[Beta] Asuswrt-Merlin 380.59 Beta 2 is available

[barbara]

it could still break with a future GPL update

I was incorrect to say that CONFIG_MD does not work with Tuxera. I managed to compile a Linux kernel for Asuswrt-Merlin 380.59, with the following features and it works with the Tuxera filesystem drivers.

Multiple devices (MD), Dm-crypt
XTS block mode
Crypto digests, ciphers and compressions
HID raw
USB HID device
USB port serial drivers, no USB modem, no WiMAX
Pulse per second (PPS)

As proof, I've attached "config_rt-ac68u" and "Module.symvers", so you can see that the module symbol versions are compatible with the Asus stock Linux kernel. Checking the "Module.symvers" file against a good one for compatibility, before flashing the firmware to the router, saves much time and potential trouble.

~/asuswrt-merlin/release/src-rt-6.x.4708/linux/linux-2.6.36/config_rt-ac68u
https://cryptobin.co/a0i307c2, password is lol

~/asuswrt-merlin/release/src-rt-6.x.4708/linux/linux-2.6.36/Module.symvers
https://cryptobin.co/w6s847d2, password is lol

Now, when I try to enable CONFIG_BRIDGE_NETFILTER kernel feature, it breaks compatibility with the Tuxera filesystem drivers in Asuswrt-Merlin 380.59. Oh well, I mostly have all the extra kernel features I need, for the moment anyway.

 

[M.Janse]

Hey,

First of all I'd like your work on the modified firmware.
But I found an error with the samba shares on the router.
When i write a lot of small sized files to a share, the share locks for a second and creates an error in the program that uses the shared directory.
I noticed this error messages in the routers log file:

kernel: tntfs error (device sda1, pid 1006): ntfs_mft_data_extend_allocation_nolock(): Not enough space in this mft record to accommodate extended mft data attribute extent. Cannot handle this yet.
kernel: tntfs error (device sda1, pid 1006): ntfs_mft_record_alloc(): Failed to extend mft data allocation.
kernel: tntfs error (device sda1, pid 1006): ntfs_create_helper(): Failed to allocate a new on-disk inode (error 95).
kernel: tntfs error (device sda1, pid 1006): ntfs_create_helper(): Failed (error 95).

I've connected a 1 TB harddisk on the USB2 port and I have more than 500 GB of free space left.
I've mounted the share from a linux based os withe the command:

smbmount //192.168.1.1/"rec3 (at BACKUP)" /mnt/smb2 --verbose -o username=user,password=pass,rw,sec=ntlm,domain=WORKGROUP,nosetuids,noperm,cache=loose

But I also noticed the problem from a windows 7 computer using total commander to access the share.
I'm using the merlin-380-59-beta-2 firmware on a Asus RT-AC68U.

It would be nice if there was a fix or a way to work around for this problem.

 

[alistair]

What really disturbs me, is that my VoIP only works properly when I set the SIP Passthrough to Enable. As soon as it is set to Enable + NAT helper, it stops working. While it is awesome that the Merlin firmware now allows me to properly use VoIP, it disturbs me that it is now impossible for me to return to the official Asus firmware, because the official firmware lacks this feature. I'd really like to know what is causing this problem and Asus should fix/change this, since I shouldn't be dependent on alternative firmware for such an essential feature to work right.

Edit: I'm using a AC68u.

 

[abescalamis]

I noticed that the icon that represents a laptop is gone, but it does show if the router detects it as laptop, I have 3 laptops and only one was detected as laptop and assigned the correct icon, I tried to do it manually for the other 2 laptops but the icon is not there.

 

[RMerlin]

kernel: tntfs error (device sda1, pid 1006): ntfs_mft_data_extend_allocation_nolock(): Not enough space in this mft record to accommodate extended mft data attribute extent. Cannot handle this yet.

Looks like a filesystem-level issue. Try doing a chkdsk on your disk.

Otherwise, nothing more that I can do, the filesystem handling is done by the Tuxera module.

 

[Broken Hope]

Is it possible to have a way to select the different NAT acceleration levels in the drop down box, rather than just auto and disabled?

My AC68U keeps incorrectly stating it can do CTF+FA, if I do a factory reset it correctly states it can do CTF only as it should be.

Since updating to beta 2 it's incorrectly reporting again.

 

[RMerlin]

Is it possible to have a way to select the different NAT acceleration levels in the drop down box, rather than just auto and disabled?

My AC68U keeps incorrectly stating it can do CTF+FA, if I do a factory reset it correctly states it can do CTF only as it should be.

Auto will automatically enable everything that can be enabled. Various features are not compatible, and therefore the firmware will automatically disable these as necessary.

Forcing them to be enabled wouldn't work any better. If your router doesn't show it as being enabled, then it's because it actually can't enable it.

FA is currently only supported on the AC3200, AC68 (certain revisions only - there must be a /proc/fa entry present to confirm HW-level support) and AC87, with a long list of limitations. It's incompatible with QoS, VLANs, IPTV, STP, Repeater Mode, Static WAN IPs, and possibly others I might have missed.


Last I heard, FA was being phased out because it was causing too many compatibility issues. None of the newer routers have FA support for instance.

 

[Broken Hope]

I was actually wanting to just force CTF since I know my version of the 68U doesn't support FA, so I'm not sure why it keeps saying it's enabled.

 

[MarkRH]

I noticed that the icon that represents a laptop is gone, but it does show if the router detects it as laptop, I have 3 laptops and only one was detected as laptop and assigned the correct icon, I tried to do it manually for the other 2 laptops but the icon is not there.

Yeah, I noticed the firmware sees my Samsung TV as a "Samsung" icon and device initially but at some point thinks it's a Netgear NTV300 device with a different Icon. There is no Samsung icon to change it to. So I downloaded the blue PNG one from here: http://www.iconarchive.com/search?q=samsung and uploaded it to the router. So, you can probably upload your own Laptop icon to the router.

 

[stambeccuccio]

RT-AC68U
380.59_beta2
Firefox 46.0.1

A probable bug: Reboot Scheduler strange relationship with Master Password.

1) Enable Reboot Scheduler: Yes
2) Apply
3) Enable Reboot Scheduler: No
4) Apply

The router says Retype Password *Should be matched.

I tried several times and there is always this "strange" problem / bug with Firefox

Instead problem does not occur with Chrome 50.0.2661.94
Bah!

 

[john9527]

The router says Retype Password *Should be matched.

Turn off any password managers you have active for the router address (it's most likely autofilling a wrong password for you).

 

[stambeccuccio]

Turn off any password managers you have active for the router address (it's most likely autofilling a wrong password for you).

Thanks John, you're right as usual.
Resolved. I deleted passwords stored.
Now this problem no longer occurs.
Thank You again.

 

[dfran1]

I was seeing some wifi dead spots on this version so I went back to .59 alpha 2 4-5-16 on my RT-AC5300.
My 5300 would also not switch off channel 6 in smart connect mode on this version, no matter what setting's I changed, .59 alpha 2 4-5-16 went to channel 11 in smart connect as 6 is used by a neighbor
This could be related to nvram save r23 or not ?? after factory default (that I did in the past with the 87). I did not factory default and configure from scratch without the nvram save, I just went back to the version known best for me for now.
EDIT:
The wifi dead spots consisted of weaker 5.2GHz (or dead spots, that lead to switching to 2.4, not seen on the alpha), also a 2.4GHz dead spot (also not seen on the alpha)

 

[RMerlin]

I have a Samsung TV which connects over 2.4Ghz WiFi on my RT-AC87U, and although I've set the name in the Network Map; the Wireless Clients log (System Log -> Wireless Log) shows it as having no name, the name field is just blank. Network map works fine, other devices also seem fine.

*EDIT: I noticed that when it switches over from 2.4Ghz to 5Ghz, then the name shows as a "*" instead of just blank... which also isn't really correct either.

It's probably always been this way, only noticed it now though.

Samsung's engineers are doing some pretty silly things in their SmartTVs, such as trying to request the use of "localhost" as their hostname...

 

[rljames]

This new scheduler is only available on ARM models (that means no RT-N66U or RT-AC66U. Their kernel is simply too old.)

Is Merlin v380_59 the first release with features not supported by the ARM AC66 router model?
If so, are the related non-functional options "hidden" from use?

Thanks,
Bob

 

[johtaja]

Is Merlin v380_59 the first release with features not supported by the ARM AC66 router model?
If so, are the related non-functional options "hidden" from use?

Thanks,
Bob

AC66U is MIPS, not ARM

 

[RMerlin]

Is Merlin v380_59 the first release with features not supported by the ARM AC66 router model?
If so, are the related non-functional options "hidden" from use?

Thanks,
Bob

The setting won't be visible for models that don't support it.

 

[vibroverbus]

How come Smartifi is still not compatible with AC56U? Is this model missing the ASSIA agent or the CloudCheck app simply needs an update to support this router?
The app is saying the only supported asus routers are AC3200, AC66R/U/W, AC68P/R/U/W and N66R/U/W.

I just tried this out for the heck of it. Would not start on my Beta2 AC3200.
Manually nvram'd enable_cloudcheck=1, had to reboot for it to take effect, now works fine.

On the fence whether I like it better or worse than the Asus app (using iOS here...). Clearly the "optimization" thing is silly stuff - all mine recommends is that some wifi devices have faster capabilities than the network (2.4G) they happen to be connected to. Not really useful in my case, only 3 devices connect to the 2.4G, 2 do NOT have faster capabilities (legacy) and the other one is a bit far away so it indeed is sensibly choosing the 2.4G instead of the 5G network.

Sorry to digress. Bottom line cloudcheck does in fact work fine w/ 380.59.beta2 on my AC3200.

 

[RMerlin]

Spent even more time trying to get ingress QoS to work, using Tomato's IFB implementation. Turns out you CANNOT use IFB to apply QoS if relying upon packet marking - the marking occurs only after the traffic gets handled by the traffic scheduler. So basically, Tomato's IFB-based inbound QoS doesn't work either - all inbound traffic is then thrown into the default class... IFB can only be used if the classification is done in some other method (DiffServ would probably work, provided any ISP actually used that, which I doubt).

The only way around this is to use IMQ, which is not possible as it requires kernel-level changes that would break the wireless driver.

 

[vaszago]

QoS will never rational functioning.