[Testers Needed] GD-77 Firmware Loader for Linux and OS-X

[KC7RBW]

Well, it took me 8 hours to install mono-complete on a RPi 3 A+

Replace the udev rules file with the attached one, then execute the following (or reboot):

sudo udevadm control --reload-rules

Cheers.

Yikes. 8 hours is a long time. Possibly a faster MicroSD would work better. I'm sure it didn't take me that long.

I tried the rules file you sent, and it didn't work either, but this time I found where the logs went.

Code: Select all

Dec 01 19:30:26 pi systemd-udevd[12613]: Process '/bin/bash -c 'ID=$(echo /devices/platform/soc/3f980000.usb/usb1/1-1/1-1.4/1-1.4:1.0/usbmisc/hiddev0 | cut -d/ -f8); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'' failed with exit code 1.

Looking at what's going on there, I wonder if grabbing the third in from the end would work better than grabbing the 7th or 8th depending on platform. You can do this using cut by using rev before and after. The following rule worked for me. If it works on non-Rasbian installations that would be great. I don't know what $devpath is on other Linuxes, but you may.

Code: Select all

KERNEL=="hiddev*", ATTRS{idVendor}=="15a2", ATTRS{idProduct}=="0073", MODE="0666", GROUP="plugdev", RUN+="/bin/bash -c 'ID=$(echo %p | rev | cut -d/ -f3 | rev); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'"

I also tried to find a better way, but I haven't found it yet.

[VK3KYY]

This is probably a stupid question, but why don't you just use the Python loader I wrote ?
I presume that would work on the RPi

[F1RMB]

Hi,

Well, it took me 8 hours to install mono-complete on a RPi 3 A+

Replace the udev rules file with the attached one, then execute the following (or reboot):

sudo udevadm control --reload-rules

Cheers.

Yikes. 8 hours is a long time. Possibly a faster MicroSD would work better. I'm sure it didn't take me that long.

I tried the rules file you sent, and it didn't work either, but this time I found where the logs went.

Code: Select all

Dec 01 19:30:26 pi systemd-udevd[12613]: Process '/bin/bash -c 'ID=$(echo /devices/platform/soc/3f980000.usb/usb1/1-1/1-1.4/1-1.4:1.0/usbmisc/hiddev0 | cut -d/ -f8); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'' failed with exit code 1.

Looking at what's going on there, I wonder if grabbing the third in from the end would work better than grabbing the 7th or 8th depending on platform. You can do this using cut by using rev before and after. The following rule worked for me. If it works on non-Rasbian installations that would be great. I don't know what $devpath is on other Linuxes, but you may.

Code: Select all

KERNEL=="hiddev*", ATTRS{idVendor}=="15a2", ATTRS{idProduct}=="0073", MODE="0666", GROUP="plugdev", RUN+="/bin/bash -c 'ID=$(echo %p | rev | cut -d/ -f3 | rev); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'"

I also tried to find a better way, but I haven't found it yet.

Okay. I guess you're using a RPi 4, beast I don't own. Yep, reversing should work regardless platforms.

So, are you able to flash your GD now ?

Cheers.
---
Daniel

[VK3KYY]

I have a Rpi 4 but not really any time to test this.

[F1RMB]

Hi,

Well, it took me 8 hours to install mono-complete on a RPi 3 A+

Replace the udev rules file with the attached one, then execute the following (or reboot):

sudo udevadm control --reload-rules

Cheers.

Yikes. 8 hours is a long time. Possibly a faster MicroSD would work better. I'm sure it didn't take me that long.

I tried the rules file you sent, and it didn't work either, but this time I found where the logs went.

Code: Select all

Dec 01 19:30:26 pi systemd-udevd[12613]: Process '/bin/bash -c 'ID=$(echo /devices/platform/soc/3f980000.usb/usb1/1-1/1-1.4/1-1.4:1.0/usbmisc/hiddev0 | cut -d/ -f8); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'' failed with exit code 1.

Looking at what's going on there, I wonder if grabbing the third in from the end would work better than grabbing the 7th or 8th depending on platform. You can do this using cut by using rev before and after. The following rule worked for me. If it works on non-Rasbian installations that would be great. I don't know what $devpath is on other Linuxes, but you may.

Code: Select all

KERNEL=="hiddev*", ATTRS{idVendor}=="15a2", ATTRS{idProduct}=="0073", MODE="0666", GROUP="plugdev", RUN+="/bin/bash -c 'ID=$(echo %p | rev | cut -d/ -f3 | rev); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'"

I also tried to find a better way, but I haven't found it yet.

Okay. I guess you're using a RPi 4, beast I don't own. Yep, reversing should work regardless platforms.

So, are you able to flash your GD now ?

Cheers.
---
Daniel

Try with this line as replacement:

Code: Select all

KERNEL=="hiddev*", ATTRS{idVendor}=="15a2", ATTRS{idProduct}=="0073", MODE="0666", GROUP="plugdev", RUN+="/bin/bash -c 'ID=$(IFS=/; read -a array <<< %p; echo ${array[-3]}); echo $ID > /sys/bus/usb/drivers/usbhid/unbind'"

[F1RMB]

Hi Roger,

This is probably a stupid question, but why don't you just use the Python loader I wrote ?
I presume that would work on the RPi

That was the first thing I tried under Linux. I fixed and extended a bit the python script, it flash ATM, but the GD won't boot after that.
I need to check why (probably mistake from my side).

[F1RMB]

Roger,

I have a Rpi 4 but not really any time to test this.

Yeah, no problem

Since Andrew idea to use the last third field from the end, that should work (hopefully) for all Linux platforms.

Cheers.
---
Daniel

[F1RMB]

Hi Roger,

Hi Roger,

This is probably a stupid question, but why don't you just use the Python loader I wrote ?
I presume that would work on the RPi

That was the first thing I tried under Linux. I fixed and extended a bit the python script, it flash ATM, but the GD won't boot after that.
I need to check why (probably mistake from my side).

Here is a working version of the python loader.

The PR has also been created.

EDIT: updated the attachement with a version that works with python2 and python3


Cheers.
---
Daniel

[KC7RBW]

Python loader worked nicely for me too, and didn't need to unbind the HID driver, though it did (without any udev rules) require root on the box I tested it on. I think that probably this means that we don't actually need to unbind the HID driver, but need to point the .NET loader at a different device node.

[F1RMB]

Python loader worked nicely for me too, and didn't need to unbind the HID driver, though it did (without any udev rules) require root on the box I tested it on. I think that probably this means that we don't actually need to unbind the HID driver, but need to point the .NET loader at a different device node.

Nice to read that

About unbinding, if I can succeed to make the CPS working with Mono, firmware flashing from within the CPS will require usbhid unbind. Codeplug reading/writing, with OpenGD77 will work as is, without unbinding, since it access it using a CDC ACM device. About official firmware, that will also require unbinding.

Cheers.
---
Daniel