What's new

Asuswrt-Merlin 378.50 Beta 2 is out

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

QoS Adaptive possible bug?

I have an Asus ac68u with a selective routing script, openvpn-event from Janosek, and I noticed that every time I hit QoS Adaptive Menu the script stop working and all my internet traffic start going thru the VPN.

So basically if QoS Adaptive is turned off and I hit QoS adaptive menu, selective routing stop working and all traffic start going thru VPN, the logs show these messages:
Feb 2 18:04:43 kernel: * Make sure sizeof(struct sw_struct)=160 is consistent
Feb 2 18:04:44 kernel: IDPfw: TrendMicro forward module ver-1.0.26
Feb 2 18:04:44 kernel: IDPfw: Apply module param dev_wan=ppp0
Feb 2 18:04:44 kernel: IDPfw: Apply module param sess_num=30000
Feb 2 18:04:44 kernel: IDPfw: Init chrdev /dev/idpfw with major 191
Feb 2 18:04:44 kernel: IDPfw: IDPfw is ready
Feb 2 18:04:44 kernel: sizeof forward param = 160

If QoS Adaptive turned on and I hit QoS adaptive menu, selective routing working ok.

Another thing that I noticed is if QoS Adaptive is On and I start using heavy internet traffic like torrent, it's seems that is not capable to handle intense traffic, all connection stop for a brief moment and them return to normal, and when this happens I see these messages in log.
Feb 2 17:27:18 kernel: htb: htb qdisc 17: is non-work-conserving?
Feb 2 17:28:14 kernel: htb: htb qdisc 13: is non-work-conserving?
Feb 2 17:31:43 kernel: net_ratelimit: 1 callbacks suppressed

I don't know if these messages have something to do with dropping all connections or if the processor 1 is running about 95% to 100% under heavy traffic that could be the cause.

So QoS Adaptive is supposed to handle heavy internet traffic?

Thanks
 
I have an Asus ac68u with a selective routing script, openvpn-event from Janosek, and I noticed that every time I hit QoS Adaptive Menu the script stop working and all my internet traffic start going thru the VPN.

So basically if QoS Adaptive is turned off and I hit QoS adaptive menu, selective routing stop working and all traffic start going thru VPN, the logs show these messages:
Feb 2 18:04:43 kernel: * Make sure sizeof(struct sw_struct)=160 is consistent
Feb 2 18:04:44 kernel: IDPfw: TrendMicro forward module ver-1.0.26
Feb 2 18:04:44 kernel: IDPfw: Apply module param dev_wan=ppp0
Feb 2 18:04:44 kernel: IDPfw: Apply module param sess_num=30000
Feb 2 18:04:44 kernel: IDPfw: Init chrdev /dev/idpfw with major 191
Feb 2 18:04:44 kernel: IDPfw: IDPfw is ready
Feb 2 18:04:44 kernel: sizeof forward param = 160

If QoS Adaptive turned on and I hit QoS adaptive menu, selective routing working ok.

Another thing that I noticed is if QoS Adaptive is On and I start using heavy internet traffic like torrent, it's seems that is not capable to handle intense traffic, all connection stop for a brief moment and them return to normal, and when this happens I see these messages in log.
Feb 2 17:27:18 kernel: htb: htb qdisc 17: is non-work-conserving?
Feb 2 17:28:14 kernel: htb: htb qdisc 13: is non-work-conserving?
Feb 2 17:31:43 kernel: net_ratelimit: 1 callbacks suppressed

I don't know if these messages have something to do with dropping all connections or if the processor 1 is running about 95% to 100% under heavy traffic that could be the cause.

So QoS Adaptive is supposed to handle heavy internet traffic?

Thanks

I was able to break 550 Mbps of throughput last year when I tested Adaptive QoS in the early days of the AC87U. There will be a performance hit, but it will still give better throughput than if you were using traditionnal QoS (which is incompatible with NAT acceleration).

The DPI engine does some changes to the iptables. When the service starts it flushes at least one of the PREROUTING tables (which is why I had to drop my original NAT loopback implementation). Unfortunately this is closed source, so I cannot change that behaviour.
 
Last edited:
I second the thought. The Traffic Analyzer is very nice. Let it run for a few hours or days and it provides some great, detailed information.

I have been running Beta3_ta since early this morning on an RT-87R. No issues, works very well.

I must say that I am increasingly impressed by both Asus and Merlin. For a relatively low cost, consumer grade product, the new features Merlin adds along with Asus providing things like the TrendMicro DPI for the RT-68U (I also own one of these) and now Traffic Analyzer in incredible. I wish Cisco, Juniper and other network equipment manufactures would take a lesson.

I have been using Asus WiFi routers since the RT-66U (with merlins fw of course). No looking back. This combination is consumer friendly and great value.
 
Just flashed beta 2 on my AC68U (from beta 1). Under tools, CFE version is still 1.0.1.6, though under the firmware section, signature is 1.028 (which I originally thought corresponded to the CFE since it used to be 1.016... but maybe that's separate).

Should I go ahead and flash with the Asus stock, or should I hold off to help with any testing of the CFE update?
 
Thank you!

For the analytic minds out there (and to thank those who take the time to read those beta threads), a little gift for you to play with:

http://www.mediafire.com/download/ccl3d333970yc3a/RT-AC87U_3.0.0.4_378.50_beta3_ta.zip

I'll leave it to you to figure out what's special about this build (it should be fairly visible ;) )

It is working for me. I used John9527's nvram save & restore utility after performing a factory reset and it seems to be working great. BTW, I also have my AC87R over-clocked to 1200MHz.
 
Just flashed beta 2 on my AC68U (from beta 1). Under tools, CFE version is still 1.0.1.6, though under the firmware section, signature is 1.028 (which I originally thought corresponded to the CFE since it used to be 1.016... but maybe that's separate).

Should I go ahead and flash with the Asus stock, or should I hold off to help with any testing of the CFE update?

Try rebooting the router once to see if it can display the new CFE.
 
OpenVPN is single-threaded, so it can only use one single CPU. The fix in Beta 1 simply made OpenVPN server/client #1 run on the second CPU, so the first CPU could be loaded with handling the actual NAT/network traffic.

There has been no changed to it in beta 2. Make sure you do run Server or Client 1 to ensure it gets run on the second core.

Ok, that makes sense. Would it be possible to run sending (including encryption) on one core and receiving (including decryption) on the other core?
 
Hi Merlin, thanks for the response. That was the first thing I tried. Still displaying 1.0.1.6, and tried in a different browser/cleared cache.

Then I have no idea. Someone else confirmed that the upgrade worked properly for him, and the code there is 100% identical to what Asus uses in 3626. And your syslog probably also shows all the cfe patches that were succesfully applied.

UPDATE: I think I found out why the new bootloader wasn't being flashed - the flash tool isn't set to be executable. I'll fix it for the final release, you can try again then.
 
Last edited:
Ok, that makes sense. Would it be possible to run sending (including encryption) on one core and receiving (including decryption) on the other core?

OpenVPN is not multithreaded. You cannot chose what part of OpenVPN runs on what CPU - it ALL runs on the same CPU.
 
Same here.... French, Italian, and 2 more i think.

Pages loaded fine with Italian. I fixed the two unescaped quotes I could find on the French dict.

If there are any more of them then Asus will have to fix them, as I have no intention of spending hours browsing through 10+ dictionary files each containing over 2100 lines, and recompiling the firmware every time I change something to test it.

Asuswrt's dictionaries are in need of a lot of rework, and I'm not the one who's going to do it...
 
Last edited:
UPDATE: I think I found out why the new bootloader wasn't being flashed - the flash tool isn't set to be executable. I'll fix it for the final release, you can try again then.
Aw, bummer. After I read your initial reply, I went ahead and flashed with the most recent Asus firmware. And sure enough, now I'm on 1.0.2.0. Sorry I won't be able to test with the final update, but hopefully some other people can hold out.

Thanks for all that you do, Merlin!
 
There looks to be a bug in the DST section. I cannot configure the router for DST end on March 8, 2015.

I am using a RT-AC66U running the 378.50_beta2

I don't have any problem changing the end date of DST here. What timezone? What actually happens when you try to change it?
 
UPDATE: I think I found out why the new bootloader wasn't being flashed - the flash tool isn't set to be executable. I'll fix it for the final release, you can try again then.

This is what I call "shirt happend" :)
 
I had the same problem that the bootloader stayed on old version of my AC68U.
Flashed a rescent official firmware first from Asus so it went up to .20. (from .16)

Is the traffic monitor also coming to AC68U soon ? :)

Everything works fine (OpenVPN/ ...)
No issues so far.
Only that my notebook can't reach the internet (dns) after update. (beta 1 and beta 2). After notebook reboot everything fine again.
 

Support SNBForums w/ Amazon

If you'd like to support SNBForums, just use this link and buy anything on Amazon. Thanks!

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top