cvs commit: src/sys/kern sched_ule.c

[David Xu]

Jeff Roberson wrote:

>> if idlethread is preempted, who will clear its idle bit in 
>> idle_cpus_mask ?
> 
> 
> idle_cpus_mask was broken before for all schedulers.  This commit didn't 
> change that.  ULE doesn't use idle_cpus_mask and it's idlethread doesn't 
> set or clear it.  The idle thread for the other schedulers remains 
> unchanged.
> 
I havn't read ULE code, but 4BSD has following code:

void
sched_idletd(void *dummy)
{
         struct proc *p;
         struct thread *td;
#ifdef SMP
         cpumask_t mycpu;
#endif

         td = curthread;
         p = td->td_proc;
#ifdef SMP
         mycpu = PCPU_GET(cpumask);
         mtx_lock_spin(&sched_lock);
         idle_cpus_mask |= mycpu;
         mtx_unlock_spin(&sched_lock);
#endif
         for (;;) {
                 mtx_assert(&Giant, MA_NOTOWNED);


how about if the idlethread is preempted here ? mycpu is not cleared.


                 while (sched_runnable() == 0)
                         cpu_idle();

                 mtx_lock_spin(&sched_lock);
#ifdef SMP
                 idle_cpus_mask &= ~mycpu;
#endif
                 mi_switch(SW_VOL, NULL);
#ifdef SMP
                 idle_cpus_mask |= mycpu;
#endif
                 mtx_unlock_spin(&sched_lock);
         }
}

> Julian wants to set idle_cpus_mask in sched_switch() which would make it 
> accurate no matter how idlethread switched out.  that seems much more 
> reasonable to me.
> 

I haven't seen sched_switch clears it, so at least, it should be fixed 
for current 4BSD scheduler.

> Cheers,
> Jeff

Regards,
David Xu