My laptop is showing a highly irritating tendency of shutting itself off
after reading false temperatures. See the following entry
in /var/log/syslog :

Aug 5 00:00:03 localhost /usr/bin/crontab[8488]: (root) LIST (no
Aug 5 00:02:25 localhost kernel: [10312.760000] ACPI: Critical t
rip point
Aug 5 00:02:25 localhost kernel: [10312.760000] Critical tempera
ture reached (2159 C), shutting down.
Aug 5 00:02:25 localhost shutdown[11489]: shutting down for syst
em halt
Aug 5 00:02:25 localhost init: Switching to runlevel: 0
Aug 5 00:02:27 localhost kernel: [10314.787000] Critical tempera
ture reached (44 C), shutting down.

First it sees 2159C (which is impossible as silicon melts before that) and 2
seconds later it has cooled to 44C. I have seen other equally or more
unbelievable temperatures.

I first thought that the issue was with the hardware (sensors reporting junk
data). It is probably not - I have had the motherboard and the processor
replaced. And after a few days, these problems have returned.

Which suggests a software issue. Whatever it is, I want to disable shut
downs on ACPI temperature trip points entirely. Having your laptop shut
down because it is having a brain fart in the middle of work or a
presentation is not very amusing to say the least.

Any ideas ? If it takes reconfiguring the kernel (like disabling IPMI
poweroff) etc., I will happily do it. Suggestions welcome.