[ALPHA] 386.9_alpha1 for WiFi 5 (802.11ac) Models (GT-AC2900, RT-AC68U, and RT-AC86U)

[Howard_inGA]

Wished ASUS pushed their gpl for the ac5300.

Never mind...;-)

So, removed foot from mouth, applied update, works like a charm both with and without Nord VPN.
Only missing piece was the IPv6 function. That never worked in NATIVE, but I don't use it due to Nord being required from time to time...

 

[Tech9]

Hmm... this AC86U alpha on stable 386 base works better than AX86U release on alpha 388 base.

 

[Tech9]

The logs are quiet, but haven't seen this one for a while on a clean system:

Dec 28 18:39:59 kernel: dcd[2728]: unhandled level 3 translation fault (11) at 0x00000000, esr 0x92000007
Dec 28 18:39:59 kernel: pgd = ffffffc012df0000
Dec 28 18:39:59 kernel: [00000000] *pgd=0000000012eb2003, *pud=0000000012eb2003, *pmd=0000000013a89003, *pte=0000000000000000
Dec 28 18:39:59 kernel: CPU: 1 PID: 2728 Comm: dcd Tainted: P           O    4.1.27 #2
Dec 28 18:39:59 kernel: Hardware name: Broadcom-v8A (DT)
Dec 28 18:39:59 kernel: task: ffffffc0192caa40 ti: ffffffc012eec000 task.ti: ffffffc012eec000
Dec 28 18:39:59 kernel: PC is at 0x29d34
Dec 28 18:39:59 kernel: LR is at 0x29fb4
Dec 28 18:39:59 kernel: pc : [<0000000000029d34>] lr : [<0000000000029fb4>] pstate: 200f0010
Dec 28 18:39:59 kernel: sp : 00000000ffe2c460
Dec 28 18:39:59 kernel: x12: 00000000000a211c
Dec 28 18:39:59 kernel: x11: 0000000000081d6c x10: 000000000000002d
Dec 28 18:39:59 kernel: x9 : 0000000000000003 x8 : 000000000000002d
Dec 28 18:39:59 kernel: x7 : 00000000f63ff65c x6 : 0000000000000036
Dec 28 18:39:59 kernel: x5 : 000000000000023f x4 : 00000000f63ff6bc
Dec 28 18:39:59 kernel: x3 : 0000000000000000 x2 : 000000000000002a
Dec 28 18:39:59 kernel: x1 : 000000000000000c x0 : 0000000000000000
Dec 28 18:39:59 kernel: potentially unexpected fatal signal 11.
Dec 28 18:39:59 kernel: CPU: 1 PID: 2728 Comm: dcd Tainted: P           O    4.1.27 #2
Dec 28 18:39:59 kernel: Hardware name: Broadcom-v8A (DT)
Dec 28 18:39:59 kernel: task: ffffffc0192caa40 ti: ffffffc012eec000 task.ti: ffffffc012eec000
Dec 28 18:39:59 kernel: PC is at 0x29d34
Dec 28 18:39:59 kernel: LR is at 0x29fb4
Dec 28 18:39:59 kernel: pc : [<0000000000029d34>] lr : [<0000000000029fb4>] pstate: 200f0010
Dec 28 18:39:59 kernel: sp : 00000000ffe2c460
Dec 28 18:39:59 kernel: x12: 00000000000a211c
Dec 28 18:39:59 kernel: x11: 0000000000081d6c x10: 000000000000002d
Dec 28 18:39:59 kernel: x9 : 0000000000000003 x8 : 000000000000002d
Dec 28 18:39:59 kernel: x7 : 00000000f63ff65c x6 : 0000000000000036
Dec 28 18:39:59 kernel: x5 : 000000000000023f x4 : 00000000f63ff6bc
Dec 28 18:39:59 kernel: x3 : 0000000000000000 x2 : 000000000000002a
Dec 28 18:39:59 kernel: x1 : 000000000000000c x0 : 0000000000000000

 

[n3cr0shark]

Thanks for the RT-AC5300. Some feedback:

Tried to dirty flash while part of my mesh and although it said it was successful, the router did not see the new code version. Soft rebooted and same situation (Note: I did not hard reboot by removing power). Wound up removing it from the mesh and updating it as a standalone. It took the upgrade this way without issue.

Then had issues readding to the mesh. It did not want to make use of the Ethernet uplink for the readd - until I did a hard power cycle of the router (removed power and replugged it in). After a hard power reset, it used the Ethernet port fine and it was added back into the mesh without issue.

 

[RMerlin]

Tried to dirty flash while part of my mesh and although it said it was successful, the router did not see the new code version.

The GT-AC5300 is not supported...

 

[n3cr0shark]

The GT-AC5300 is not supported...

Brainfreeze....I meant RT-AC5300 (I was staring at the GT-AX11000 when I typed that).

 

[Tech9]

This firmware is working very well on AC86U. I don't remember the Wi-Fi drivers version in 386.7_2 and run my standard Wi-Fi tests on it - the ones that fail on 388.1 firmware for AX86U. Everything connects quickly, stays connected, re-connects at router and client reboots, good link speeds - Wi-Fi is perfect.

 

[heysoundude]

This firmware is working very well on AC86U. I don't remember the Wi-Fi drivers version in 386.7_2 and run my standard Wi-Fi tests on it - the ones that fail on 388.1 firmware for AX86U. Everything connects quickly, stays connected, re-connects at router and client reboots, good link speeds - Wi-Fi is perfect.

This is good news because you're a tough one to please....but I'm still going to wait until it's a release rather than an alpha.
Maybe it'll be a New Year's treat from Merlin...

 

[JohnD5000]

This is good news because you're a tough one to please....but I'm still going to wait until it's a release rather than an alpha.
Maybe it'll be a New Year's treat from Merlin...

Is there going to be a general release before a beta? Alpha -> Release? And in a few days? Seems kind of fast even with the few positive responses so far.

 

[Tech9]

This is good news because you're a tough one to please...

I'm a perfectionist. Somewhat working is not acceptable for me. It's either working or not.

I have run all basic features so far and the router is stable. It has DHCP reservations, DoT (Google, OpenDns, Cloudflare), UPnP (not tested in Double NAT yet) and Port Forwarding, DDNS service to No-IP, VPN Server (both IPv4 and IPv4/IPv6), VPN Client to NordVPN (~160Mbps), AiProtection and Parental Controls enabled, tested both Adaptive (can't test correct categories though) and Cake QoS, Guest Network, DNS Director (to few different servers + router), Traffic Monitor with history on USB, Dnsmasq torture test (pass), no NTP issues, LED lights on/off test (pass), AiMesh page shows correct info, CPU temp shows correct values and graph... what most people will use or see. No scripts - clean firmware.

What's not working properly (the same as previous releases):
- AiMesh wireless node discovery, wired initial setup is needed
- The GUI kicks you out at preset in Administration, System time even if there is sliders movement and clicks on the screen
- Kernel buggy errors with IPv6 enabled + NAT acceleration, hidden
- Wrong DST time settings for GMT-05:00 zone
- IPv6 leak with on-router VPN Client enabled (IPv6 speed test is good enough)

Can't test all the rules in VPN Director, sorry. This one will take days. I do target testing based on my previous experience with the router and previous firmware versions. Something new I've never seen before in logs:

Dec 28 21:36:15 dnsmasq[1319]: reducing DNS packet size for nameserver 208.67.222.222 to 1232

 

[laracroftonline]

Wifi really has improved on 386.9 i'm pleasantly surpised.

 

[heysoundude]

Is there going to be a general release before a beta? Alpha -> Release? And in a few days? Seems kind of fast even with the few positive responses so far.

Merlin's year-end wrap-up post had some indicators...I'm not sure it's happened before, but as the focus seems to be on AX models, I won't be surprised if it does. Key word being IF

@Tech9 's testing leads me to believe Asus is cleaning up AC-model loose ends before moving (further) on into AX+ development, as Merlin has gently warned us will be his case as well.

 

[JohnD5000]

Merlin's year-end wrap post had some indicators...I'm not sure it's happened before, but as the focus seems to be on AX models, I won't be surprised if it does. Key word being IF

@Tech9 's testing leads me to believe Asus is cleaning up AC-model loose ends before moving (further) on into AX+ development, as Merlin has gently warned us will be his case as well.

Merlin's YE wrap he said "386.9 is currently in development, I am awaiting one missing GPL before the beta cycle can begin for this release. I don't expect 386.9 to be finalized until sometime in January."

Sounds like sometime in January GPL missing piece will be in and final and we can move to beta cycle from alpha. Although, I could be misreading it. Anyhow, I'm not expecting a final release on or near New Year's day.

 

[SBSTNGRSS]

Wifi really has improved on 386.9 i'm pleasantly surpised.

Can agree and running the alpha since release and no issues at all. Thx @RMerlin and other contributors!

 

[heysoundude]

Merlin's YE wrap he said "386.9 is currently in development, I am awaiting one missing GPL before the beta cycle can begin for this release. I don't expect 386.9 to be finalized until sometime in January."

Sounds like sometime in January GPL missing piece will be in and final and we can move to beta cycle from alpha. Although, I could be misreading it. Anyhow, I'm not expecting a final release on or near New Year's day.

The missing GPL is in - see @laracroftonline 's post confirming improved wifi for their AC5300. I would think beta is imminent

 

[Tech9]

@Tech9 's testing leads me to believe Asus is cleaning up AC-model

Even before playing with 386.9 I'm 95% sure this version is going to be the last Asuswrt-Merlin release compatible with AC-class routers. Playing with AC86U is more enjoyable for me at the moment because it works better than AX86U with 388.1 firmware. Early alpha release is already stable. I'm also sure very few people can do the testing I do simply because they actually use their routers as network core. I can reboot and reset as many times as I need.

I can also confirm for people with good working AC86U routers and mostly AC-class clients that upgrading to AX86U has practically zero investment return with exception of future support and only if the user is willing to go through the entire Asus firmware beta testing cycle all over again. I don't see any "vast performance improvements" and "lower latency" in AX86U when coming from AC86U. It only runs better >1 VPN clients and more user scripts.

Let's make AC86U great again!

 

[John Fitzgerald]

Let's make AC86U great again!

You have said that this unit is buggy and prone to hardware failure.

 

[Tech9]

You have said that this unit is buggy and prone to hardware failure.

Yes, exactly as the "hardware" original quote was about.

I wouldn't recommend getting one in 2023, but if you already have good working one - no rush to upgrade.

 

[RMerlin]

Even before playing with 386.9 I'm 95% sure this version is going to be the last Asuswrt-Merlin release compatible with AC-class routers.

Considering how widely used some models like the RT-AC68U are, I do plan on continuing with at the very least component updates for the time being. And Asus hasn't announced that they were EOL'ing any of these models at this time (not even in our internal discussions). They day they officially put them on EOL, I intend to ask for one final GPL update.

 

[Tech9]

I do plan on continuing with at the very least component updates

Thank you for the good intentions, but get ready for questions "is latest CVE-xxxx included" and "when is feature X coming"... as usual.