I want to initiate a shutdown if the temperature gets too high. I have been using sensorsd(8), but sensorsd(8) only reacts once to the "high" (or low) event, leaving it up to the program/script to run timers to keep checking if the temperature gets worse. For my satisfaction, the timers would have to keep running until the system cooled down below the "high" temperature, so that sensorsd(8) will pick up the monitoring from there.
When the temperature gets to a warning level, I would like sensorsd(8) to notify logged in users (me), mail root, step down the CPU with apm -L, and then let the kernel do a shutdown, with acpitz(4), if the temperature continues to rise to critical. This would be easier and more simple for me than using sensorsd(8) alone (no timers). I checked this out a little bit today. Some laptop manufacturers release Windows programs to control these temperature settings. I don't know if the setting is permanent/saved in BIOS, but if it is then I could run it from a Windows Livecd to reset the critical temperature. Another idea was installing Coreboot (free-bios), but I doubt my mainboard is supported, and it could brick my system. Or, configure the OpenBSD kernel to ignore the BIOS setting, and use my hard coded temperature instead. Or, use sensorsd(8) and a script. On Mon, Jun 18, 2012 at 9:35 AM, Mike Larkin <[email protected]> wrote: > On Mon, Jun 18, 2012 at 06:35:58AM -0700, Robert Connolly wrote: > > Hello. > > > > During boot I see: > > acpitz0 at acpi0: critical temperature is 200 degC > > > > The acpitz(4) man page mentions that the system will power down if this > > critical temperature is reached. I assume this temperature is retrieved > > from BIOS, but I do not have an option in BIOS setup for it. > > > > Can I hard code this temperature in sys/dev/acpi/acpitz.c to a saner > > number? If so, it looks like I need to define sc->sc_crt, or possibly > _CRT. > > > > Or is there another way to do this? > > > > Thanks > > > > Why do you want to do this? > > -ml
