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

[klidec]

How would i be able to prioritize homekit/homebridge activity on the network?
Would this tweak help me do so? I'm not sure where this activity gets categorized today or how to check.
Any help would be appreciated.

 

[Vexira]

Can anyone say why the script starts twice? Or is this behavior normal, have I perhaps done something wrong?

Sep 24 10:41:04 87u: Adaptive QOS: Modification Script Started
Sep 24 10:41:13 87u: Adaptive QOS: Modification Script Started
Sep 24 10:41:35 87u: Adaptive QOS: Changing container for Unidentified Traffic & Applying Custom Rules
Sep 24 10:41:35 87u: Adaptive QOS: Changing minimum alloted bandwidth per QOS category to user defined percentages
Sep 24 10:41:45 87u: Adaptive QOS: No change required for Unidentified Traffic Container or Custom Rules
Sep 24 10:41:45 87u: Adaptive QOS: Changing minimum alloted bandwidth per QOS category to user defined percentages

thats not normal are you using manual bandwith

 

[Vexira]

How would i be able to prioritize homekit/homebridge activity on the network?
Would this tweak help me do so? I'm not sure where this activity gets categorized today or how to check.
Any help would be appreciated.

you can do it by default in adaptive qos,under bandwirh monitor

 

[klidec]

you can do it by default in adaptive qos,under bandwirh monitor

I'd like to prioritize homekit communications traffic coming to and from any device in my network. How could i use bandwidth monitor to do that? Almost all my devices have the potential to communicate via homekit.

 

[strangeluck]

I'd like to prioritize homekit communications traffic coming to and from any device in my network. How could i use bandwidth monitor to do that? Almost all my devices have the potential to communicate via homekit.

I don't know much about homebridge or homekit, but it seems the default port homebridge uses is 51827, while homekit communicates with icloud on port 5223. Not sure this helps, but for example to make rules putting all traffic on ports 51827 and 5223 into the "others" container use the following:

##DOWNLOAD
${tc} filter add dev br0 protocol all prio 1 u32 match ip dport 51827 0xffff flowid ${Others}
${tc} filter add dev br0 protocol all prio 1 u32 match ip dport 5223 0xffff flowid ${Others}
##UPLOAD
${tc} filter add dev eth0 protocol all prio 1 u32 match ip sport 51827 0xffff flowid ${Others}
${tc} filter add dev eth0 protocol all prio 1 u32 match ip sport 5223 0xffff flowid ${Others}

 

[klidec]

I don't know much about homebridge or homekit, but it seems the default port homebridge uses is 51827, while homekit communicates with icloud on port 5223. Not sure this helps, but for example to make rules putting all traffic on ports 51827 and 5223 into the "others" container use the following:

##DOWNLOAD
${tc} filter add dev br0 protocol all prio 1 u32 match ip dport 51827 0xffff flowid ${Others}
${tc} filter add dev br0 protocol all prio 1 u32 match ip dport 5223 0xffff flowid ${Others}
##UPLOAD
${tc} filter add dev eth0 protocol all prio 1 u32 match ip sport 51827 0xffff flowid ${Others}
${tc} filter add dev eth0 protocol all prio 1 u32 match ip sport 5223 0xffff flowid ${Others}

Thanks! I'll start experimenting.

 

[Vexira]

is there a way to get the, script to automatically calculate 90% of the bandwidth entered.

 

[jpclarke]

To do floating point calculations in bash you will need to use something like bc available on optware.

 

[FreshJR]

To do floating point calculations in bash you will need to use something like bc available on optware.

It doesn't make a difference for your statement but we are using ash not bash.

Vexira just multiply by 90 then divide by 100 using the expr command. It will truncate the floating point. Problem solved.

If you can't figure it out, I'll send u an updated version

 

[eclp]

thats not normal are you using manual bandwith

Is set to manual bandwith.
However, the many syslog entries come only after router restart. Later only three entries.

 

[eddiez]

Got these:
22:57:53 kernel: HTB: quantum of class 10011 is big. Consider r2q change.
Oct 9 22:57:53 kernel: HTB: quantum of class 10012 is big. Consider r2q change.
Oct 9 22:57:53 kernel: HTB: quantum of class 10013 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10011 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10015 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10012 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10016 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10013 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10015 is big. Consider r2q change.
Oct 9 22:57:55 kernel: HTB: quantum of class 10016 is big. Consider r2q change.

Started popping up in 68_4, both in manual and automatic setting.

Did not modify anything. No mentioning in earlier firmware, already updated to 1.92, no difference.

Anyone know what's the issue?

 

[Jack Yaz]

Got these:
22:57:53 kernel: HTB: quantum of class 10011 is big. Consider r2q change.
Oct 9 22:57:53 kernel: HTB: quantum of class 10012 is big. Consider r2q change.
Oct 9 22:57:53 kernel: HTB: quantum of class 10013 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10011 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10015 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10012 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10016 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10013 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10015 is big. Consider r2q change.
Oct 9 22:57:55 kernel: HTB: quantum of class 10016 is big. Consider r2q change.

Started popping up in 68_4, both in manual and automatic setting.

Did not modify anything. No mentioning in earlier firmware, already updated to 1.92, no difference.

Anyone know what's the issue?

No issue. Warning that can be ignored.

 

[Vexira]

Got these:
22:57:53 kernel: HTB: quantum of class 10011 is big. Consider r2q change.
Oct 9 22:57:53 kernel: HTB: quantum of class 10012 is big. Consider r2q change.
Oct 9 22:57:53 kernel: HTB: quantum of class 10013 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10011 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10015 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10012 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10016 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10013 is big. Consider r2q change.
Oct 9 22:57:54 kernel: HTB: quantum of class 10015 is big. Consider r2q change.
Oct 9 22:57:55 kernel: HTB: quantum of class 10016 is big. Consider r2q change.

Started popping up in 68_4, both in manual and automatic setting.

Did not modify anything. No mentioning in earlier firmware, already updated to 1.92, no difference.

Anyone know what's the issue?

Also use manual bandwidth

 

[Vexira]

It doesn't make a difference for your statement but we are using ash not bash.

Vexira just multiply by 90 then divide by 100 using the expr command. It will truncate the floating point. Problem solved.

If you can't figure it out, I'll send u an updated version

ok appreciated its a bit confuseing for me lol, havent been on the pc for a while.

 

[eddiez]

No issue. Warning that can be ignored.

This one as well?

Oct 10 15:51:23 kernel: htb: htb qdisc 14: is non-work-conserving?
Oct 10 15:52:05 kernel: htb: htb qdisc 11: is non-work-conserving?

 

[Vexira]

This one as well?

Oct 10 15:51:23 kernel: htb: htb qdisc 14: is non-work-conserving?
Oct 10 15:52:05 kernel: htb: htb qdisc 11: is non-work-conserving?

Annoying error the it's not a big issue.

 

[eclp]

Can I do it remove the call /jffs/scripts/FreshJR_QOS completely out of firewall-start and move it to services-start? I want to do this to avoid double calls of the script.

Sep 24 10:41:04 87u: Adaptive QOS: Modification Script Started
Sep 24 10:41:13 87u: Adaptive QOS: Modification Script Started
Sep 24 10:41:35 87u: Adaptive QOS: Changing container for Unidentified Traffic & Applying Custom Rules
Sep 24 10:41:35 87u: Adaptive QOS: Changing minimum alloted bandwidth per QOS category to user defined percentages
Sep 24 10:41:45 87u: Adaptive QOS: No change required for Unidentified Traffic Container or Custom Rules
Sep 24 10:41:45 87u: Adaptive QOS: Changing minimum alloted bandwidth per QOS category to user defined percentages

 

[Jack Yaz]

Can I do it remove the call /jffs/scripts/FreshJR_QOS completely out of firewall-start and move it to services-start? I want to do this to avoid double calls of the script.

Sep 24 10:41:04 87u: Adaptive QOS: Modification Script Started
Sep 24 10:41:13 87u: Adaptive QOS: Modification Script Started
Sep 24 10:41:35 87u: Adaptive QOS: Changing container for Unidentified Traffic & Applying Custom Rules
Sep 24 10:41:35 87u: Adaptive QOS: Changing minimum alloted bandwidth per QOS category to user defined percentages
Sep 24 10:41:45 87u: Adaptive QOS: No change required for Unidentified Traffic Container or Custom Rules
Sep 24 10:41:45 87u: Adaptive QOS: Changing minimum alloted bandwidth per QOS category to user defined percentages

Or could I use this in firewall_start:

ps w |  grep -v grep |  grep /jffs/scripts/FreshJR_QOS.sh ||  sh /jffs/scripts/FreshJR_QOS.sh

I borrowed the Check_Lock function from @Adamm skynet project, to prevent duplicate runs

 

[eclp]

@Jack Yaz ...
And where did you insert the function, can you possibly make it available?

 

[Jack Yaz]

@Jack Yaz ...
And where did you insert the function, can you possibly make it available?

https://gist.github.com/jackyaz/c6675a0ec59c0aa328258cb2c894694b
Lines 33-43, then line 227