Hi,
I'm now designing new vCPU scheduler in Xen, and trying to implement
the scheduler based on the Credit scheduler in Xen-4.5.1. But I encountered
come problems when debuging the code.
Most of the code modification is done in function csched_schedule() in
file: xen/common/csched_schedule.c . And the core code is as followed:
if( vcpu_runnable(current) )
{
if( match the migration contition )
{
cpu_affinity = pick_pcpu_runq(); // this function is defined
by myself
pcpulock = pcpu_schedule_lock_irqsave(cpu_affinity,
&pcpulock_flag);
TRACE_3D(TRC_CSCHED_STOLEN_VCPU, cpu_affinity , domain_id,
vcpu_id);
SCHED_VCPU_STAT_CRANK(scurr, migrate_q);
SCHED_STAT_CRANK(migrate_queued);
WARN_ON(scurr->vcpu->is_urgent);
scurr->vcpu->processor = cpu_affinity;
__runq_insert(cpu_affinity, scurr);
pcpu_schedule_unlock_irqrestore(pcpulock, pcpulock_flag,
cpu_affinity );
}
else
__runq_insert(cpu, scurr);
}
else
BUG_ON( is_idle_vcpu(current) || list_empty(runq) );
I try to run the modified Xen. But according to the log I found that,
although I insert the vCPU into the runqueue of another pCPU, the
vCPU still appears at the old pCPU in the following scheduling period.
Now I have a few questions:
1. Does the Xen scheduler framework support changing the pCPU of a
vCPU after using out the scheduling time slice, but not just to steal one
vCPU from runqueue of other pCPU in load_balance period?
2. If yes, what status of the vCPU should be changed before inserting
the vCPU into the destination pCPU?
Thank you very much!
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
