[Release] FreshJR Adaptive QOS (Improvements / Custom Rules / and Inner workings)

[Jack Yaz]

I can.

But there are some users on older firmwares before the services trigger was introduced. I didn’t want to break compatibility.

You could use firmware version detection code to selectively enable the feature. Yes it would create a little support overheard, but would it benefit the majority of the user base?

 

[FreshJR]

Could be an additional option to improve robustness for people using up-to-date firmware. Check the changelog to see which version added that script, and look at the buildno nvram to determine whether or not tu use it.

Realistically, even listening for "restart_qos" still wouldn't reliably catch each QoS structure reset.

Ever since like 384.6, Qos likes to restart a few times after the "restart_qos" event is initially triggered.

For reliable QoS reset detection, fakeTC would have to send a trigger each time it intercepted "qdisc del dev br0 root"

(The QoS structure used to only be created one time. Then they introduced the stupid "tc ... show | wc -l" check and the system goes crazy for a little bit)

--

But yes, it would decrease gaps in coverage if listening for the service trigger.

 

[RMerlin]

Ever since like 384.6, Qos likes to restart a few times after the "restart_qos" event is initially triggered.

Might be worth configuring a temporary services-event script that will log all called events. Maybe a pattern, or a specific event could be spotted to help handle these restarts.

But there's a chance that we are fighting against direct libbwdpi calls from within the code, for instance from the rules_agent...

For reliable QoS reset detection, fakeTC would have to send a trigger each time it intercepted "qdisc del dev br0 root"

I try to keep faketc as streamlined as possible for performance reasons (because of their constant "tc show" polling... )

 

[Sinner]

Realistically, even listening for "restart_qos" still wouldn't reliably catch each QoS structure reset.

Ever since like 384.6, Qos likes to restart a few times after the "restart_qos" event is initially triggered.

For reliable QoS reset detection, fakeTC would have to send a trigger each time it intercepted "qdisc del dev br0 root"

(The QoS structure used to only be created one time. Then they introduced the stupid "tc ... show | wc -l" check and the system goes crazy for a little bit)

--

But yes, it would decrease gaps in coverage if listening for the service trigger.

I think this was the very reason i didnt move past 384.5. Something didnt seem right with all the reports i was reading. On 384.5 your current script appears to work perfectly. I had the old script set as low as 23secs without issue but settled on 30sec for this version to give some headway and a no worries setup.

 

[FreshJR]

I think this was the very reason i didnt move past 384.5. Something didnt seem right with all the reports i was reading. On 384.5 your current script appears to work perfectly. I had the old script set as low as 23secs without issue but settled on 30sec for this version to give some headway and a no worries setup.

It's fine after it settles down. I wouldn't worry about it or have it keep you from upgrading.

 

[Tekneek]

I think this was the very reason i didnt move past 384.5. Something didnt seem right with all the reports i was reading. On 384.5 your current script appears to work perfectly. I had the old script set as low as 23secs without issue but settled on 30sec for this version to give some headway and a no worries setup.

It's fine after it settles down. I wouldn't worry about it or have it keep you from upgrading.

Once you understand how to put it back on track, in the rare circumstance it goes off the rails (we did enough debugging the other night to establish there's no showstopper happening), it's definitely nothing that should keep you from embracing the enhancements that have come along since 384.5 rolled out.

 

[Sinner]

It's fine after it settles down. I wouldn't worry about it or have it keep you from upgrading.

Well in all honestly I purchased the asus for the advanced features expecting them to work. Asus doesn't pay me to foreveertest theyre software so when I found one that was just solid and worked for everything I needed, that's good enough for me. I wouldn't have even upgd your script except your new gui was the dream and too good to pass up.. and works very well

I do understand security updates are important but my days as a secret agent are over ya know.. Someone wants to hack my google history theyre welcome to it =p

 

[L&LD]

Well in all honestly I purchased the asus for the advanced features expecting them to work. Asus doesn't pay me to foreveertest theyre software so when I found one that was just solid and worked for everything I needed, that's good enough for me. I wouldn't have even upgd your script except your new gui was the dream and too good to pass up.. and works very well

I do understand security updates are important but my days as a secret agent are over ya know.. Someone wants to hack my google history theyre welcome to it =p

What advanced features did you expect to work? When did you buy the router with these expectations? Asus doesn't pay any of us here either, but we find ways to make the hardware we purchased and supplied firmware work as well as it can. Yes, this is a moving target. But so is anyone else's products in the consumer space. And the competitors fall far short of Asus' efforts, IMO.

Security updates are not just to protect you, per se. But to also protect the equipment too (including all the devices on your network) from being put to use in a larger attack as part of a botnet network that does more harm than spying or collecting your google history.

Your priorities should always be security first. Browsing history hijacking is 1990's. I'm sure there are more important digital assets that would make you a valuable target. Even if those assets are just your devices put to more nefarious purposes.

 

[Sinner]

What advanced features did you expect to work? When did you buy the router with these expectations? Asus doesn't pay any of us here either, but we find ways to make the hardware we purchased and supplied firmware work as well as it can. Yes, this is a moving target. But so is anyone else's products in the consumer space. And the competitors fall far short of Asus' efforts, IMO.

Security updates are not just to protect you, per se. But to also protect the equipment too (including all the devices on your network) from being put to use in a larger attack as part of a botnet network that does more harm than spying or collecting your google history.

Your priorities should always be security first. Browsing history hijacking is 1990's. I'm sure there are more important digital assets that would make you a valuable target. Even if those assets are just your devices put to more nefarious purposes.

Well keep up the great work bro! Ill be using mine for things other than searching this forum for fixes 24/7. Im happy with it as is

 

[L&LD]

Well keep up the great work bro! Ill be using mine for things other than searching this forum for fixes 24/7. Im happy with it as is

Nice way to side skirt the questions asked.

Good luck with your head in the sand philosophy.

 

[FreshJR]

This is very off topic. Take it up via PM. You guys are both corect.

- Forced updates are terrible, think iOS.
- I too expect something to work as advertised after I purchase it.
- I personally keep up-to-date, but only out of concern for security.
- If it ain't broken don't fix it, so just leave this decision up to the user.

--

Now keep everything else QoS related.

 

[skeal]

Hey @FreshJR I installed the new @Jack Yaz spdmerlin and the script runs while interfering with Streaming. When the test runs the download is not categorized at all and the upload is in the file transferring category. I played with the script download and upload priority and this didn't help. I raised the video streaming category up above web surfing. This seems to help. Can you shed some light on this or point me in the right direction please?

 

[Jack Yaz]

Hey @FreshJR I installed the new @Jack Yaz spdmerlin and the script runs while interfering with Streaming. When the test runs the download is not categorized at all and the upload is in the file transferring category. I played with the script download and upload priority and this didn't help. I raised the video streaming category up above web surfing. This seems to help. Can you shed some light on this or point me in the right direction please?

Having thought about this, it could be because the traffic is being terminated on the router, rather than forwarded to another device?

 

[skeal]

Having thought about this, it could be because the traffic is being terminated on the router, rather than forwarded to another device?

That is what I was thinking. It may be something nothing can be done about.

 

[Jack Yaz]

That is what I was thinking. It may be something nothing can be done about.

I suspect you're right - if it's not even being classified its probably not passing through tc?

 

[smooth]

Hi all,
Just thought I would pop in and share my appreciation.

I purchased the Asus 86U on Tuesday, already installed @RMerlin Firmware plus @FreshJR adaptive QoS.
Thank you both, great work.

 

[skeal]

Hi all,
Just thought I would pop in and share my appreciation.

I purchased the Asus 86U on Tuesday, already installed @RMerlin Firmware plus @FreshJR adaptive QoS.
Thank you both, great work.

Welcome aboard!!

 

[FreshJR]

Hey @FreshJR I installed the new @Jack Yaz spdmerlin and the script runs while interfering with Streaming. When the test runs the download is not categorized at all and the upload is in the file transferring category. I played with the script download and upload priority and this didn't help. I raised the video streaming category up above web surfing. This seems to help. Can you shed some light on this or point me in the right direction please?

Probably same issue as hosting a VPN server.

The INCOMMING speedtest traffic is routed towards the router itself and terminates after the INPUT chain. This makes the INCOMMING download traffic never hit traffic control/qos as QOS is performed on egress right after POSTROUTING.

See this link for traffic flow. https://i.pinimg.com/originals/e3/7c/f7/e37cf7795695af5ab9f8f4387d557bb7.png

Probably no fix without hacking up the inner structure with IFBs. Alternativly, not sure if there’s a simple way to merge QOS ingress and egress traffic evaluation. If there is (or that traffic already it is merged/muxed within TC) then all we’d need is a simple iptable rule.

Id be nice to figure how to do either of the above to bring QoS to local traffic and virtual interfaces. I dedicated a chuck of time researching the above some time back but never mastered the topic. I do understand the issue clearly.

 

[L&LD]

FreshJR, I am getting an error on a fully reset RT-AC3100 on the latest RMerlin firmware 384.10_2.

The script seems to install okay, except for the menu in the GUI.

admin@RT-AC3100:/tmp/home/root# curl "https://raw.githubusercontent.com/FreshPr/FreshJR_QOS/master/FreshJR_QOS.s
h" -o /jffs/scripts/FreshJR_QOS --create-dirs && sh /jffs/scripts/FreshJR_QOS -install
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  102k  100  102k    0     0   193k      0 --:--:-- --:--:-- --:--:--  197k
Firewall-start not detected, creating firewall-start
Placing FreshJR_QOS entry into firewall-start
mount: mounting /jffs/scripts/www_FreshJR_QoS_Stats.asp on /www/QoS_Stats.asp failed: No such file or directory

Any thoughts on this? User error? This is the command from your GitHub page.

Edit: I used the command from post 2 in this thread and no errors! Maybe the GitHub page needs to be updated?

 

[FreshJR]

Yes github command needs to be updated. But I’ll actually upgrade the script so original command works.

Was reported last page.