What's new
By:

  • 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!

Adaptive QoS and HW acceleration

Morpheus2020

Morpheus2020

Regular Contributor
As @RMerlin said, Adaptive QoS of all ASUS WiFi 7 routers is Broken.
www.snbforums.com

We have liftoff

Are those WiFi 7 models still broken? Will the QoS Classification page function again on 3006.102.4? You'd have to ask Asus and/or TrendMicro. The problem isn't related to the QoS Classification page. It's the TrendMicro bwdpi engine itself flat out failing to do proper classification on that...
www.snbforums.com www.snbforums.com

In a previous firmware of RT-AX86U, the reason why Adaptive QoS was broken was incompatibility with Broadcom's hardware acceleration. The solution is to manually turn off hardware acceleration in SSH.

Code: 
fc disable
fc flush
runner disable
nvram set runner_disable="1"
nvram set runner_disable_force="1"
nvram set fc_disable="1"
nvram set fc_disable_force="1"
nvram commit

Turn on hardware acceleration or Restore to default:
Code: 
nvram set runner_disable="0"
nvram unset runner_disable_force
nvram set fc_disable="0"
nvram unset fc_disable_force
nvram commit
runner enable
fc enable

Users of ASUS BE routers can try turning off hardware acceleration and then test whether Adaptive QoS is available. I'm not sure if the reason why Adaptive QoS of ASUS BE routers is broken is related to hardware acceleration. I'm also not sure if the above method is useful.
If someone is willing to test and tell me the results of whether this method is useful, I would be very grateful.
.
 
Morpheus2020 said:
As @RMerlin said, Adaptive QoS of all ASUS WiFi 7 routers is Broken.
www.snbforums.com

We have liftoff

Are those WiFi 7 models still broken? Will the QoS Classification page function again on 3006.102.4? You'd have to ask Asus and/or TrendMicro. The problem isn't related to the QoS Classification page. It's the TrendMicro bwdpi engine itself flat out failing to do proper classification on that...
www.snbforums.com www.snbforums.com

In a previous firmware of RT-AX86U, the reason why Adaptive QoS was broken was incompatibility with Broadcom's hardware acceleration. The solution is to manually turn off hardware acceleration in SSH.

Code: 
fc disable
fc flush
runner disable
nvram set runner_disable="1"
nvram set runner_disable_force="1"
nvram set fc_disable="1"
nvram set fc_disable_force="1"
nvram commit

Turn on hardware acceleration or Restore to default:
Code: 
nvram set runner_disable="0"
nvram unset runner_disable_force
nvram set fc_disable="0"
nvram unset fc_disable_force
nvram commit
runner enable
fc enable

Users of ASUS BE routers can try turning off hardware acceleration and then test whether Adaptive QoS is available. I'm not sure if the reason why Adaptive QoS of ASUS BE routers is broken is related to hardware acceleration. I'm also not sure if the above method is useful.
If someone is willing to test and tell me the results of whether this method is useful, I would be very grateful.
.
Click to expand...
What am I looking for? After doing the above and agreeing to the agreement that popped up I got a warning that it would take several minutes for QoS to be active. Everything "appears" to be working
From the log:
Code: 
Mar 18 16:35:46 ripshod SIG_UPDATE: current sig : 2066
Mar 18 16:35:46 ripshod SIG_UPDATE: latest sig : 2448
Mar 18 16:35:46 ripshod SIG_UPDATE: < sig_ver, Do upgrade
Mar 18 16:35:46 ripshod SIG_UPGRADE: wget trf FULL : Official
Mar 18 16:35:46 ripshod SIG_UPGRADE: wget sig, exit code: 0
Mar 18 16:35:46 ripshod SIG_UPGRADE: wget rsa, exit code: 0
Mar 18 16:35:46 ripshod SIG_UPGRADE: Download trf/rsa OK
Mar 18 16:35:47 ripshod SIG_UPGRADE: sig check OK
Mar 18 16:35:47 ripshod SIG_UPGRADE: mv /jffs/signature/rule.trf to jffs
Mar 18 16:35:47 ripshod SIG_UPGRADE: Do restart_wrs
Mar 18 16:35:47 ripshod rc_service: rc 5519:notify_rc restart_wrs
To me that could be a normal firmware update check.
 
Okay. Slight decrease in download speed (from 520 to 501Mbps) and no decrease in wireguard speeds. Looks like you have it
 
These commands would help determine if it’s working or not.
Code: 
tc qdisc show dev br0
tc -s class show dev br0 parent 1:
tc -s filter show dev br0
grep -oE " mark=[0-9a-f]+" /proc/bw_cte_dump | sort | uniq -c
 
Any good for you?
Code: 
Gilly@ripshod:/tmp/home/root# tc qdisc show dev br0
qdisc noqueue 0: root refcnt 2
Gilly@ripshod:/tmp/home/root# tc -s class show dev br0 parent 1:
Gilly@ripshod:/tmp/home/root# tc -s filter show dev br0
Gilly@ripshod:/tmp/home/root# grep -oE " mark=[0-9a-f]+" /proc/bw_
cte_dump | sort | uniq -c
     31  mark=0
    170  mark=10002
      2  mark=1400000
    112  mark=149003d
     18  mark=1490076
      1  mark=2400000
     12  mark=289003d
      1  mark=3c00000
      9  mark=400000
   2077  mark=410002
     26  mark=490076
     34  mark=800000
      2  mark=80000000
    144  mark=80010002
     54  mark=80400000
    239  mark=80410002
      1  mark=804d00de
      4  mark=805400b9
      6  mark=80800000
      5  mark=8080007f
      1  mark=80830005
      2  mark=808d000f
     18  mark=808d00be
      2  mark=808d00de
      2  mark=8092003f
     29  mark=809400b9
      2  mark=809400ba
      1  mark=809400c2
      1  mark=809400c3
      1  mark=80980003
      1  mark=80c40032
      1  mark=80c400c1
      2  mark=80cd00ea
    102  mark=80d2003f
     15  mark=80d400b9
      1  mark=81000000
     43  mark=81400000
      5  mark=814d000f
      2  mark=814d0086
     19  mark=814d00de
      2  mark=8153005e
      1  mark=815400ba
     51  mark=815400bd
      1  mark=81800000
      1  mark=81c00000
      1  mark=82000000
      2  mark=8245005f
      1  mark=824d0017
     35  mark=825400b9
      1  mark=825400ba
      5  mark=829400b9
      3  mark=82c00000
      2  mark=82d400b9
      2  mark=83000000
      1  mark=831400b9
      1  mark=83400000
      1  mark=83c00000
      2  mark=8d00de
      2  mark=9400b9
      2  mark=c00000
 
Ripshod said:
That's a shame :(
Let me know if I can do anything else ;)
Click to expand...
@Ripshod The output you provided..was that with H/W acceleration enabled or disabled? I don't think it matters just curious if the output would be different if H/W was enabled (or vice versa). Thanks!
 
Last edited:
Ripshod said:
That's a shame
Click to expand...
On the plus side, it seemed to identify many BitTorrent connections on your network, but had no QoS structures (qdiscs, classes, filters) setup to manage them.
 
Kingp1n said:
@Ripshod The output you provided..was that with H/W acceleration enabled or disabled? I don't think it matters just curious if the output would be different if H/W was enabled (or vice versa). Thanks!
Click to expand...
That was with the hardware acceleration disabled, as per your first block. Just for you I've run the commands again but with H/W acceleration enabled along with Adaptive QoS so you can compare (if there is any difference).
Code: 
Gilly@ripshod:/tmp/home/root# tc qdisc show dev br0
qdisc noqueue 0: root refcnt 2
Gilly@ripshod:/tmp/home/root# tc -s class show dev br0 parent 1:
Gilly@ripshod:/tmp/home/root# tc -s filter show dev br0
Gilly@ripshod:/tmp/home/root# grep -oE " mark=[0-9a-f]+" /proc/bw_
cte_dump | sort | uniq -c
     10  mark=0
      1  mark=1000000
     40  mark=10002
      1  mark=1800000
      6  mark=1c00000
     32  mark=1c9003d
      4  mark=1c90076
      1  mark=2000000
      1  mark=2400000
      1  mark=2800000
      2  mark=2c00000
      1  mark=3000000
      2  mark=306000c
      1  mark=3400000
      1  mark=349003d
      1  mark=3c00000
      4  mark=400000
    268  mark=410002
      1  mark=490076
      6  mark=800000
      9  mark=80010002
      6  mark=80400000
     12  mark=80410002
     16  mark=8092003f
      1  mark=809400b9
      1  mark=80d400b9
      5  mark=815400b9
      3  mark=81c00000
      1  mark=81cd000f
      1  mark=81cd0086
      4  mark=81cd00de
      9  mark=81d400bd
      1  mark=831400b9
     21  mark=c00000
      1  mark=cd00de
 
You must log in or register to reply here.

Similar threads

S
Routers with 1) configurable QoS with hw acceleration at 1Gbps and 2) multiple vlans?
Replies
11
Views
442
Tech9
Tech9
A
Flow cache vs tcpdump and ntopng
Replies
0
Views
450
arjay199
A
A
RT-AC3100 Adaptive QoS Broken?
Replies
0
Views
3K
asusrunshot
A
Corey Vidal
Level 2 hardware acceleration with new firmware (2021)
Replies
18
Views
10K
Corey Vidal
Corey Vidal
N
Connectivity Issues on IOT Devices with ROG Rapture GT-AXE16000 and ASUS ZenWiFi AX (XT8) Mesh
Replies
11
Views
4K
Krzysztofz
K
Similar threads
Thread starter Title Forum Replies Date
B GT-BE98 FW 3.0.0.6.102_34372/Adaptive QoS Broken ASUS BE Routers & Adapters (Wi-Fi 7) 8

Similar threads

Latest threads

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!

Members online

Total: 606 (members: 40, guests: 566)
Back
Top