:47:08PM +0200, Christian Borntraeger wrote: >> Paul, >> >> Michael J. Sullivan from our performance team told me that >> commit 9111b9361f2 EXP: Fine-grained timer diagnostics breaks cpu hot unplug for him. >> >> I can confirm that. > > Interesting! > > This commit is not, repeat -not-, ever going upstream. It is strictly > diagnostic for a timer problem that I hit from time to time in runs of > rcutorture. The fact that it breaks your tests should provide a clue > to me, but I am not immediately seeing what that clue might be. > > For whatever it is worth, here are the symptoms I am seeing: > [...] >> Reverting the patch fixes the issue, but I do not yet understand why. > > Welcome to my world! ;-) > > Hmmm... I have to ask... Have you tried this with lockdep? The spinning > on CPUs is suspicious, though not something that I have seen. Of course and with lockdep the problem goes away but no message comes up. Might be related to the fact that the timer lock is a raw spin lock, which does not have lockdep IIRC. -- To unsubscribe from this list: send the line "unsubscribe linux-s390" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html