Mailinglist Archive: opensuse-bugs (6458 mails)
| < Previous | Next > |
[Bug 259992] kernel acpi read wrong temperature - critical shutdown
- From: bugzilla_noreply@xxxxxxxxxx
- Date: Tue, 29 May 2007 06:23:04 -0600 (MDT)
- Message-id: <20070529122304.DB62725C88C@xxxxxxxxxxxxxxxxxxxxxx>
https://bugzilla.novell.com/show_bug.cgi?id=259992
trenn@xxxxxxxxxx changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jdelvare@xxxxxxxxxx
Status|ASSIGNED |NEEDINFO
Info Provider| |shanti@xxxxxxx
------- Comment #15 from trenn@xxxxxxxxxx 2007-05-29 06:23 MST -------
Yes, looks scary, but could be harmless, if e.g. BIOS exports trip point
table(s), but the fans are still handled by BIOS.
I more expect wrong/weird EC values (110 C and 80 C values should be simply
wrongly reported), than real fan failure.
You may want to monitor this a bit:
watch -n1 cat /proc/acpi/fan/*/*
/proc/acpi/thermal_zone/*/{temperature,trip_points}
(Also use your ears for monitoring :), if fan control is done by BIOS, you can
at least guess from temperature and fan activity if things seem to work
normal).
To increase temp, you can e.g. do: "cat /dev/zero >/dev/null &" (this should
keep one core busy and let increase temperature or fan activity fast.
Do you still get the critical shutdowns with the smbus accessing module
disabled?
You should also be able to trigger this bug more often if you set
THERMAL_POLLING_FREQUENCY="1" in /etc/powersave/thermal and
/etc/sysconfig/powersave/thermal and restart powersaved (rcpowersaved restart).
You should then be able to verify the critical shutdown more easily (only if
the smbus module is loaded?).
Jean Delvare has some ideas about getting legacy sensor modules and ACPI
working together -> adding to CC.
Christoph: If you don't see critical shutdowns anymore without legacy smbus
modules, I'd like to close this one (hmm, it's worth to still make sure not
loading them per default for 10.2, maybe Daniel or Jean could give me a hint
how to do that, the modules should also taint the kernel IMO).
Christoph: It would also be very interesting to see how this machine behaves
with latest 10.3 Alpha (it's the 4th AFAIK). If you still have a free partition
on your machine, it would be great to give it a test. Now we still can fix
things, as soon as 10.3 is released it's getting difficult.
Christoph: Could you also attach acpidump output please.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
trenn@xxxxxxxxxx changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jdelvare@xxxxxxxxxx
Status|ASSIGNED |NEEDINFO
Info Provider| |shanti@xxxxxxx
------- Comment #15 from trenn@xxxxxxxxxx 2007-05-29 06:23 MST -------
Yes, looks scary, but could be harmless, if e.g. BIOS exports trip point
table(s), but the fans are still handled by BIOS.
I more expect wrong/weird EC values (110 C and 80 C values should be simply
wrongly reported), than real fan failure.
You may want to monitor this a bit:
watch -n1 cat /proc/acpi/fan/*/*
/proc/acpi/thermal_zone/*/{temperature,trip_points}
(Also use your ears for monitoring :), if fan control is done by BIOS, you can
at least guess from temperature and fan activity if things seem to work
normal).
To increase temp, you can e.g. do: "cat /dev/zero >/dev/null &" (this should
keep one core busy and let increase temperature or fan activity fast.
Do you still get the critical shutdowns with the smbus accessing module
disabled?
You should also be able to trigger this bug more often if you set
THERMAL_POLLING_FREQUENCY="1" in /etc/powersave/thermal and
/etc/sysconfig/powersave/thermal and restart powersaved (rcpowersaved restart).
You should then be able to verify the critical shutdown more easily (only if
the smbus module is loaded?).
Jean Delvare has some ideas about getting legacy sensor modules and ACPI
working together -> adding to CC.
Christoph: If you don't see critical shutdowns anymore without legacy smbus
modules, I'd like to close this one (hmm, it's worth to still make sure not
loading them per default for 10.2, maybe Daniel or Jean could give me a hint
how to do that, the modules should also taint the kernel IMO).
Christoph: It would also be very interesting to see how this machine behaves
with latest 10.3 Alpha (it's the 4th AFAIK). If you still have a free partition
on your machine, it would be great to give it a test. Now we still can fix
things, as soon as 10.3 is released it's getting difficult.
Christoph: Could you also attach acpidump output please.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
| < Previous | Next > |