Asuswrt-Merlin 378.55 is now available

[RMerlin]

AMPDU RTS
DTIM Interval
Beacon Interval
Enable WMM APSD

Surely there was no need to disable "everything" for 5ghz. But it's not worth the trouble for me to investigate further for now, because everything's fine even with all those settings off.

I recommend re-enabling or setting back to their default values at least those settings above, to prevent other issues.

 

[joegreat]

I recommend re-enabling or setting back to their default values at least those settings above, to prevent other issues.

Hi,

I had similar issues with 5 GHz WLAN on the AC68U router after updating to .55 release. Disabling 'beamforming' made it better, but did not fully correct the issue.

I got things back to work fully only after I did
- a factory reset and
- restored all user values with User NVRAM Save/Restore Utility, plus
- forgot (removed) the configuration on the client side and reconnected.

Now the 5 GHz WLAN works again also with enabled 'beamforming'!

With kind regards
Joe

 

[Phylion]

Today I finally made the upgrade to 378.55 on my AC66 router. I came from 378.53_0 release. Upgraded first to 378.54 and then to 378.55.
Unfortunately things did not went smoothly. A reboot by power flip made the router not be able to connect to WAN. To fix that I've tried several things ( rebooting, cable pulling, flashing back 378.54 and so on). Finally I tried to disable all WAN settings ( port, NAT & upnp), that trick made the box come up again.

But now I've have an issue with the DDNS service of Asusscom. It keeps saying :"unauthorized registration request". Does anyone know how to fix this? I need the ddns name because of the use of it in a few VPN configs.
The log file:
Aug 20 11:34:26 ddns update: ez-ipupdate: starting...
Aug 20 11:34:26 ddns update: connected to nwsrv-ns1.asus.com (103.10.4.108) on port 80.
Aug 20 11:34:29 ddns update: Asus update entry:: return: HTTP/1.1 401 |Authorization failed^M Date: Thu, 20 Aug 2015 10:34:25 GMT^M Server: Apache/2.4.9 (Unix) PHP/5.5.14 OpenSSL/1.0.1h^M X-Powered-By: PHP/5.5.14^M Content-Length: 0^M Content-Type: text/html^M ^M
Aug 20 11:34:29 ddns update: retval= 2, ddns_return_code (,401)
Aug 20 11:34:29 ddns update: asusddns_update: 2

 

[joegreat]

Today I finally made the upgrade to 378.55 on my AC66 router. I came from 378.53_0 release. Upgraded first to 378.54 and then to 378.55.

Hi,

I did not find the magic words "factory reset" in your update chain...

Due to the jump from multiple releases (and my own experience) I strongly suggest to you a factory reset and re-configuration of the router. This will make sure that any of the many hidden settings is reset to the correct value of the last firmware version.

To avoid long re-typing of settings, you can save your specific configuration to an USB drive with the User NVRAM Save/Restore Utility and get it back with the restore script after the factory reset.

With kind regards
Joe

 

[NomadTech]

Asuswrt-Merlin 378.55 is now available for download.

The highlights of this new release:

This build merges with Asus's GPL 378_6117, wiwth a few fixes backported from 6975. This new GPL includes a few important changes:

• New login scheme, based on a token rather than on HTTP basic authentication. This is far more secure against crosssite exploits.
• OpenVPN keys and certificates migrated to JFFS the first time you start an OpenVPN client or server. This frees up a good amount of nvram.
• NAT acceleration for PPTP/L2TP WAN users (used by some Russian ISPs). Only on ARM-based routers at this time.
• New client list on the networkmap page

Other important changes in this release:

• All System Log pages have been re-designed as tables, to make them much easier to read.
• Various components updates: OpenVPN, OpenSSL, ffmpeg.
• Additionally, minidlna was updated to the latest upstream version, and some of Asus's customizations were reverted, fixing the stability issues of the recent releases.
• To deal with OpenVPN clients that upgraded to the latest OpenSSL version, the automatically generated Diffie-Hellman is now 2048-bit instead of 512-bit. For performance reasons, a pre-generated one is used (this is perfectly safe). Server instances configured with any DH below 1024-bit is automatically upgraded to this 2048-bit one.
• Resource optimizations on the RT-AC87U (4 MB of RAM saved) and other ARM models (buffers are no longer regularly flushed, for better overall performance - that buffer memory will get automatically freed if required by actual applications, so higher memory usage is perfectly normal).
• A few other bug fixes and improvements - see the changelog for more info

Downloads are here.
Changelog is here.

Discovered a BUG where if you disable 5GHz-1 in AC-3200 it will also disable temperature reading for 2.4 GHz radio.
Let me know if you need screenshots or any other info.

 

[m_listed]

I don't know if anyone else has reported this, but since going from .54.2 to .55, my hard drives spin down after like 5 minutes of idle use, instead of the 8 hours that I set it to, or any other time setting. Normally this would be a good thing, since having Download Master enabled previously would prevent hard drives from spinning down, but since .55 the hard drives just spin down with a mind of their own instead of any set time. Just a report.

 

[AndyMS]

Third crash this evening after ensuring NAT acceleration and all beamforming options were disabled. Performed a factory reset, again disabling beamforming and NAT acceleration. Will report back here if it crashes again.

 

[RMerlin]

Discovered a BUG where if you disable 5GHz-1 in AC-3200 it will also disable temperature reading for 2.4 GHz radio.
Let me know if you need screenshots or any other info.

Thanks, fixed with this commit.

 

[RMerlin]

I'll have to see if I can actually reproduce the issue. Could be a bug in the code that checks for the presence of an existing certificate, and causes a buffer overrun. That's the only explanation I could see for a kernel panic to occur there.

The wiki article should probably simply be taken down by now, since both Asus and I are no longer storing certificates in nvram. This article will simply confuse a lot of users out there.

There was a bug in the code checking the validity of the DH that could cause preinit to crash if the DH was corrupted/invalid. I suspect that's what was hitting you.

https://github.com/RMerl/asuswrt-merlin/commit/06878dafa6df5df2a6c5c48b42c893bb54918cec

 

[kklee]

Hi RMerlin,

In this latest version do I have to block TOR node incoming connection manually as per your wiki instruction, Or all TOR node incoming connections are blocked by default?

Hope to hear from you soon.

 

[pcfr33k]

As I keep saying, wireless performance is tied to the closed source driver used by Asus in their firmware, and totally out of my control Go through the troubleshooting steps posted on the forums, a LOT of issues can be resolved through these steps. Just disabling beamforming alone will resolve the vast majority of issues with 5 GHz clients.

Are both the 2.4/5ghz chipsets broadcom on the 66u,68u,87u and 3200? I read on one of your posts the 5ghz chipset was either non broadcom or used a different manufacturer driver?

 

[L&LD]

The RT-AC87U uses an additional Quantenna radio and cpu chipset.

 

[tsunami2311]

So i just removed a Mac address for "name-phone" add a new mac address "name-phone" all is good mac address show only the new phones mac address,

But in the QOS> user-defined QOS rules

Source iP drop box shows 2 "name-phone" mac address the old one I removed from router settings in mac addresses and the new "name-phone" mac address. I already rebooted the router but i cant get user-defined QOS rules to stop showing the OLD mac address in source ip

by "name-phone" i mean when Some one changes there phone I keep "name-phone" but old mac address is removed, but in User-Define QOS rules I will have 2 "name-phones" 1 with the OLD mac and new mac.

 

[tsunami2311]

So i found out the client List is what lists "Source ip or Mac" in QOS user-defined QOS rules.

Thing is there dont seem to be away to DELETE from the client list if the DEVICE is not actual Connected atlest not that i can see?? If said device is say phone and is completely dead, there is no way of removing it from the client list, short of actually factory defaulting.

As far as I can tell Client List only show connected devices so offline devices or broke devices(that dont power on) can be seen or removed, or Am i missing something

 

[RMerlin]

So i found out the client List is what lists "Source ip or Mac" in QOS user-defined QOS rules.

Thing is there dont seem to be away to DELETE from the client list if the DEVICE is not actual Connected atlest not that i can see?? If said device is say phone and is completely dead, there is no way of removing it from the client list, short of actually factory defaulting.

As far as I can tell Client List only show connected devices so offline devices or broke devices(that dont power on) can be seen or removed, or Am i missing something

378.55 actually does have a Delete button when you edit a client. Asus implemented it, but had kept it disabled possibly because it was unfinished/untested. I've enabled it in 378.55, but afterward it was reverted, because Asus did a lot of changes to that code. So since the client list code is constantly changing these past few months, I ain't touching that again until it stabilizes.

 

[SeCrEt_BoY]

I have a problems with 378.55 on AC87R.
I can not load some website. (Enable ZenMate on browser >>> loading OK) I try to reset my AR87R but after 1 or 2 weeks, that problems happens again and my network is so slow. I reflash official FW and everything works fine (more than 3 weeks...) I still dont know what caused that problems...

 

[octopus]

RMerlin

When DDNS update "Forced refresh interval (in days)" then watchdog get crazy and update every 30sek after that.
Im using afraid.org and custom script.

Aug 24 19:47:38 watchdog: start ddns.
Aug 24 19:47:38 rc_service: watchdog 770:notify_rc start_ddns
Aug 24 19:47:38 custom script: Running /jffs/scripts/ddns-start (args: 81.xxx.xxx.xxx)
Aug 24 19:47:39 ddns: Completed custom ddns update
Aug 24 19:48:08 watchdog: start ddns.
Aug 24 19:48:08 rc_service: watchdog 770:notify_rc start_ddns
Aug 24 19:48:08 custom script: Running /jffs/scripts/ddns-start (args: 81.xxx.xxx.xxx)
Aug 24 19:48:09 ddns: Completed custom ddns update
Aug 24 19:48:22 kernel: nvram: consolidating space!
Aug 24 19:48:38 watchdog: start ddns.
Aug 24 19:48:38 rc_service: watchdog 770:notify_rc start_ddns
Aug 24 19:48:38 custom script: Running /jffs/scripts/ddns-start (args: 81.xxx.xxx.xxx)
Aug 24 19:48:39 ddns: Completed custom ddns update

 

[tsunami2311]

378.55 actually does have a Delete button when you edit a client. Asus implemented it, but had kept it disabled possibly because it was unfinished/untested. I've enabled it in 378.55, but afterward it was reverted, because Asus did a lot of changes to that code. So since the client list code is constantly changing these past few months, I ain't touching that again until it stabilizes.

delete client works and the option is there the problem is IF the client is not connected to router Client list dont show it there for you cant delete it.

Client list only show Online devices not Offline devices, is what i trying to say so long is it online it can be deleted cause it shows the option to delete it, but if its offline and or can not longer powered on there is no way to remove the device cause it wont be listed.

It not huge issue as it this not done often, it just annoying. if that is also ASUS issue then i guess it just wait game.

 

[VANT]

i found small bug.
In wireless/professional tab i have set tx power adjustment to 50%

but when i go to Time Setting in Enable wireless scheduler, and press cancel or apply, slider is set to max but still indicated 50%.

 

[lbradio]

Not sure, I suspect it's related either to IPv6, or to TrendMicro's kernel modules. The error message is too vague to be of any help in tracking it down at this time.

A lot of

kernel: [truncated] L/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/npch ==NULL/np

as well in the syslogfile.
Router is performing well.
Did anyone by accident find the solution yet?