On Tue, Dec 15, 2020 at 09:04:07PM -0800, Paul E. McKenney wrote: > Hello, Frederic, > > Are you seeing rcutorture writer stalls? Please see attached for an > example from testing, search for "Call Trace". I am running an overnight > test, which should get me some idea of frequency. My thought is to run > on the same version, but without the rcutorture.nocbs_nthreads=0 and > rcutorture.nocbs_toggle=1000 kernel boot parameters. > > This sort of error can occur when grace periods are progressing, but > callback processing is not. Things do seem to recover. > > Other thoughts?
Ok I'm trying to reproduce that. Thanks.