IMQ@NAT kernel modul issue

[RomanHK]

Hello @pbix (and/or @Lantis), can you please check it out?
I came to this problem as the first and progressed time is found error for multiple users.
viewtopic.php?f=14&t=11829&start=160#p52149
viewtopic.php?f=14&t=11829&start=180#p53152
viewtopic.php?f=14&t=11829&start=200#p53265

P.S. For Issue #818 from GitHub I divided this problem into a separate thread.

Every help is given. Thank you.

[Lantis]

Can you please change the "sleep 2" in /etc/hotplug.d/iface/21-gargoyle_firewall to "sleep 30" and see if the issue goes away? This should confirm if it is a collision of the modules.

[Lantis]

It would also be helpful to have a full bootlog when the problem occurs and when it doesn't. Preferably on a vanilla system where the imq module has not been re-ordered etc.

[RomanHK]

Thank you for your first ideas, check out after the weekend.

I bought a new router (Linksys WRT3200ACM) and the problem was also there.

Turris Omnia - helped postpone loading by changing priority - but it is not reliable.
Linksys WRT3200ACM - error also occurred when changing priority, also when loading from /etc/modules-boot.d, I also modified the script /etc/init.d/boot and here I put a 2s delay and it didn't help.

The bad thing about this is that the error is not constant and sometimes it manifests itself and sometimes not.

[RomanHK]

So I have the first results of your proposal - unsuccessful. The NAT was not loaded after the third restart.
5x restart = 2x failure.

In my opinion, it won't even happen to /etc/hotplug.d . The error will be somewhere near the boot loading kernel mods.

More ideas please?

[Lantis]

And if you disable IMQ from loading at all?

Just for reference, I’ve never seen this error on any of my devices so I can’t debug it.

[RomanHK]

Yes, I confirm. When I delete /etc/modules.d/46-ipt-imq, everything goes well. Day 3 without file, several reboots and everything works (of course, IMQ-dependent services don't work).

[tester]

Just for reference, I’ve never seen this error on any of my devices so I can’t debug it.

Here you said that got it once: viewtopic.php?f=14&t=11829&start=80#p51660

Can you plz at least re-check yours IMQ kernel/iptables patches comparing with originals?

[Lantis]

I also don’t know if that behaviour is the same, I didn’t look at the time. It could be, but I don’t know.

I’m already looking and have changes to be applied when I have finished testing them. I don’t expect them to fix this issue, but should stop the crash message on first unload of the module. It isn’t as straight forward as comparing the two. The patch is customised specifically for Gargoyle. Comparing apples and elephants.

Just trust that I don’t sit on my hands with these things. I do actively try to fix things, even problems that I can’t reproduce.

Thank you for continue research. I hope that Gargoyle developers join it one day.

Comments like these are quite frankly self entitled and disrespectful.

[tester]

but should stop the crash message on first unload of the module.

It's a good news too.

I do actively try to fix things,

Thank you for all the hard work, I didn't knew that you are watching this bug already.

even problems that I can’t reproduce.

RomanHK have this bug on WRT3200ACM, I've seen that you also have WRT3200ACM and you can't reproduce it on it? Very strange if so.