[Fork] Asuswrt-Merlin 374.43 LTS releases (Archive)

[il2]

Many thanks John for all updates in Update-11E1.
My he.net 6in4 tunnel config has lived through the update and it is still 10/10 on test-ipv6.com as it was before, fonts become anti-aliased and look cool.
I only had to reset/reformat jffs and restore it from the backup.

 

[Builder71]

...
I only had to reset/reformat jffs and restore it from the backup.

Same here.

 

[Builder71]

Thx for the works John.

I found one bug.
If you go to "Tools->Sysinfo" bottom of the page.
"Wireless clients" should always count all wireless clients.
It now doesn't count wireless clients using guest networks.

 

[john9527]

I only had to reset/reformat jffs and restore it from the backup.

The code image grew by 80-100K between V10 and V11 depending on the router model. For the MIPs routers, this was probably enough to 'push' the jffs start location.

 

[john9527]

I found one bug.
If you go to "Tools->Sysinfo" bottom of the page.
"Wireless clients" should always count all wireless clients.
It now doesn't count wireless clients using guest networks.

Looks like you found a latent behavior from the original code (I even loaded a couple of old levels to double check). Problem is that it makes a system call into the closed source driver to get the info, so not sure I can do anything (at least easily) to change it.

 

[ml70]

I run HE 6in4 tunnel without any problems....that's what I use for all my testing since I don't have native IPv6 support from my ISP.

But this turns ctf off and limits the router to 150-200 mbps performance? Even before you run anything else on it...

 

[john9527]

But this turns ctf off and limits the router to 150-200 mbps performance? Even before you run anything else on it...

Not for me.....ctf remains active....
I only have a 50Mbps line, but my understanding is that a ctf conflict would show up regardless of the line speed.
Anyway, the IPv6 stack is not going to change. If you are having problems with it in your environment, the main Merlin build may be a better choice for you.

 

[RMerlin]

Looks like you found a latent behavior from the original code (I even loaded a couple of old levels to double check). Problem is that it makes a system call into the closed source driver to get the info, so not sure I can do anything (at least easily) to change it.

I haven't looked at this yet (I had a similar report a few months ago) but I suspect the fix would require querying the wireless driver for every active sub-interface as well as the main one. The relevant code is in httpd/sysinfo.c:get_wifi_clients().

 

[john9527]

I haven't looked at this yet (I had a similar report a few months ago) but I suspect the fix would require querying the wireless driver for every active sub-interface as well as the main one. The relevant code is in httpd/sysinfo.c:get_wifi_clients().

yes...which in turn calls wl_ioctl...

 

[RMerlin]

yes...which in turn calls wl_ioctl...

What I mean is, wl_ioctl might need to be called with wl0.1, wl0.2, wl0.3, etc... This is the parameter stored in the "name" variable.

 

[ml70]

Not for me.....ctf remains active.....

Now this is a case for some serious wtf: every time i turn on an 6in4 tunnel (he.net) on AC66 & your fork (or any 374.43 fw for that matter) my ctf goes off, as reported by Tools>Sysinfo.

 

[john9527]

Now this is a case for some serious wtf: every time i turn on an 6in4 tunnel (he.net) on AC66 & your fork (or any 374.43 fw for that matter) my ctf goes off, as reported by Tools>Sysinfo.

Maybe the difference in the kernel levels between MIPS and ARM (your AC66 vs my AC68)?

 

[john9527]

I haven't looked at this yet (I had a similar report a few months ago) but I suspect the fix would require querying the wireless driver for every active sub-interface as well as the main one. The relevant code is in httpd/sysinfo.c:get_wifi_clients().

OK.....so it was easy , but cumbersome......

@RMerlin, let me know if you want me to send you a patch file.

 

[Chrysalis]

john thanks for not moving to newer ipv6 implementation as I think radvd is superior.

 

[Builder71]

OK.....so it was easy , but cumbersome......

@RMerlin, let me know if you want me to send you a patch file.

Great you guys found it.
Will there be a -11E2, or wait for -12?

 

[john9527]

Right now for V12....it's been there forever, a little longer shouldn't hurt too bad

 

[RMerlin]

OK.....so it was easy , but cumbersome......

@RMerlin, let me know if you want me to send you a patch file.

I'm on a development break these days,, so go for it if you want to. You will need to determine if there's an nvram that contains the list of currently active virtual subinterface, and then start iterating through them.

One place to look as a starting point would be the httpd/sysdep/web-broadcom-wl6.c to see how Asus does it when building the list of clients for the Wireless Client interface. That should provide a good starting point.

The whole content of sysinfo.c was written by me, so we can change it as much as we want without any fear of running into future GPL merge issues.

 

[john9527]

I'm on a development break these days,, so go for it if you want to. You will need to determine if there's an nvram that contains the list of currently active virtual subinterface, and then start iterating through them.

One place to look as a starting point would be the httpd/sysdep/web-broadcom-wl6.c to see how Asus does it when building the list of clients for the Wireless Client interface. That should provide a good starting point.

The whole content of sysinfo.c was written by me, so we can change it as much as we want without any fear of running into future GPL merge issues.

I already coded and tested.....keep it simple approach. Just modified get_wifi_clients to accept calls for each of the guest network interfaces as well as the primary wireless, and then just check them all from the tools page.

 

[Builder71]

Right now for V12....it's been there forever, a little longer shouldn't hurt too bad

Sounds fair to me.

 

[Builder71]

Looking at the Wireless Log I do see some weird RSSI numbers.
Not sure if I've seen that before.

Crazy high (invalid) numbers like +39dBm.
If I hit "Refresh" it sometimes go back to all negative numbers.
But most of the time the high numbers stay.