[Release] Asuswrt-Merlin 384.16 (and 384.13_6) are now available

[GaselK]

Asus RT-AX56U 384.16

It seems adaptive QoS is Hardware acceleration compatible, at least for 56U

 

[L&LD]

@GaselK, is this after a reboot?

 

[Deleted member 22229]

@GaselK, is this after a reboot?

It's after the defibulation but before the reboot.

 

[Torson]

@GaselK - therein lies the rub - we're comparing different entities (or adaptive QoS settings in this case.)

Mine are Manual bandwidth settings, fq_codel and customized mode. That being said, my test results were quite similar for automatic vs manual bandwidth setting. I chose the latter because it offers a subjective feel-good state of mind knowing that VoIP is at the top of the heap followed by streaming etc.

 

[GaselK]

@GaselK, is this after a reboot?

384.16 flashed. Hardware defaults restored. Settings set manually.
Of course it is after reboot, stays on from day 1. In beta there was some eratic behaviour of hw acc but with GPL of ASUS RT-AX56U for firmware 3.0.0.4.384.8253 can't be sure that everything is fine (more testing is needed) but they DO seem fine.
@Torson my line is relatively limited at 50mbs so adaptive QoS seems more "right". Nevertheless my Voip, Webcam comunication, streaming and online gaming working pretty good atm.

 

[Kingp1n]

@GaselK - therein lies the rub - we're comparing different entities (or adaptive QoS settings in this case.)

Mine are Manual bandwidth settings, fq_codel and customized mode. That being said, my test results were quite similar for automatic vs manual bandwidth setting. I chose the latter because it offers a subjective feel-good state of mind knowing that VoIP is at the top of the heap followed by streaming etc.

@Torson, on you screenshot, I believe you download/uploads numbers are mixed. Is your upload really 150?

 

[maxbraketorque]

...

https://www.asuswrt-merlin.net/about

I think that the final sentence on that page describing the ASUS-Merlin firmware, i.e., "New feature addition is very low on the list of priorities for this project.", may not apply so much any more. Ok with me though.

 

[RMerlin]

"New feature addition is very low on the list of priorities for this project.", may not apply so much any more.

Quite the opposite, it's even more valid today than 5 years ago. Back in the day I added a lot of major new features, like the whole OpenVPN support. Now, I haven't added any new big features since DoT, and that was last year.

 

[nev39]

Good evening,
I see very often following strings in System log - General Log tab which repeat very often again and again.

Apr 16 19:57:43 syslog: WLCEVENTD wlceventd_proc_event(401): eth1: Disassoc 80:19:34:41:38:27, status: 0, reason: Disassociated because sending station is leaving (or has left) BSS (8)
Apr 16 19:57:43 syslog: WLCEVENTD wlceventd_proc_event(420): eth1: Auth 80:19:34:41:38:27, status: 0, reason: d11 RC reserved (0)
Apr 16 19:57:43 syslog: WLCEVENTD wlceventd_proc_event(449): eth1: Assoc 80:19:34:41:38:27, status: 0, reason: d11 RC reserved (0)
Apr 16 19:57:43 dnsmasq-dhcp[357]: DHCPREQUEST(br0) 192.168.1.8 80:19:34:41:38:27
Apr 16 19:57:43 dnsmasq-dhcp[357]: DHCPACK(br0) 192.168.1.8 80:19:34:41:38:27 Andrei-PC

What can cause this?

 

[Deleted member 22229]

What can cause this?

Asus left debugging log enabled been talked about for awhile now. Nothing to sweat over ignore it.

 

[nev39]

Asus left debugging log enabled been talked about for awhile now. Nothing to sweat over ignore it.

Thanks!

Ok, one more question. Taken from same logs

Apr 16 21:51:21 rc_service: cfg_server 3946:notify_rc email_info
Apr 16 21:51:21 custom_script: Running /jffs/scripts/service-event (args: email_info)

What is this and how can i disable this script? I have latest Merlin and Skynet installed.

 

[gotdade]

384.13_6 running well on RT-AC3200 with AiProtection OFF. Enabling AIProtection causes Wifi connection drops.

With AiProtection OFF, the wifi connections seems to be stable even with Tri-Band Smart Connect.

Thanks Merlin and I hope Asus continues to update the RT-AC3200 firmware a few more years.

 

[maxbraketorque]

Quite the opposite, it's even more valid today than 5 years ago. Back in the day I added a lot of major new features, like the whole OpenVPN support. Now, I haven't added any new big features since DoT, and that was last year.

I'd consider incorporation of AMTM to be fairly major. I know its easy to add, but it takes away a barrier that many probably feel is difficult to surmount. Anyhow, everything that you (and everyone else) are doing for these routers is great.

 

[dave14305]

Thanks!

Ok, one more question. Taken from same logs

What is this and how can i disable this script? I have latest Merlin and Skynet installed.

The service-event script is needed for add-ons such as Skynet to pass commands from the GUI to the underlying scripts. If the service-event script isn’t written to “handle” an event like email_info, it will just end with no harm done. All service events are passed to this script if it exists, but not all are actioned within the script.

 

[RMerlin]

I'd consider incorporation of AMTM to be fairly major.

It's not a new feature, it's just that an existing script was added to the firmware image. There hasn't been a single line of code change by me for that. Code changes are the main problem with adding new features - the further away from Asus's code I get, the harder it is for me to keep my code in sync with theirs. Therefore, I try to do a minimum of code changes to anything that is linked to their own code.

 

[MDM]

It's not a new feature, it's just that an existing script was added to the firmware image. There hasn't been a single line of code change by me for that. Code changes are the main problem with adding new features - the further away from Asus's code I get, the harder it is for me to keep my code in sync with theirs. Therefore, I try to do a minimum of code changes to anything that is linked to their own code.

And it's better like that! You don't add bloatware, just nice and useful tings

 

[axaro1]

His ISP may be like many were there modem/Gateway is required for authentication there is no true bridge mode. ATT is like that all they offer is IP Pass or DMZ mode. It's like a semi bridge mode but not a true bridge mode.

I called my ISP, basically all I can do is enable DMZ.
I enabled it on my Modem and I used the AX56U's WAN IP.
I tried portchecker to check if there was any port issue, apparently every single port appears to be closed but everything is working the same as with DMZ disabled, I honestly didn't notice any difference.
@Torson There wasn't any difference in bufferbloat from enabling DMZ.

 

[axaro1]

Asus RT-AX56U 384.16
View attachment 22734 View attachment 22735
It seems adaptive QoS is Hardware acceleration compatible, at least for 56U

My AX56U automatically disable Runner with Qos Enabled.
-QOS Enabled Runner: Disabled (QoS) - Flow Cache: Enabled
-QOS Disabled Runner: Enabled - Flow Cache: Enabled

Edit: I just tried Traditional QoS which appears to be able to limit download efficiently (2-8ms download bufferbloat and staying under the limit), upload bufferbloat still can't be limited (100ms average with 400ms spikes).
Adaptive QoS is not limiting download neither upload, I'll probably just stick with Traditional QoS.

Edit 2: I ran a dslreport test with sfq and traditional QoS and it scored A A A+, I ran the same the same test after 2 minutes and it scored B C A+.
I can't understand why traditional qos is so inconsistent, It cuts my download speed in half every 2 minutes

 

[Blokie]

My AX56U automatically disable Runner with Qos Enabled.
-QOS Enabled Runner: Disabled (QoS) - Flow Cache: Enabled
-QOS Disabled Runner: Enabled - Flow Cache: Enabled

Edit: I just tried Traditional QoS which appears to be able to limit download efficiently (2-8ms download bufferbloat and staying under the limit), upload bufferbloat still can't be limited (100ms average with 400ms spikes).
Adaptive QoS is not limiting download neither upload, I'll probably just stick with Traditional QoS.

Edit 2: I ran a dslreport test with sfq and traditional QoS and it scored A A A+, I ran the same the same test after 2 minutes and it scored B C A+.
I can't understand why traditional qos is so inconsistent, It cuts my download speed in half every 2 minutes

I came to post something similar to this with my new RT-AX58U (384.16).

I've not tried Traditional QoS yet, but adaptive seems to do nothing at all, and I get full download and upload speed no matter what the limits I have entered. I even dropped to 10%, and it still ran at full speed, and same behaviour for sfq and fq_codel.

I also see the same behaviour with Runner being disabled when QoS is enabled.

 

[Blokie]

I came to post something similar to this with my new RT-AX58U (384.16).

I've not tried Traditional QoS yet, but adaptive seems to do nothing at all, and I get full download and upload speed no matter what the limits I have entered. I even dropped to 10%, and it still ran at full speed, and same behaviour for sfq and fq_codel.

I also see the same behaviour with Runner being disabled when QoS is enabled.

I've just been having a play with Traditional QoS, results as follows:

~10% hit on max speeds
Bufferbloat is practically eliminated
Speed limiting still doesn't seem to be working as dropping to 10% max still runs at max speed
Runner and Flow Cache both still reporting Disabled (QoS)