386.5 Alpha1?

[JGrana]

Great news on reverting dnsmasq. My remote AX86U does a dnsmasq Tainted panic then reboot daily...

 

[RMerlin]

Check out this PR #805 , if it's merged, we might see huge VPN performance improvements on HND routers running 386.5.

No you won't. OpenSSL doesn't use it.

As for Strongswan, the only thing that can make use of it, benchmarks I ran a few years ago showed that bcmspu was faster than using these cryptos.

 

[JGrana]

No you won't. OpenSSL doesn't use it.

As for Strongswan, the only thing that can make use of it, benchmarks I ran a few years ago showed that bcmspu was faster than using these cryp

No you won't. OpenSSL doesn't use it.

As for Strongswan, the only thing that can make use of it, benchmarks I ran a few years ago showed that bcmspu was faster than using these cryptos.

How about Wireguard? Its fairly fast as it is, just wondering what the kernel module uses.

 

[heysoundude]

Check out this PR #805 , if it's merged, we might see huge VPN performance improvements on HND routers running 386.5.

#804 will also bring performance improvements

#804 will have effects across the range, if it's not straying too far off the beaten path: Asus chose to follow Broadcom's lead for a reason, remember. there are engineers and designers who know why they did what they've done and we may end up outsmarting ourselves in the pursuit of performance gains.

#805 might be better for OpenVPN users, but I think everyone is after WireGuard's sheer speed...if it plays nicely with WG, then we're good.

 

[dbell]

2x RT-AX86U's running in wireless backhaul mode via 2.5GbE Ports with 2x QNAP QSW-1105-5T 5-Port Unmanaged 2.5GbE Switches.

Was on 386.4 (since release), flashed to 386.5 Alpha 1 using the following steps.

• Flashed AiMesh node waited 2 minutes for client devices to connect to the node.
• Flashed the main router. Needed to wait a few extra moments for the 5GHz band to come back up.
• Waited ~20 minutes after the main router had rebooted with new firmware. Rebooted system via AiMesh, System Settings, System Reboot. I did not reboot because of any network issues, this is what prevents them, ime.
• After an additional 60 + minutes, I performed another System Reboot as above.
  
  Network stable.

Wireless backhaul via 2.5 GbE ?

 

[L&LD]

Good catch! Wired, of course.

I'll edit my original post above.

 

[RMerlin]

How about Wireguard?

Wireguard uses Chacha20, not AES.

 

[ComputerSteve]

@RMerlin you think you can fix the issue with avahi-daemon changing names after reboot. I have it working on the current firmware 386.4 by creating a service-start containing "service restart_timemachine".. I'm hoping you can implement a fix on why that needs to happen after a reboot & it just doesn't work automatically. If I don't have that script created. After a normal reboot it changes names saying "Host name conflict, retrying with" Then it assigns a new name with numbers at the end. Only if I restart the service does it work correctly. I can either restart it with that script or manually (toggle) time machine off then on in the web interface.

 

[brambil]

Will this solve the bug about Wi-Fi 5Ghz slow that is in 386.4?

 

[RMerlin]

@RMerlin you think you can fix the issue with avahi-daemon changing names after reboot

I don't know because I can't reproduce the problem, so it's impossible for me to test anything. I did upgrade avahi to the latest version however.

 

[kernol]

@RMerlin - any idea when you plan to merge 46065 branch with the Master so self compiles can take advantage of code updates you have done subsequent to releasing the 386.5 Alpha1. I see some useful fixes which would be great to test. many thnx

 

[RMerlin]

@RMerlin - any idea when you plan to merge 46065 branch with the Master so self compiles can take advantage of code updates you have done subsequent to releasing the 386.5 Alpha1.

When I have the GPLs for all the other models. Right now I only have three, and a forth one that the upload got truncated.

Nothing prevents people from building from master. That's precisely why I keep the 46065 merges in a separate branch, so they don't prevent building other models.

 

[kernol]

When I have the GPLs for all the other models. Right now I only have three, and a forth one that the upload got truncated.

Nothing prevents people from building from master. That's precisely why I keep the 46065 merges in a separate branch, so they don't prevent building other models.

Many thanks for prompt response - all understood and will remain patient .

 

[tarzan2000]

AX86U - everything is ok.

IpoE 500Mbit
4 VPN client.s OpenVPN + VPN Director (on router)
2 VPN Servers OPVN on router
21 LAN clients (5 AX clients, 6 AC, 10 LAN)
GUEST 1 for IoT Device 2,4g
2.4N for Printer's
5 AC\AX for Home User's devivce
AiMesh paired with AS66U_b1 (386.4)
DualWAN (WAN + USB Stick)

 

[dsring]

Running on RT-AX86U for the last 24 hours with no issues. For me, this alpha is running better than 386.4. I started getting 'nf_conntrack expectation table full' errors on 386.4 which affected my network, and I ended up rolling back to 386.4_alpha2. Since flashing 386.5, I have not seen this error, or had any other issues. Thanks.

 

[albinoman887]

When I have the GPLs for all the other models. Right now I only have three, and a forth one that the upload got truncated.

Nothing prevents people from building from master. That's precisely why I keep the 46065 merges in a separate branch, so they don't prevent building other models.

On the subject of self-compiles I know that when one build for themselves there are no trend micro features but what exactly are those?

 

[RMerlin]

On the subject of self-compiles I know that when one build for themselves there are no trend micro features but what exactly are those?

Everything that uses their DPI engine. So AiProtection, Adaptive QoS, Traffic Analyzer, etc...

 

[kernol]

Running on RT-AX86U for the last 24 hours with no issues. For me, this alpha is running better than 386.4. I started getting 'nf_conntrack expectation table full' errors on 386.4 which affected my network, and I ended up rolling back to 386.4_alpha2. Since flashing 386.5, I have not seen this error, or had any other issues. Thanks.

New wireless drivers may well have helped ...

and a bunch of other fixes and improvements in the stock code ...

Spoiler: Asus Stock 3.0.0.4.386.46065 firmware - change.log

ASUS Firmware version 3.0.0.4.386.46065
Security
- Fixed string format stacks vulnerability
- Fixed cross-site-scripting vulnerability
- Fixed informational vulnerability.
Thanks to Howard McGreehan.

-Fixed SQL injection vulnerability
-Fixed json file traversal vulnerability
-Fixed plc/port file traversal vulnerability
-Fixed stack overflow vulnerability
Thanks to HP of Cyber Kunlun Lab

-Fixed authenticated stored XSS vulnerability
Thanks to Luke Walker – SmartDCC

-Fixed LPD denial of service vulnerability
-Fixed cfgserver heap overflow vulnerability
-Fixed cfgserver denial of service vulnerability
Thanks to TianHe from BeFun Cyber Security Lab.

Added more ISP profile
Digi 1 - TM
Digi 2 - TIME
Digi 3 - Digi
Digi 4 - CTS
Digi 5 - ALLO
Digi 6 - SACOFA
Maxis - CTS
Maxis - SACOFA
Maxis - TNB/ALLO

Fixed AiMesh guest network issues.
Fixed DDNS issues where the WAN IP is IPv6
Fixed UI bugs in Administration --> feedback.
Fixed time zone error.
Improved the connection stability.

 

[agbommarito]

New wireless drivers may well have helped ...
View attachment 39254

Are you sure there is a new wireless driver? I just looked at my router running 386.4 and it shows:

The timestamps are different, but the version numbers are the same. Could just be a recompile of the same code as before.

I've been hoping for a new wireless driver to fix the 2.4 connection issues.

 

[shabbs]

Are you sure there is a new wireless driver? I just looked at my router running 386.4 and it shows:

View attachment 39257

The timestamps are different, but the version numbers are the same. Could just be a recompile of the same code as before.

I've been hoping for a new wireless driver to fix the 2.4 connection issues.

Off topic question - where are you pulling this wireless driver info from? Don't see this on my Tools screen.