-
Notifications
You must be signed in to change notification settings - Fork 7
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Module crashing on 5.10.11-v7+ Kernel #12
Comments
I get the same error. Did you find a workaround or downgraded the kernel? |
Same error here, RPI 3 B+, kernel 5.10.17-v7+. Had to downgrade back to 5.4.79-v7+ (as strangely the latest 5.4 kernel 5.4.83 is not in rpi-update). If someone has any solution please let me know. |
my RPI died on downgrading :( . Now I have to reinstall everything. |
Same error here on 5.10.17+ (RPi 1 B), and this is the error you can see in dmesg upon boot:
|
ok i'm not a linux pro but it seems that the "set_fs" and "get_fs" function had been removed in Kernel 5.10 for security reasons. I made a very dirty hack which compiles. I simply replaced the function with: |
Does it work for you after 'hardcoding' the return value? The module compiles and installs, but I can't get ebusd to work: 2021-03-24 12:00:00.005 [mqtt debug] received topic ebusd/bai/WaterPressure/get with data 1 But I get nothing in return. Like the communication with the adapter/heating system does not work. |
unfortunatly I couldn't test my Workaround yet because I don´t have access to an Ebus device at the moment. What ist the output of?
does it acquire the signal and find the devices? |
Looks like the answer is: no |
seems so. Issue #8 mentioned a changed Interrupt Adress. I only have remote access to a raspi which is connected to an ebus. I didn´t tried to run my previos fix of the ttyebus on it because I feared of crashing it with no way to fix it from remote. At the moment I use the normal ttyAMA0 device which works partially. But I ran the
It looks like the IRQ is now 114. Maybe It works when you change Line 196 from 81 to 114. Could you change this line, reinstall and try again? |
Have to try later. Also only have remote access now :) |
Tried running with both changes, i.e. raspi_model 3 and irq 114. Raspi crashed and was not accessible. Had to manually delete the ttyebus module from lib/moduls using my laptop. |
Mine also changed (RPi 2):
|
Has the raspi crashed right after installation or after connecting with the ebus(d)? I could install the module with both changes and it didn't crash after a reboot but I couldn't try it with ebus yet. |
As far as I remember the raspi crashed immediately after rmmoding the old module and modprobing the new one. That is, with irq changed to 114. Before that, i.e. with only the raspi_model hack, the module would load but ebusd did not provide data (from my ventilation Wolf CWL Excellence 300). To be sure that I did not make a mistake in the code: could you provide the hacked .c file? I would be willing to make another test. |
ok let's try that. I forked the repo and added my changes to a new branch: https://github.com/unamehere/ttyebus/tree/kernel_5_10_hack |
I compiled with just the model hack. Result: [ 75.852366] ttyebus: Found RASPI model 3 No connection: ebusctl Therefore I will not even try with irq set to 114, there must be something else that keeps the module from working with kernel 5.10. |
This was to be expected. Without the right IRQ, the kernel module is not able to communicate with the adapter. Sadly I can't try it at the moment because I'm not at home . I could do it remote but I have to be careful because the Pi is also my DNS (PiHole). But I will try to run ebusd with both changes compiled into the module this evening. |
I would try a different irq but how to know which one? As I have physical access to my raspi I would take a little more risk. |
As promised I tried changed IRQ and harcoded return value. Dirty hack but working fine :)
|
@mdk2412
@unamehere |
No problem it was fun. But I wouldn't close this issue until a good replacement for the model identification is found. Right now it's dirty. Also the IRQ adresses of the other RPI models could have changed. This issue should be open until these things are resolved. |
I agree. |
I checked using the steps mentioned by cyablo. Indeed the IRQ has changed to 114 (Raspi 3B, kernel 5.10.17-v7+). I can confirm that with the two hacks (return raspi_model and IRQ set to 114) ebus works on this kernel. Why this setup did crash on me before I don't know. Good work! |
When I try to compile and install the module for 5.10.11-v7+ Kernel on my RPi2, I get a crash stacktrace:
The text was updated successfully, but these errors were encountered: