>>>>> On Mon, 11 Apr 2005 18:10:02 -0700,
	Jonathan Stone <jonathan@dsg.stanford.edu> said:

>> are you sure that runqueue is protected by biglock?
>> afaik it's protected by sched_lock, not biglock.

> No, I'm not sure. I took Soda-san's word for it.

I wrote that article without closely looking at it... ;-)
Actually, I didn't notice that the kernel_lock isn't held,
if the interrupt priority is higher or equal to IPL_SCHED.
So, yamamoto-san is right.
--
soda