OpenCPN Version 5.7.1 Beta Test #2

[Rae Simpson]

Possible clue on the cause of the hang up on exit issue:

Comparing OpnCPN logs in the problem Raspberry with a functioning windows OpenCPN log:

Raspberry OpenCPN log ends at ---> 17:47:59 890 'Message Chart1.cpp:4022 gFrame destroyed."

Windows OpenCPN.log continues with further cleanup message like:

16:50:34.819 MESSAGE chart1.cpp:4022 gFrame destroyed.
16:50:34.839 MESSAGE chart1.cpp:2634 opencpn::MyApp starting exit.
16:50:34.839 MESSAGE chart1.cpp:2673 LOGBOOK: 2023-01-16 00:50:34 UTC OFF: Lat 49.01080 Lon -123.82290
16:50:34.845 MESSAGE TCDataSource.cpp:41 UnLoading Tide/Current data source: C:\Program Files (x86)\OpenCPN\tcdata\harmonics-dwf-20210110-free.tcd
16:50:34.854 MESSAGE TCDataSource.cpp:41 UnLoading Tide/Current data source: C:\Program Files (x86)\OpenCPN\tcdata\HARMONICS_NO_US.IDX
16:50:34.857 MESSAGE chart1.cpp:2697 opencpn::MyApp exiting cleanly...

It sure looks like the issue is tide and/or current table related.

Turning tide / current tables on took ~15 minutes with heavy CPU during that time. After the Tides/currents are displayed, right clicking the current arrows triggers the heavy cpu and slow response. Same with exiting current table display boxes.

Thoughts?

[Hakan]

Originally Posted by Hakan

About RPi and modprob can0 crash at shut down


I though experienced if I start OCPN without modprobe installed he will crash on shut down.
So we better make sure the modprobe is installed in a way it reloads after reboot.

Just to inform this issue is now solved in current git master. Also was my wordings not correct. O didn't crashed, he hanged.


We may need a common convenient way to have the can0 driver installed permanently. Here we may need some help from all Debian gurus?
I've the MCS board on a another RPi so there all was installed by help of the GeDab SignalK-app.

Thanks.

[bdbcat]

Rae Simpson....
re:
"Turning tide / current tables on took ~15 minutes with heavy CPU during that time. After the Tides/currents are displayed, right clicking the current arrows triggers the heavy cpu and slow response. Same with exiting current table display boxes."


On RPi4/64/Bullseye, I cannot reproduce anything like this. Turning on Tide/Currents from lower right "hamburger" takes about .5 secs to fully redraw the (fully zoomed out) chart. Right click on a station brings up the data curve more or less instantly.


Please tell us more about your config. Anything potentially non-standard, or otherwise useful.
Anyone else seeing this effect?
Thanks
Dave

[verkerkbr]

Originally Posted by Hakan

Just to inform this issue is now solved in current git master. Also was my wordings not correct. O didn't crashed, he hanged.

We may need a common convenient way to have the can0 driver installed permanently. Here we may need some help from all Debian gurus?
I've the MCS board on a another RPi so there all was installed by help of the GeDab SignalK-app.

Thanks.

Hi Hakan,

I have given an instruction how to generate the can0 at the boot of a Linux system.
The instruction is in the etc/rc.local file. Has no effect on the working of the system if a NMEA2000 item is not connected. Sometimes rc.local is allready available but is not working. If it is not available you can simply add a file rc.local.

And adjust the rights.

This works very well with these small and cheap Canable boards. Not yet tried with the Actisense NGT-1 adapter. If you are interested I can try to make a setup with the Actisense.

Regards,

Bram (no Guru, only a user)

As I mentioned before the can0 does not show up in the Flatpak O version.

[kevinvh]

Originally Posted by bdbcat

Rae Simpson....
re:
"Turning tide / current tables on took ~15 minutes with heavy CPU during that time. After the Tides/currents are displayed, right clicking the current arrows triggers the heavy cpu and slow response. Same with exiting current table display boxes."


On RPi4/64/Bullseye, I cannot reproduce anything like this. Turning on Tide/Currents from lower right "hamburger" takes about .5 secs to fully redraw the (fully zoomed out) chart. Right click on a station brings up the data curve more or less instantly.


Please tell us more about your config. Anything potentially non-standard, or otherwise useful.
Anyone else seeing this effect?
Thanks
Dave

Tides & Currents working perfectly with OpenCPN 5.7.1.b2 for MacOS

[Hakan]

Originally Posted by verkerkbr

Hi Hakan,

I have given an instruction how to generate the can0 at the boot of a Linux system..

Bram..
That instruction I've missed. What commands do you add to the rc.local file?

Now I run these after a reboot. Is it as simple as adding these rows to rc.local?

modprobe can_dev
modprobe can
modprobe can_raw
ip link set can0 type can bitrate 250000ip link set up can0

Thanks
Håkan

[Rae Simpson]

Regarding the high CPU at shutdown and when accessing tide functions.....

Hardware is RPi 4 with PiCAN-M from Copper Hill. The 64 bit version...I think that is called bullseye with OpenCPN 5.6.2 as distributed with Openplotter.

I'm unable to install the Beta....but hoping to understand and sort this high CPU issue.

Back to 32 bit...??

[sanner]

Originally Posted by Hakan

Bram..
That instruction I've missed. What commands do you add to the rc.local file?

Now I run these after a reboot. Is it as simple as adding these rows to rc.local?

modprobe can_dev
modprobe can
modprobe can_raw
ip link set can0 type can bitrate 250000ip link set up can0

Thanks
Håkan

rc.local can work well but udev is another option which should also update when the device is plugged in after booting.

For my can device I created a file called /etc/udev/rules.d/90-canable.rules with the following line:

SUBSYSTEMS=="usb", ATTRS{idVendor}=="1d50", ATTRS{idProduct}=="606f", ACTION == "add", RUN+="/sbin/ip link set can0 down", RUN+="/sbin/ip link set can0 up type can bitrate 250000 loopback off listen-only off"

The two ATTRS values were gleaned from output of dmesg. You may see something similar depending on your device.

[ 2.572703] usb 1-1.3: New USB device found, idVendor=1d50, idProduct=606f, bcdDevice= 0.00
[ 2.572725] usb 1-1.3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 2.572742] usb 1-1.3: Product: canable gs_usb

Seems to work well, brings a can0 at boot / when device is plugged into usb before/after boot time.

[verkerkbr]

Originally Posted by sanner

rc.local can work well but udev is another option which should also update when the device is plugged in after booting.

For my can device I created a file called /etc/udev/rules.d/90-canable.rules with the following line:

SUBSYSTEMS=="usb", ATTRS{idVendor}=="1d50", ATTRS{idProduct}=="606f", ACTION == "add", RUN+="/sbin/ip link set can0 down", RUN+="/sbin/ip link set can0 up type can bitrate 250000 loopback off listen-only off"

The two ATTRS values were gleaned from output of dmesg. You may see something similar depending on your device.

[ 2.572703] usb 1-1.3: New USB device found, idVendor=1d50, idProduct=606f, bcdDevice= 0.00
[ 2.572725] usb 1-1.3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 2.572742] usb 1-1.3: Product: canable gs_usb

Seems to work well, brings a can0 at boot / when device is plugged into usb before/after boot time.

Just tested the udev setup and can confirm that this also works. Not only at boot time but als by simply connecting the usb to the computer system or the RPI4.

Now to see if this also works on the flatpak versions.


Bram

[Hakan]

sanner.. Bram..
Thanks for the tips and for testing.
I'll make a try as well.

[Hakan]

sanner.. bram..
No it didn't worked on my system: arm64 Bullseye.
- The rules for the SG-lock dongle, 65-..., in the same folder was disturbed so no o-charts.

- can0 wasn't detected.

The suggested command was about "usb". My can0 has to what I know nothing to do with any USB?

Deleting the 90-... file, reboot and apply my "normal" modprob commands works as usual though.

[verkerkbr]

Originally Posted by Hakan

sanner.. bram..
No it didn't worked on my system: arm64 Bullseye.
- The rules for the SG-lock dongle, 65-..., in the same folder was disturbed so no o-charts.

- can0 wasn't detected.

The suggested command was about "usb". My can0 has to what I know nothing to do with any USB?

Deleting the 90-... file, reboot and apply my "normal" modprob commands works as usual though.

Hi Hakan,

tested the setup on my Linux Mint Vera AMD64 in combination wih the Oesenc plugin: Works

tested on my RPi4 with Raspian Bullseye 64 version: Works

Connected it to the Actisense NGT-1. Does not work no Can0 available. It works perfect in the SignalK server.

No modprobe action here.

Bram

I'am using the Candlelight firmware. on the Canable versions.

[Be Free]

Windows 10 pro
Route and Mark Manager crashes the program intermittently while stepping through a track. It usually happens when I move from the first to the second entry on the track.



Bill

[verkerkbr]

Originally Posted by verkerkbr

Hi Hakan,

tested the setup on my Linux Mint Vera AMD64 in combination wih the Oesenc plugin: Works

tested on my RPi4 with Raspian Bullseye 64 version: Works

Connected it to the Actisense NGT-1. Does not work no Can0 available. It works perfect in the SignalK server.

No modprobe action here.

Bram

I'am using the Candlelight firmware. on the Canable versions.

Ubuntu Kinetic for the RPI4 also work.


How to get the Can0 in the flatpak version working.


Bram

[sanner]

Originally Posted by Hakan

sanner.. bram..
No it didn't worked on my system: arm64 Bullseye.
- The rules for the SG-lock dongle, 65-..., in the same folder was disturbed so no o-charts.

- can0 wasn't detected.

The suggested command was about "usb". My can0 has to what I know nothing to do with any USB?

Deleting the 90-... file, reboot and apply my "normal" modprob commands works as usual though.

I'm using the Candlelight firmware as well. If your device is using something else you'll have to adjust the ATTRS values in the udev conf file to get it to trigger.

Do you see your device in the output from dmesg ? (is the kernel module being loaded? check lsmod ?)