CC1101 died. Kernel worker thread takes up all CPU


It seems that my CC1101 has died after unplugging and plugging the raspi power connector. :cry:
Now I can see a kernel worker thread “kworker/0:1” which takes all CPU time if homegear is running.

Is there anything inside the homegear CC101 driver which would wait forever if something with the chip is not ok?
For example: Is there an endless loop while powering up the oscillator which waits until a stable condition is signalled?

I will go to Conrad today to buy a HM-CFG-USB-2. :unamused:

Hi nanosonde,

No, not in the Homegear source. But the SPI driver functions can block. That happens in normal operation on high interrupt load for example. The only place where Homegear is blocking is on reading the GPIO state. I’m using the C++ function “poll” on the GPIO. If the GPIO signals that a packet was received, I’m reading the packet. That’s basically it. You can enable debug level 6, to see the actual SPI communication in the Homegear log. There you can see if any sensible SPI communication happens at all (that’s as low level as you can get).




this is happend if you use the Modul from pollin.
I have the same problem with that. After change to a other modul no problems.
Regards Achim

It was no module from Pollin. It was a module from an old original Homematic device.