Re: EeePC, 5.0: acpitz gets wrong temperature

2011-11-05 Thread Frédéric Perrin
Hi Henri,

Le mercredi 2 C  20:48, Henri Kemppainen a C)crit :
>>> I just installed a snapshot (dated Oct 19) of -current on a new
>>> EeePC 1001PXD. The installation itself went fine. However, on the
>>> first boot, even before I can see the login prompt, acpitz decides
>>> to shutdown the machine:
>>> acpitz0: critical temperature exceeded 255C (5282K), shutting down
>
>> `dmesg' with this patch is at:
>> 
>
> From your dmesg:
>> acpiec _REG failed, broken BIOS
>
> ASUS has a BIOS update (0702; 2011.04.13) with the this description:
>> Update EC firmware
>
> I'd try that.  Tell me if it fixes the problem.

Thanks for the suggestion. I updated the BIOS. The only relevant
difference I see in dmesg is the new version number of the BIOS; I still
have the message about a broken BIOS and acpitz finding the temperature
too high. For the record:



I've been paying closer attention to the dmesg from Linux and FreeBSD,
and indeed, in both cases, I see messages about flaky things in the
BIOS. However, both Linux and FreeBSD are able to see the correct
temperature. For instanace, under FreeBSD:

# sysctl -a
...
hw.acpi.thermal.tz0.temperature: 57.0C




Linux: 

[0.168433] ACPI: EC: EC description table is found, configuring boot EC
(I don't know if the EC table was available before the BIOS update; all
I know is that Linux was able to get the correct temperature/)

[0.249652] [Firmware Bug]: ACPI: BIOS _OSI(Linux) query ignored
...
[0.272921]  pci:00: Requesting ACPI _OSC control (0x1d)
[0.272930]  pci:00: ACPI _OSC request failed (AE_NOT_FOUND), returned
control mask: 0x1d

Please do tell me if there is more relevant info I can provide. I am far
from an expert in ACPI; I have no idea what an EC table is, for
instance, and maybe I'm missing something obvious.

--
Fred



Re: EeePC, 5.0: acpitz gets wrong temperature

2011-11-02 Thread Henri Kemppainen
> > I just installed a snapshot (dated Oct 19) of -current on a new EeePC
> > 1001PXD. The installation itself went fine. However, on the first boot,
> > even before I can see the login prompt, acpitz decides to shutdown the
> > machine:
> > acpitz0: critical temperature exceeded 255C (5282K), shutting down

[..]

> `dmesg' with this patch is at:
> 

>From your dmesg:
> acpiec _REG failed, broken BIOS

ASUS has a BIOS update (0702; 2011.04.13) with the this description:
> Update EC firmware

I'd try that.  Tell me if it fixes the problem.



Re: EeePC, 5.0: acpitz gets wrong temperature

2011-11-02 Thread Frédéric Perrin
Hello,

Le samedi 22 C  11:15, FrC)dC)ric Perrin a C)crit :
> I just installed a snapshot (dated Oct 19) of -current on a new EeePC
> 1001PXD. The installation itself went fine. However, on the first boot,
> even before I can see the login prompt, acpitz decides to shutdown the
> machine:
> acpitz0: critical temperature exceeded 255C (5282K), shutting down

Same thing happens with 5.0. If somebody wants to investigate:

`dmesg' from 5.0's ramdisk is at :


`dmesg' from 5.0's kernel with ACPI disabled is at:

`sysctl -a' is at:


I applied the following, and rebooted without disabling ACPI:

--- acpitz.c.~1.43.~2011-06-25 07:35:06.0 +0200
+++ acpitz.c2011-11-01 19:40:24.0 +0100
@@ -326,11 +326,11 @@ acpitz_refresh(void *arg)
}
/* critical trip points */
if (sc->sc_crt != -1 && sc->sc_crt <= sc->sc_tmp) {
-   /* do critical shutdown */
-   printf("%s: critical temperature exceeded %dC (%dK), shutting "
-   "down\n",
-   DEVNAME(sc), KTOC(sc->sc_tmp), sc->sc_tmp);
-   psignal(initproc, SIGUSR2);
+   printf("%s: temperature %dC (%dK) >= %dC (%dK)\n",
+   DEVNAME(sc), KTOC(sc->sc_tmp), sc->sc_tmp,
+   KTOC(sc->sc_crt), sc->sc_crt);
+   printf("%s: for the record, _HOT is %dC (%dK)\n",
+   DEVNAME(sc), KTOC(sc->sc_hot), sc->sc_hot);
}

acpitz spams about every 10 seconds the following:

acpitz0: temperature 255C (5282K) >= 98C (3712K)
acpitz0: for the record, _HOT is -273C (-1K)

`dmesg' with this patch is at:


`sysctl -a' is at:


Is there anything else I can try?

> Of course, the temperature of the laptop is nowhere near 255C. This
> happened with both amd64 and i386. I tried again this morning after
> shutting down the laptop last evening, so the temperature should have
> been much above 20C. Under Linux, I see the following (temperatures in
> milliCelsius):
>
> % uname -or
> 3.0.0-1-amd64 GNU/Linux
> % grep '' /sys/class/thermal/thermal_zone*/*
> /sys/class/thermal/thermal_zone0/cdev0_trip_point:1
> /sys/class/thermal/thermal_zone0/cdev1_trip_point:1
> /sys/class/thermal/thermal_zone0/mode:enabled
> /sys/class/thermal/thermal_zone0/temp:5
> /sys/class/thermal/thermal_zone0/trip_point_0_temp:98000
> /sys/class/thermal/thermal_zone0/trip_point_0_type:critical
> /sys/class/thermal/thermal_zone0/trip_point_1_temp:95000
> /sys/class/thermal/thermal_zone0/trip_point_1_type:passive
> /sys/class/thermal/thermal_zone0/type:acpitz


--
Fred