On Fri, 2022-04-08 at 11:20 +0000, Dario Faggioli wrote: > On Thu, 2022-04-07 at 15:27 +0200, Jan Beulich wrote: > > > > Credit2 moving the vCPU-s off of their initially assigned ones > > right > > away of course renders sched_select_initial_cpu()'s use of > > cpu_cycle() > > pretty useless. > > > Mmm... you mean that Credit2 is moving the vCPUs off they're assigned > ones _right_now_, or that it will, with this patch? > > If you mean the former, I'm not sure it is. In fact, when > sched_select_initial_cpu() is called for dom0, dom0's node affinity > is > just all nodes, isn't it? > Actually, it's what results from dom0_nodes. I was forgetting that we set d->node_affinity while parsing the parameter.
That said, I've added some logging, to better understand what is going on, and there are some not obvious (at least, not to me) things that are happening, and it seems that what is causing problems is the call to sched_insert_unit(), within sched_init_vcpu(). I have to leave now, but I will report what I've found ASAP (probably tomorrow, though). Regards -- Dario Faggioli, Ph.D http://about.me/dario.faggioli Virtualization Software Engineer SUSE Labs, SUSE https://www.suse.com/ ------------------------------------------------------------------- <<This happens because _I_ choose it to happen!>> (Raistlin Majere)
signature.asc
Description: This is a digitally signed message part
