Hi!
Thanks for answering.
Eventually the system will be an embedded system (The OS is Linux 3.4.47 with an RT-patch).
The language is C or C++
It's not clear right now if overheating is an issue. So, at the moment I simply want to monitor the temprature.
The idea is simple : if the system overheats do something (Actually send a message for the moment).
I know that the temperature can be seen by "cat /sys/bus/acpi/devices/LNXTHERM\:00/thermal_zone/temp" (See
http://www.linuxquestions.org/questi...n-linux-598658)
There are 4 cores so there 4 entries to query. There is also the GPU to query (which I do not know how at the moment).
But how often should I check these fields ? Every second ? Every millisecond ? Every minute ?
So, I thought that a hook will be the best solution.
Thanks
zmau