• 10 dec 2017: forum version update. In case of issues use this topic.
  • 30 nov 2017: pilight moved servers. In case of issues use this topic.
Hello There, Guest! Login Register

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Kernel Fault - Raspberry Pi 2
I have had the following error message three times recently:
Message from syslogd@raspberrypi at Apr 29 13:58:00 ...
kernel:[ 1257.134992] Disabling IRQ #49

Each time it happened when stopping the daemon with "sudo service pilight stop".

After the error it is still possible to start and stop the daemon, but there is no response from pilight-receive or pilight-debug.

It is necessary to restart the raspberry pi to get it working again. I am using a raspberry pi 2.

On two of the three times I was running the stable version, and recently the same fault whilst running the development version.
Can you please post:
- Linux Kernel / Firmware
- OS Distribution / Last Dist update/-grade
- SW Packages last update/-grade
- pilight version (sha code)
- Pi2 board revision
- Powersupply
Could you help me with the commands (where applicable) which would produce the information you are looking for. I can Google it, but if you have it at your finger tips it will allow me to get the results to you quickly.
(05-03-2015, 10:36 AM)simonsmart99 Wrote: Could you help me with the commands (where applicable) which would produce the information you are looking for. I can Google it, but if you have it at your finger tips it will allow me to get the results to you quickly.
Firmware / Kernel:
uname -a

git log

To perform an:
a) Application update/grade
apt-get update
apt-get upgrade

b) Distribution up-date/upgrade:
apt-get dist-upgrade

P2Board / Powersupply: Check the documentation you got with those devices.
Firmware / Kernel:
Linux raspberrypi 3.18.7-v7+ #755 SMP PREEMPT Thu Feb 12 17:20:48 GMT 2015 armv7l GNU/Linux

commit 2ce14f3c1e06569cf547453fc5d5fcfb3817d3c0
Author: CurlyMoo <curlymoo1@gmail.com>
Date:   Sat May 2 11:52:09 2015 +0200

    Only accept positive numbers

commit 782664327bbebee285ff9439631b3a3d75d090ba
Author: CurlyMoo <curlymoo1@gmail.com>
Date:   Sat May 2 10:43:34 2015 +0200

    Only check gpio on capable machine

commit e762ec32550f23b828d9c485ae8e696e4d84ab5f
Author: CurlyMoo <curlymoo1@gmail.com>
Date:   Fri May 1 18:38:06 2015 +0200

    Fixed small type on label action

commit 8a98001c33159a6604834a92abe21878bbac3d19
Author: CurlyMoo <curlymoo1@gmail.com>
Date:   Fri May 1 15:54:33 2015 +0200

    Allow functions and device properties in action arguments
    Before action arguments are passed to the actual function modules
    they are first parsed by the eventing library. This allows the usage
    of functions and device parameters in the action part of a rule.
    A few examples:
    IF 1 == 1 THEN dim DEVICE dimmer TO RANDOM(0, 10)
    IF 1 == 1 THEN sendmail TO ... SUBJECT Doorbell rang @ DATE_FORMAT(datetime,
    IF 1 == 1 THEN switch DEVICE switch TO on FOR RANDOM(0, 10) MINUTE

commit 68a963e4343e4ba7185afc622398b0b02ade2228
Author: CurlyMoo <curlymoo1@gmail.com>
Date:   Fri May 1 11:11:19 2015 +0200

    Prevent timezone shifts in date functions

Do you want me to perform an upgrade now?

And the power supply as per the attached .jpg.

Attached Files Thumbnail(s)
To me it seems that you get too many interrupts on IRQ #49.
It is easier said than done to catch the corresponding error message with dmesg.

Definitely you should upgrade the kernel.

What is the content of /proc/irq/49/spurious ?
Read more about Spurious Interrupts.
My spurious file has the following content:
count 4427
unhandled 0
last_unhandled 0 ms

As you can trace the error message back to shutting down pilight, the cause may be that interrupts are no longer handled and the kernel shuts down interrupt support for IRQ #49. (read more in the referenced link and google yourself a little bit).

It may be a bug due to the move to device trees.
- let us keep this out for the time being.

It may be a bug in the signal level of your Data IN line, or too much noise on your power supply or improper grounding of devices and blocking of voltage spikes.
(In short it may be that your equipment is creating a lot of noisy interrupts and when shutting down pilight the interrupt handler is gone .... and the kernel is doing its job to keep the reminder of the system healthy)

How do you connect your external hardware to your PI
What kind of hardware ?
Are you using the approved equioment ?
Do you use a LPF ?
Do you use resistors on your incoming data line ? (to ensure proper voltage levels ?
Do you use 5V or 3.3V for the receiver equipment ?
Do you use capacitors on the power lines for your external hardware to block spikes from your external hardware to cause problems ?
Do you use a proper star wiring structure on your ground cables ?
(one cable from the power supply to the PI, one cable to the LPF and a 100nF capacitor between GND and +, one cable from the LPF to the Receiver module and again a 100nF capacitor between GND and +, one cable from the LPF to the Transmitter module and again a 100nF capacitor between GND and +)

D) If you have other USB equipment is it connected yia a powered HUB ? If not does it improve if you do so (and vice versa) - You may have competing power supplies via the USB port, or you may unintentionally create grounding problems here, remove what you can remove in order to see if it fixes the problem.
My spurious file reads as follows:

count 90903
unhandled 0
last_unhandled 0 ms

which is interesting considering the error I received.

For the interim I will use another type of power supply, also from Modmypi, but there top end supply.

Regarding the TX and RX units, they are cheap units off of amazon:

The receiver has a 5v output, but I have dropped the voltage to 3.3v for the pi input 18.

I had ordered the full kit from pilight a while back now, so will wait until that arrives and implement according to spec at that point. Until then I will keep a record of the error if it happens again and let you know. Until the kit arrives I am running with no filter (CPU at around 11% on a pi2).

Thank you once again.
You should not allow that the input circuit of the PI's GPIO can draw more than 500µA. Thus put a 10K resistor in line (or inbetween) the output of the receiver and the GPIO input and operate the receiver with 5V.

Possibly Related Threads...
Thread Author Replies Views Last Post
  [Fixed] Segmentation fault when using Socket API wrong tomk 2 539 08-19-2019, 07:09 PM
Last Post: curlymo
  Odroid C2 support on Kernel 4.14 Jojo 5 1,315 06-01-2018, 03:19 PM
Last Post: curlymo
  433 mhz receiver working on Raspberry pi with a hack knex 3 2,612 01-21-2018, 06:28 PM
Last Post: curlymo
  AdHoc network with 2 Raspberry and 2 GPIO transmitter not working webdandy 29 9,677 09-06-2017, 02:39 PM
Last Post: wseifert
  glibc 2.24 / kernel 4.8.x breaks things on raspberry pi? Forumkel 0 1,040 12-11-2016, 02:25 PM
Last Post: Forumkel
  Cannot move configuration from old to new raspberry koos147 6 2,993 03-07-2016, 08:10 PM
Last Post: curlymo
Bug Bug report: segmentation fault using rule with AFTER woutput 3 3,065 11-22-2015, 12:27 PM
Last Post: woutput
  Raspberry Pi 2 darkday 26 16,016 05-07-2015, 11:18 AM
Last Post: curlymo
  Raspberry Pi Temperature Sensor Nelus82 3 2,706 03-22-2014, 11:48 PM
Last Post: creamers

Forum Jump:

Browsing: 1 Guest(s)