Asuswrt-Merlin 376.49 is out

[wgentine]

For RT-AC87U owners: I did a quick recompile of the firmware with the newest Quantenna SDK taken from the recent Asus release (since updating the SDK was a fairly simple thing to do). Grab it here:

http://www.mediafire.com/download/5z5zlb4z6yz412e/RT-AC87U_3.0.0.4_376.49_5_newqtn.zip

This uses the new Quantenna v36.7.0.31 driver which should resolve compatibility issues with Apple devices. Asus recommends doing a factory default reset when updating to this new driver - I suggest people first test it without doing so. If you still have connectivity issues with your Apple devices, then go ahead and do the recommended factory default reset + manual reconfiguration.

Been running it on my own router for the past few hours now while using the laptop connected over 5 GHz wifi without any problem.

You're awesome. Thanks! Happy new Year!

 

[wgentine]

Try disabling everything related to AiProtection and AiQoS, as the issue could be related to the TrendMicro DPI engine.

Also try by disabling NAT acceleration.

Both are closed source, and beyond my control.

Positively reporting the problem is CTF! Using regular NAT fixes it.
Thanks for giving these hypothetic causes.

Do you have ways to report it to ASUS or should I do it by myself?

 

[vdemarco]

Perhaps I spoke too soon, is this a sign of the compatibility issue (wrong address on an iPhone 5S)?

Jan 1 13:21:35 dnsmasq-dhcp[2365]: DHCPREQUEST(br0) 192.168.161.25 dc:9b:9c:xx:xx:xx
Jan 1 13:21:35 dnsmasq-dhcp[2365]: DHCPNAK(br0) 192.168.161.25 dc:9b:9c:xx:xx:xx wrong address
Jan 1 13:21:36 dnsmasq-dhcp[2365]: DHCPREQUEST(br0) 192.168.1.213 dc:9b:9c:xx:xx:xx
Jan 1 13:21:36 dnsmasq-dhcp[2365]: DHCPACK(br0) 192.168.1.213 dc:9b:9c:xx:xx:xx iPhone5S

This isn't a problem, the iPhone tries to ask for the address it already has, that fails then it asks for a new address.

This was done so it can get online faster.

 

[jgrove]

I have exactly what you describe, a 'Static Range' from 2-19 and the rest (20-254) on DHCP, but this goes against RMerlin's recommendations..

Address 192.168.161.25 in the log I showed in a previous post is definitely not in my Static Range, not even in my network (192.168.1.x) at all!

It was my wife's iPhone that had that strange address, I wonder if it is some address she picked up in a shop before coming home.

Try re-setting the iPhone network settings and re try, i think my ipad had this issue but i cant remember how i stopped it. Sorry

 

[highwire]

Perhaps I spoke too soon, is this a sign of the compatibility issue (wrong address on an iPhone 5S)?

Jan 1 13:21:35 dnsmasq-dhcp[2365]: DHCPREQUEST(br0) 192.168.161.25 dc:9b:9c:xx:xx:xx
Jan 1 13:21:35 dnsmasq-dhcp[2365]: DHCPNAK(br0) 192.168.161.25 dc:9b:9c:xx:xx:xx wrong address
Jan 1 13:21:36 dnsmasq-dhcp[2365]: DHCPREQUEST(br0) 192.168.1.213 dc:9b:9c:xx:xx:xx
Jan 1 13:21:36 dnsmasq-dhcp[2365]: DHCPACK(br0) 192.168.1.213 dc:9b:9c:xx:xx:xx iPhone5S

This is normal. My iPhone picks up a 10.128.x.x address when I'm at work. When I get home, I get similar logs as it attempts to refresh the work IP address on my home network - I get the 'wrong address' log with my work IP address, and then it gets an address assigned in on my home network (10.0.1.x).

 

[RMerlin]

@merlin i have installed this test build and since then I have sometimes connection but no internet. Disabling wifi on the iPhone en enable again fixes that. Normal _4 build does not have that issue. So the new driver is something off I guess.
I have ac87u and iPhone 6 connected in ac mode.

Then do the factory default reset Asus recommends to do with this new driver. Also delete and recreate any existing wireless profile on your iPhone.

 

[RMerlin]

Positively reporting the problem is CTF! Using regular NAT fixes it.
Thanks for giving these hypothetic causes.

Do you have ways to report it to ASUS or should I do it by myself?

You'll have to first test with their newly released RT-AC87U firmware to confirm that the issue still exists in this latest version (it was released yesterday). If it does, open a support ticket on their website.

 

[LifeboatDave]

She who must be obeyed has just complained that she can't get on the internet.
Looked in the log and found this:

Jan 1 20:09:06 dnsmasq-dhcp[665]: DHCPDISCOVER(br0) f0:27:65:13:a2:9b
Jan 1 20:09:06 dnsmasq-dhcp[665]: DHCPOFFER(br0) 192.168.1.21 f0:27:65:13:a2:9b
Jan 1 20:09:06 dnsmasq-dhcp[665]: DHCPREQUEST(br0) 192.168.1.21 f0:27:65:13:a2:9b
Jan 1 20:09:06 dnsmasq-dhcp[665]: DHCPNAK(br0) 192.168.1.21 f0:27:65:13:a2:9b no leases left

It's interesting that it's a static IP device and the only wireless device trying to connect on 5Ghz. It's an Andriod phone.
I tried connecting my phone on 5Ghz (again an android with a static IP) and got the same

Jan 1 20:15:53 dnsmasq-dhcp[665]: DHCPDISCOVER(br0) cc:3a:61:b0:6c:f3
Jan 1 20:15:53 dnsmasq-dhcp[665]: DHCPOFFER(br0) 192.168.1.14 cc:3a:61:b0:6c:f3
Jan 1 20:15:53 dnsmasq-dhcp[665]: DHCPREQUEST(br0) 192.168.1.14 cc:3a:61:b0:6c:f3
Jan 1 20:15:53 dnsmasq-dhcp[665]: DHCPNAK(br0) 192.168.1.14 cc:3a:61:b0:6c:f3 no leases left

Wireless log shows (under the 5Ghz section):

MAC IP Address Name RSSI Rx/Tx Rate Connected Flags
CC:3A:61:B0:6C:F3 192.168.1.14 -59dBm 390/390 Mbps 00:00:00 S
F0:27:65:13:A2:9B 192.168.1.21 -56dBm 433/390 Mbps 00:00:00 S
70E:E2:8D:90B 192.168.1.13 DavesIpad -57dBm 65/65 Mbps 00:00:00

DavesIpad is in airplane mode and not connected. .14 and .21 are our phones! Weird!

Trying to connect on 2.4Ghz ends in the same results.

Time for a reboot!

It's an RT- AC68U running 376.49_4

Cheers, and Happy New Year to all!

 

[lancer73]

If you read a few pages back, you'll see that you have probably hit a fundamental change in dhcp behaviour.

Are your static leases in the range defined for the dhcp server?


Sent from my iPhone using Tapatalk

 

[wgentine]

You'll have to first test with their newly released RT-AC87U firmware to confirm that the issue still exists in this latest version (it was released yesterday). If it does, open a support ticket on their website.

Thanks friend! Will test it. Still didn't know about this new firmware.
Keep you guys posted.

 

[LifeboatDave]

If you read a few pages back, you'll see that you have probably hit a fundamental change in dhcp behaviour.

Are your static leases in the range defined for the dhcp server?


Sent from my iPhone using Tapatalk

Hi lancer73,

My DHCP address pool is .100 to .105, all my static addresses are below .50

Regards

 

[lancer73]

Hi lancer73,



My DHCP address pool is .100 to .105, all my static addresses are below .50



Regards

There is your problem then. Try enlarging your pool and put your static addresses within the pool range and your problems will be solved.

(Asus started mimicking Microsoft dhcp behaviour, this is not Merlin specific)


Sent from my iPhone using Tapatalk

 

[Ronv42]

Hi lancer73,

My DHCP address pool is .100 to .105, all my static addresses are below .50

Regards

Yes is the latest firmware the DHCP server (DNSmasq) is now configured with a managed IP count based on the pool defined. So you only have 5 addresses that DNSmasq will manage.

There are two ways of fixing this. Increase the range of the pool or create a startup script to edit the DNSmasq config item:

dhcp-lease-max=<number>​

 

[LifeboatDave]

Thanks guys, I'll give it a go!

Just to save my sanity, did it used to be the case that static addresses used to have to be outside the DHCP pool?

 

[Ronv42]

Thanks guys, I'll give it a go!

Just to save my sanity, did it used to be the case that static addresses used to have to be outside the DHCP pool?

No the recommendation from Merlin was to assign static addresses in the DHCP range since the firmware now calculates the max number addresses based on pool.

I am still a strong supporter that the dhpc-lease-max value should be decoupled from the pool size since DNSmasq and other DHCP software allows static addresses to be defined outside of the pool. I have been doing this for years.

 

[LifeboatDave]

No the recommendation from Merlin was to assign static addresses in the DHCP range since the firmware now calculates the max number addresses based on pool.

I am still a strong supporter that the dhpc-lease-max value should be decoupled from the pool size since DNSmasq and other DHCP software allows static addresses to be defined outside of the pool. I have been doing this for years.

Thanks Ronv42 - I wasn't going mad then!
It's been a long time since I played with router software, and I was pretty sure I used to assign the static addresses outside of the DHCP pool.
Having recently upgraded to the RT-AC68U and Merlins firmware, It's made me blow a few cobwebs off the old grey matter

Thank you for everybodys efforts and kudos to RMerlin for his continued development - it is appreciated!

 

[wgentine]

Thanks friend! Will test it. Still didn't know about this new firmware.
Keep you guys posted.

Tried new firmware and same issue. Ill open a case in Asus Support.

 

[Dragster]

For RT-AC87U owners: I did a quick recompile of the firmware with the newest Quantenna SDK taken from the recent Asus release (since updating the SDK was a fairly simple thing to do). Grab it here:

http://www.mediafire.com/download/5z5zlb4z6yz412e/RT-AC87U_3.0.0.4_376.49_5_newqtn.zip

This uses the new Quantenna v36.7.0.31 driver which should resolve compatibility issues with Apple devices. Asus recommends doing a factory default reset when updating to this new driver - I suggest people first test it without doing so. If you still have connectivity issues with your Apple devices, then go ahead and do the recommended factory default reset + manual reconfiguration.

Been running it on my own router for the past few hours now while using the laptop connected over 5 GHz wifi without any problem.

Thank you Merlin, not that I have Apple devices (living in an Apple-free environment here ) but I downloaded the firmware and updated my RT-AC87u with it... we'll see how it goes... read about the new ASUS firmware and it seems that it has fixes for several issues.. but this new firmware only includes the updated Quantenna driver, right?

Cheers,
Dragster

 

[unsynaps]

I am still a strong supporter that the dhpc-lease-max value should be decoupled from the pool size since DNSmasq and other DHCP software allows static addresses to be defined outside of the pool. I have been doing this for years.

I'm a supporter as well.

My input on the topic is a little less... technical. Basically it has always worked for me int he past so I have no idea why doing this is broken now.

I have been putting reservations outside the pool since I started teaching myself linux over 15 years ago. I also was able to do out of range reservations on the Apple basestation (first embeded router I had that supported reservations) the RT-AC87R replaced. Well actually for a short time a WRT1900AC replaced the Apple basestation. But I had a mile long list on why the WRT1900AC is a complete pile of crap (lack of being able to reserve outside the range was one of the many) so it got replaced by the current RT-AC87R.

Somewhat old but rules and practices may have changed but this is what I went by some time ago:http://support.microsoft.com/kb/196066
Oddly here is a post on Netgear of someone complaining the opposite. Netgear forced him to set reservations OUTSIDE the pool and he complained they should be INSIDE:http://forum1.netgear.com/showthread.php?t=84666
PFSense saying the reservations should be ouside because of how ISC dhcpd works:https://doc.pfsense.org/index.php/Why_can't_I_have_static_mappings_inside_my_DHCP_range

 

[chele809]

Is it posible to have the 87R Email me When someone connects to the wifi, with a scripts or something? THIS would be a nice feature. For Many many reasons. Can someone elaborate if possible. Thanks in advance.