Because I got tired of manual patching an old bug in Linux’s AX.25 kernel module (which still isn’t fixed!) I wrote a small script to do the work last year.
Now it’s easy to (re)patch my kernels on several systems without any hassle.
I still need to do some code cleanup and optimize it a bit before I think it’s good enough to publish here.
In the meantime… If you’re interested please feel free to contact me.
Hi,
I would like to receive the script.
Does this problem also cause kernel “oops”es?
The problem with / bug in the AX.25 kernel module doesn’t cause kernel panics.
Not sure if you got my email from the contact page, it ended in some form of WordPress error.
Anyways, just wondering about the status of this patch? Was it a RPi specific patch, or something that could be cross platform in the ax.25 tree. Was this submitted to the kernel team?
Seems the 6.x kernel(s) have the same issue.
Hello Nick, thanks for the info about the contact page. I’ll have a look at it.
Regarding the patch. This was to fix ‘hanging socket’ issues in kernel 4.19, not specifically for the Rpi.
I’m not aware of the current Linux AX.25 stack status but I thought this particular issue was already fixed in kernel 6.x another way around I’ve heard.
This patch is shared with the kernel team / AX.25 maintainers several times by me and some others (like N1URO RIP) some years ago. After many attempts to get it fixed we gave up…
Thanks for the reply Dave. Any chance I can get a copy of said patch, would like to try to apply it on a RPi test system and see if our packet issues continue or go away. We are playing with a console/html based winlink client called Pat. Seems if there is a problem during the winlink protocol session, it ends with an error, and 99.9% of the time you have to either restart ax.25 or reboot the machine. If we point pat at an AGWPE interface on a windows machine, it seems to run flawlessly. Am interested to see if the patch may reduce/remove the issues.
Thanks,
Nick