On Thu, Oct 1, 2009 at 6:21 PM, Thomas Backlund <tmb@xxxxxxxxxxxx> wrote: > Luca Tettamanti wrote: >> >> On Thu, Oct 01, 2009 at 08:40:59AM -0600, Robert Hancock wrote: >>> >>> On Thu, Oct 1, 2009 at 12:50 AM, Jean Delvare <khali@xxxxxxxxxxxx> wrote: >>>> >>>> On Wed, 30 Sep 2009 17:38:21 -0600, Robert Hancock wrote: >>>>> >>>>> On Tue, Sep 29, 2009 at 8:49 AM, Thomas Backlund <tmb@xxxxxxxxxxxx> >>>>> wrote: >>>>>> >>>>>> Try this one: >>>>>> http://marc.info/?l=linux-kernel&m=125421276407283&w=2 >>>>> >>>>> With that patch it's lasted almost a day with no timeouts, previously >>>>> I was getting about one every hour. Has that patch been submitted? >>>> >>>> Do you mean you're still getting timeouts but they are less frequent, >>>> or you did not get any timeout at all so far? >>> >>> So far no timeout at all (after about 36 hours). >> >> Good. I have a new patch for you then :P >> >> This version checks whether the EC is already enabled or not before >> touching it >> and also restore the state when the module is unloaded. >> You should check that the driver keeps working when is unloaded and >> reloaded. >> > > I tried your latest patch on a P7P55D Deluxe and get this: >> >> [ 10.419119] ACPI Error: Field [PR11] at 64 exceeds Buffer [NULL] size >> 32 (bits) (20090903/dsopcode-596) >> [ 10.419125] ACPI Error (psparse-0537): Method parse/execution failed >> [\_SB_.PCI0.SBRG.ASOC.GITM] (Node ffff88023f8147a0), AE_AML_BUFFER_LIMIT >> [ 10.419157] ATK0110 ATK0110:00: GITM[0x11060004] ACPI exception: >> AE_AML_BUFFER_LIMIT >> [ 10.419158] ATK0110 ATK0110:00: Unable to query EC status >> [ 10.419161] ATK0110: probe of ATK0110:00 failed with error -5 I see. GITM probably expects the same buffer structure as SITM, though in older models the upper bits were never used (hence I never noticed the problem). Working on a patch. L _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
