On Wed, May 12, 2021 at 6:58 PM Wanpeng Li <kernellwp@xxxxxxxxx> wrote: > > From: Wanpeng Li <wanpengli@xxxxxxxxxxx> > > Inspired by commit 262de4102c7bb8 (kvm: exit halt polling on need_resched() > as well), CFS_BANDWIDTH throttling will use resched_task() when there is just > one task to get the task to block. It was likely allowing VMs to overrun their > quota when halt polling. Due to PPC implements an arch specific halt polling > logic, we should add the need_resched() checking there as well. > > Cc: Ben Segall <bsegall@xxxxxxxxxx> > Cc: Venkatesh Srinivas <venkateshs@xxxxxxxxxxxx> > Cc: Jim Mattson <jmattson@xxxxxxxxxx> > Cc: David Matlack <dmatlack@xxxxxxxxxx> > Cc: Paul Mackerras <paulus@xxxxxxxxxx> > Cc: Suraj Jitindar Singh <sjitindarsingh@xxxxxxxxx> > Signed-off-by: Wanpeng Li <wanpengli@xxxxxxxxxxx> > --- > v1 -> v2: > * update patch description > > arch/powerpc/kvm/book3s_hv.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c > index 28a80d2..6199397 100644 > --- a/arch/powerpc/kvm/book3s_hv.c > +++ b/arch/powerpc/kvm/book3s_hv.c > @@ -3936,7 +3936,8 @@ static void kvmppc_vcore_blocked(struct kvmppc_vcore *vc) > break; > } > cur = ktime_get(); > - } while (single_task_running() && ktime_before(cur, stop)); > + } while (single_task_running() && !need_resched() && > + ktime_before(cur, stop)); Consider moving this condition to a helper function that can be shared between book3s and the generic halt-polling loop. > > spin_lock(&vc->lock); > vc->vcore_state = VCORE_INACTIVE; > -- > 2.7.4 >
