PRU example does not work

GoogleGroups
1

Hello everybody,

I just got my hands on a BB AI. During my testing I stumbled over an error in the Cloud 9 examples.I wanted to run the program blinkInternalLED.pru1_1.c of the PRU examples of the AI.

The following message appeares in the terminal output if I try to run it (note: I work with the cloud 9 IDE):

fatal error: pru_cfg.h: No such file or directory
#include <pru_cfg.h>

I updated the BB to the latest version. Did I miss something else to be able to run the example files? Or is the header for the pru realy missing?

Thanks alot
Stephan

2

Could you refer to which version of the image or code repository this applies?

I have a BBBlack.
Maybe I could try it.

3

Sry, I totally forgot about this one.

I just came back to playing around with the cloud9 examples. The pru examples work, if you take the right runner (which is C or C++ Beagle Makefile).

In the meantime, I did a few things to my BB AI, e.g. I patched the kernel to 4.14.108-ti-xenomai-r127. Now, everytime a halt() occurs in the code, I get an error:

  • Stopping PRU 1_1
    sh: echo: I/O error
    Cannot stop 1_1
    CC blinkExternalLED.pru1_1.c

Just tested this with the non xenomai kernel (4.14.108-ti-r127) and it doesn’t work either.
But, the installation with the newest image (Debian 9.9 lxqt, kernel 4.14.108-ti-r124), afterwards apt-get update/upgrade aswell as a git pull in the cloud9 directory works fine.

I did not make any changes to something pru related. Anyone else with the r127 release facing this?

Stephan

4

I did some further investigations.

It’s not a problem of the examples, I guess. If I directly address the PRUs via the linux system with

echo 'start' > /sys/class/remoteproc/remoteprocX/state

I get a write error: Invalid argument.
My COM connections says

remoteproc remoteprocX: Boot failed: -22

for such an event.

This occurs for all PRU cores except remoteproc1 (PRU1_1). If I start this one, the mentioned blink example starts.

I checked the present firmware and uevent entrys for every single core and they seem to be fine.

Has anyone a clue what may block the PRUs right from a restart?

Thanks
Stephan