386.5 Alpha1?

[Adam Williamson]

how come no rt-ac86u alpha

 

[octopus]

how come no rt-ac86u alpha

patience.... coming soon near you

 

[bennor]

how come no rt-ac86u alpha

Generally it means Asus hasn't (yet) posted newer build files for that model to RMerlin for him to compile.

 

[Akore]

RT-AC68U dirty flash over 386.4 Seems more stable than 386.4 on the RT-AC68U which is unheard of for an Alpha release!

 

[geobernd]

RT-AC68U dirty flash over 386.4 Seems more stable than 386.4 on the RT-AC68U which is unheard of for an Alpha release!

I started typing something similar yesterday and then promptly forgot about it but here it is:
Alpha 1 seems to be better with my RT-AC68P and FlexQoS under high QoS load scenarios:
I have 120/18 speed and activities with lots of small packets can max out the processor on download when QoS is active - 100% on core 1 with very high sirq - large packets don't do that)

With 386.4 I noticed sluggishness aka latency increase in other activities when that happens.
With 386.5 Alpha 1 everything remains snappy even when Core 1 is 100% busy....
Maybe something sirq related got optimized with the new Asus portion since all the busy comes from sirq activities.

 

[visortgw]

RT-AC68U dirty flash over 386.4 Seems more stable than 386.4 on the RT-AC68U which is unheard of for an Alpha release!

This is @RMerlin work that you are talking about — not unheard of at all!

 

[capncybo]

I started typing something similar yesterday and then promptly forgot about it but here it is:
Alpha 1 seems to be better with my RT-AC68P and FlexQoS under high QoS load scenarios:
I have 120/18 speed and activities with lots of small packets can max out the processor on download when QoS is active - 100% on core 1 with very high sirq - large packets don't do that)

With 386.4 I noticed sluggishness aka latency increase in other activities when that happens.
With 386.5 Alpha 1 everything remains snappy even when Core 1 is 100% busy....
Maybe something sirq related got optimized with the new Asus portion since all the busy comes from sirq activities.

Likely thanks to this...

Assorted TQoS fixes by kjbracey2 · Pull Request #800 · RMerl/asuswrt-merlin.ng

Fix transfer limits - undefined behaviour was stopping minimums from working Allow default rule to work when a transfer limit rule no longer applies Basic IPv6 support Fix imq0 download classificat...

github.com

EDIT: Oh & look what I noticed while I'm in here snooping...
Perhaps we'll see an alpha for the RT-AC86U sometime soon ;-) ???

https://github.com/RMerl/asuswrt-merlin.ng/commit/46f37f47720ac0b8d9f7e32e4fb02c7a626a0f24

BTW Thx RMerlin

 

[dave14305]

Likely thanks to this...

Assorted TQoS fixes by kjbracey2 · Pull Request #800 · RMerl/asuswrt-merlin.ng

Fix transfer limits - undefined behaviour was stopping minimums from working Allow default rule to work when a transfer limit rule no longer applies Basic IPv6 support Fix imq0 download classificat...

github.com

It would not change Adaptive QoS behavior.

 

[capncybo]

It would not change Adaptive QoS behavior.

LOL, well I've been wrong before... Thanks for the correction (as Nobody knows FlexQoS) much better than you.

EDIT: Hopefully my second point is still Valid, or that's two strikes for me

 

[supe]

@RMerlin As always really appreciate your work.
I was hoping you may have some insight into why the rt-ax86 LAN to LAN transfers get affected by enabling QOS besides Adaptive(transfer rates drops from around 100MBs to around 60MBs).
I don't believe its your build because it happens with the stock FW as well, but whats puzzling is if I swap in my spare RT-AC86U, this speed hit does not occur enabling regular QOS.

 

[torstein]

RT-AC68U dirty flash over 386.4 Seems more stable than 386.4 on the RT-AC68U which is unheard of for an Alpha release!

Don’t tempt me to install an alpha!!

 

[kernol]

Don’t tempt me to install an alpha!!

 

[Zastoff]

See a lot of: Since i updated to alpha 1 on AX88u
These only come from clients i have set on DNS-Filter

Feb 16 21:21:30 RT-AX88U-6C58 kernel: ** ERROR: [send_redir_page:625] # redir_url=https://192.168.1.1:8443/blocking.asp?cat_id=23&mac=*******C8428&domain=yield-op-idsync.live.streamtheworld.com
Feb 16 21:21:50 RT-AX88U-6C58 kernel: ** ERROR: [send_redir_page:625] # redir_url=https://192.168.1.1:8443/blocking.asp?cat_id=23&mac=*******B5023&domain=acds.prod.vidible.tv

Sometimes restarting DNS-Filters seems to help for awhile
&

Feb 16 21:06:41 RT-AX88U-6C58 kernel: httpds (1261): drop_caches: 1
Feb 16 21:12:13 RT-AX88U-6C58 kernel: [tdts_shell_ioctl_stat:256] Recv ioctl req with op 2

 

[skeal]

Feb 16 21:06:41 RT-AX88U-6C58 kernel: httpds (1261): drop_caches: 1

When you login to the firmware this entry happens.

Feb 16 21:12:13 RT-AX88U-6C58 kernel: [tdts_shell_ioctl_stat:256] Recv ioctl req with op 2

I have the same error and have no clue what it is.

 

[XIII]

RMerlin on Twitter:

https://twitter.com/x/status/1494104617864634368

 

[underdose]

I have the same error and have no clue what it is.

Looks like Trendmicro stuff. https://www.snbforums.com/threads/kernel-tdts_shell_ioctl_stat-256-recv-ioctl-req-with-op-2.49347/

 

[dmtg]

bug
in Main_WStatus_Content.asp RT-AX88U cannot see WIFI clients connect

 

[skeal]

bug
in Main_WStatus_Content.asp RT-AX88U cannot see WIFI clients connect

I have the same router and don't have that issue.

 

[Jeep]

Dirty flash to 386.5_alpha1 to my main router (RT-AX86U) and AiMesh node (RT-AX88U). Everything seems to be working smoothly, and my AiMesh connection quality as reported on the AiMesh page has gone from Fair to Great. Both were on 386.4 previously and I was getting IoT devices dropping from time to time, I'll keep an eye on things with the updated GPL. I'm ~90 minutes in with no issues yet, knock on wood.

On a side note, thanks for all of the encouragement in this thread to install an alpha version, something I would never do normally but as others have said Merlin is amazing!

 

[Damit1]

After all the warm recommendations here to upgrade to Alpha version which I have promised to myself in order to avoid bugs, I just upgraded to an Alpha version for the first time after I stopped it because of Asus made me a QA tester about 1 year ago.
AX88u - 2 PCs (Router and a Node)
Never thought a dirty upgrade would be so fast and further more- seamless.
I came with fear because of past experience.

Running now for 20m and 21 clients connected like a charm (did 1 reboot right after the upgrade, would do another one now)
Hope that the new WIFI drivers would assist few IOT devices I got disconnecting and re-connecting.
My huge thanks to @RMerlin for his terrific work, definitely worth the donation I did not so much while ago.