On 17.04.19 15:20, Paolo Bonzini wrote: > On 16/04/19 20:01, Christian Borntraeger wrote: >> v1->v2: >> - add kvm_vcpu parameter to kvm_arch_no_poll >> - add kvm stat in the s390 implementation >> - make the value tunable >> >> Folks, >> >> this is a very simple variant to disable halt polling when the KVM host >> is already running virtualized. We could imagine more complex variants >> (like tuning down the halt polling value) but this seems to do the trick >> for some kvm deployment scenarios on s390x. (e.g. having multiple LPARS >> with KVMs that are in itself already overcommitted). > > Ok to queue this for 5.2? Yes, after review. I can queue that in the s390 tree if you want. > (Stupid question, I thought LPARs are partitions and as such were never > overcommitted. What is it that separates an LPAR from a zVM or KVM guest?) LPAR can do CPU overcommitment as well (or dedicated whatever you configure). Its the memory that is hard partitioned on LPAR.
