On Wed, 2014-04-09 at 14:02 -0400, Prarit Bhargava wrote:
>
> On 04/09/2014 01:59 PM, Matthew Garrett wrote:
> > So... why add the driver at all? Refusing to permit the kernel to touch
> > these resources is a deliberate design choice, because of the potential
> > for these races. It'll work absolutely fine right up until the point
> > where you end up reading the wrong temperature value from an i2c hwmon
> > chip and perform a critical thermal shutdown.
>
> Because it seems like the "right" thing to do is use the ACPI interface rather
> than the i801 pci driver. At least to me that makes sense. OTOH, if Jean
> thinks there isn't a need I can drop it.
We should use any ACPI interface that allows us to perform appropriate
synchronisation with the hardware, but there's no real advantage in
using an ACPI interface that's still racy.
--
Matthew Garrett <matthew.garrett@xxxxxxxxxx>
��.n��������+%������w��{.n�����{��-��)�� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
