[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-devel] please revert c/s 17686

On Friday, June 13, 2008 3:51 PM, Keir Fraser wrote:
> On 13/6/08 08:45, "Jan Beulich" <jbeulich@xxxxxxxxxx> wrote:
>>> Will you keep the 10ms tick in this case? If that's acceptable it
>>> should be a simple patch.

We have similar considerations as Jan mentioned below. We will try
larger interval & dynamically enable/disable.

>> I think it would be nice it the tick was enabled only when at least
>> CPU actually is about to enter or in C3. And I'm not certain whether
>> it wouldn't be possible to use a larger value than 10ms - at least in
>> case where all CPUs are in C3 (but I see that this case doesn't
>> seem to be expected anyway, given the warning handle_hpet_broadcast()
>> generates when the current CPU is in the channel's mask; I'm also
>> unclear about how the warning is avoided when the CPU currently in
>> charge of handling the timer interrupt is to enter C3 - maybe I'm
>> overlooking a place where the affinity get changed).

For the current implementation, the hpet_broadcast_exit() will be
executed before irq enabled, so the handle_hpet_broadcast() will always
get executed after the mask was cleared. We will look into whether it is
better to move hpet_broadcast_exit() after local_irq_enable().

> I missed that warning printk. It does indeed look odd.

As to this warning printk, we can simply replace it with an assert.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.