On 04/28, Ingo Molnar wrote: > > * tip-bot2 for Oleg Nesterov <tip-bot2@xxxxxxxxxxxxx> wrote: > > > Another corner case is "nohz_full=0" on a machine with a single CPU or with > > the maxcpus=1 kernel argument. In this case non_housekeeping_mask is empty > > and tick_nohz_full_setup() makes no sense. And indeed, the kernel hits the > > WARN_ON(tick_nohz_full_running) in tick_sched_do_timer(). > > > > And how should the kernel interpret the "nohz_full=" parameter? It should > > be silently ignored, but currently cpulist_parse() happily returns the > > empty cpumask and this leads to the same problem. > > > > Change housekeeping_setup() to check cpumask_empty(non_housekeeping_mask) > > and do nothing in this case. > > So arguably the user meant NOHZ_FULL to be turned off - but it is de-facto > already turned off by the fact that there's only a single CPU available, > right? Or the user passes the empty "nohz_full=" mask on a multi-CPU machine. In both cases (before this patch) housekeeping_setup() calls tick_nohz_full_setup(non_housekeeping_mask) which sets tick_nohz_full_running = true even if tick_nohz_full_mask is empty. This doesn't look right to me and triggers the "this should not happen" warning in tick_sched_do_timer(). But let me repeat, I know nothing about nohz/etc. Oleg.
|