Asuswrt-Merlin - Custom firmware for RT-N16

[RMerlin]

its not really a problem, I just found it strange that asus would include default names that would give an error.

Choose any service from the drop down list that has "-" in the name, for example, "some - service" , when you click add, an error pops up saying the service name cannot contain "-" !

Wouldn't be the first time - they had Virtual Server preset that would break the iptable lists in the past

Also about the QoS, how does one differentiate between the priority of highest settings?
Say, I want VOIP/Gaming/Web to all be highest, but in that order. So if I'm downloading something, and someone on wifi is browsing, they should get priority, but if I'm browsing and someone is using VoIP, they should get priority.. how does it work if they're all set to highest?
Or should only VoIP be set t highest, then gaming to high, then HTTP to medium?

Setting both VoIP and HTTP to Highest should be fine. Note that the HTTP rule with a high priority is only for small transfers (below 512 KB), so it should still leave plenty of bandwidth left for VoIP. Large HTTP transfers are under a lower priority rule. Otherwise if you wanted to give absolute priority to one, you would have to drop the other ones to the priority setting just below.

 

[richardeid]

1) Power adjustment isn't enabled by Asus on the RT-N16. I don't know if it's because it's not adjustable on the RT-N16 hardware or just because they didn't enable it.

I haven't played with this at all, but on my RT-N16 I can access: http://RT-N16/Advanced_PerformanceTuning_Content.asp

If I were to make a change there, would that not do anything? Just curious and I have no intent on adjusting anything on this page. What I see is 80mW for 2.4GHz, blank for 5GHz, Temperature unit set to C and 'auto' for Spin duty cycle. Also, the graph on top is totally blank, for obvious reasons.

But you're saying nothing here is technically enabled for RT-N16?

 

[mstombs]

Just flashed to check it out, can see but not yet tried guest wireless - but what is meant by dual-wan? Is it the backup via 3G usb?

 

[RMerlin]

Just flashed to check it out, can see but not yet tried guest wireless - but what is meant by dual-wan? Is it the backup via 3G usb?

The regular option found in Asus's firmware is just for USB failover.

Dual Wan is a more advanced version of it that is still (AFAIK) under development by Asus. Dual Wan means that you can use not only a USB device, but also assign one of the LAN ports on the router to be connected to a second ISP. The second Dual Wan device can also be used for failover or load-balancing.

This is still fairly experimental AFAIK, since Asus hasn't yet enabled that feature in any of the available devices, so I cannot say if it's 100% working yet or not.

 

[RMerlin]

I haven't played with this at all, but on my RT-N16 I can access: http://RT-N16/Advanced_PerformanceTuning_Content.asp

If I were to make a change there, would that not do anything? Just curious and I have no intent on adjusting anything on this page. What I see is 80mW for 2.4GHz, blank for 5GHz, Temperature unit set to C and 'auto' for Spin duty cycle. Also, the graph on top is totally blank, for obvious reasons.

But you're saying nothing here is technically enabled for RT-N16?

That whole page isn't linked to the web interface on the RT-N16 specifically because it's only meant for the RT-N66U. So, it's normal that nothing works or even makes sense at all on this page when you manually access it on an RT-N16.

 

[mstombs]

The regular option found in Asus's firmware is just for USB failover.

Dual Wan is a more advanced version of it that is still (AFAIK) under development by Asus. Dual Wan means that you can use not only a USB device, but also assign one of the LAN ports on the router to be connected to a second ISP. The second Dual Wan device can also be used for failover or load-balancing.

This is still fairly experimental AFAIK, since Asus hasn't yet enabled that feature in any of the available devices, so I cannot say if it's 100% working yet or not.

... that's exactly the sort of dual-wan I am interested in, there's a ripped off Chinese version of tomatousb that does that - also allows a wireless wan as backup, but even after breaking their web page change block and translating the web pages I found it too buggy to use. From bits of Asus code I have seen the original Broadcom SDK did support multi-wan and there are dd-wrt scripts to do it, but is there anything in the new Asus web interface to configure?

 

[RMerlin]

is there anything in the new Asus web interface to configure?

It's entirely configurable through the web interface. Just need a custom build that has the feature enabled (like my 157.12 build has).

 

[AnthonyArmato]

So technically you could run two cable modems (of course you'd have to pay for two lines) and the router will load balance the two when using your firmware version ?

Have you or anyone tested it ?

 

[RMerlin]

So technically you could run two cable modems (of course you'd have to pay for two lines) and the router will load balance the two when using your firmware version ?

Have you or anyone tested it ?

That's correct.

I haven't tested it because I don't have anything to test it myself.

 

[AnthonyArmato]

This would be perfect for me. I've been wanting faster internet but Comcast currently only offers 16/2 here. Adding another line and being able to load balance the two would be amazing.

 

[RogerSC]

This would be perfect for me. I've been wanting faster internet but Comcast currently only offers 16/2 here. Adding another line and being able to load balance the two would be amazing.

Or, in my case, two DSL lines at 3Mb/s each, to replace the crappy Comcast cable that we have to use here. That would give me 2-3 times the download speed that I get in the evening via Comcast. Good thought.

 

[RMerlin]

Just bear in mind that load balancing doesn't mean your speed doubles. It means you can do two downloads at once, both at full speed.

So in essence, it could make torrents up to twice faster, but downloading a large file from a website would be only at the speed of one connection (unless using an advanced download manager that does segmented downloading).

And also, just reminding everyone that unless anyone here actually tests it, there is no guarantee that this feature is even fully working yet. So, don't jump in on getting that second connection just yet...

 

[mstombs]

It's entirely configurable through the web interface. Just need a custom build that has the feature enabled (like my 157.12 build has).

Yes I am running your build - but no sign of any config enable link on any display found! I can read the www code and see I need to enable a link to

Advanced_WANPort_Content.asp

going directly the content doesn't work properly, so its so close!

state.js needs a positive reply in one of these calls I think

state.js:var dualWAN_support = rc_support.search("dualwan");
state.js:var dualWAN_support = rc_support.search("dualwan");

BTW I know/understand why the git source is behind the beta - but the current git doesn't build RTN-16 for me - obscure errors in sqlite. Writing over with the Asus 144 sources fixes the problem for me. Asus 144 do not compile due to missing wps prebuilt which I see you have fixed in your git, but skip that step it does complete for me including sqlite to massive 18MB firmware.

 

[RogerSC]

Just bear in mind that load balancing doesn't mean your speed doubles. It means you can do two downloads at once, both at full speed.

So in essence, it could make torrents up to twice faster, but downloading a large file from a website would be only at the speed of one connection (unless using an advanced download manager that does segmented downloading).

And also, just reminding everyone that unless anyone here actually tests it, there is no guarantee that this feature is even fully working yet. So, don't jump in on getting that second connection just yet...

True, I was more thinking of being able to stream a couple of Netflix movies at one time, which could be done with two lines (or one line, if it was fast enough *smile*). The one line that I have now barely has enough throughput in the evening to stream one movie, sometimes it cannot. Comcast is evidently working on making sure that no one stops using their cable TV service, since they make it impossible to stream at night. By the way, I'm paying for 12Mb/s from Comcast, and get 1-3Mb/s at night. The only other choice I have is 3Mb/s DSL, which would be fine if I could use two lines. And I know that it's not there yet *smile*, but something to look forward to. I'd love to just flip off Comcast, but at this point I'm kind of stuck.

 

[RMerlin]

Yes I am running your build - but no sign of any config enable link on any display found! I can read the www code and see I need to enable a link to

Advanced_WANPort_Content.asp

going directly the content doesn't work properly, so its so close!

state.js needs a positive reply in one of these calls I think

My apologies - Dualwan is currently only enabled on the RT-N66U. I tend to answer messages here without checking first if I'm on the RT-N16 or RT-N66U thread, so I overlooked that part. My mistake.

I'll try to compile the next beta with dual wan enabled on the RT-N16. The next beta should be out sometime this weekend if all goes as planned.

BTW I know/understand why the git source is behind the beta - but the current git doesn't build RTN-16 for me - obscure errors in sqlite. Writing over with the Asus 144 sources fixes the problem for me. Asus 144 do not compile due to missing wps prebuilt which I see you have fixed in your git, but skip that step it does complete for me including sqlite to massive 18MB firmware.

I usually get that sqlite-related error if I try to build the rt-n16 version without first doing a "make clean" to ensure there is no leftovers from the rt-n66u previous build. The git version should compile as I based my first RT-N16 build on it, however it's also possible I forgot to push a binary file to it. Just give me the actual path of the missing file, and I'll look into it. But try first doing a "make clean" followed by "make rt-n16", that might fix it.

 

[RMerlin]

Updated OP with 3.0.0.3.157.12 Beta 2.

The big changes in this version are the new NAT loopback code, and Dual WAN being enabled (experimentally) for the RT-N16. Please give this a good test run, especially regarding the Dual WAN feature.

UPDATE: I forgot to mention that output power control was also manually enabled for the RT-N16. I'll need feedback on whether it's working or not.

 

[mstombs]

Gosh weekend comes early your side of the pond! Many thanks.

I now have guest wifi working on lan, and dual dhcp connections on wan in load-sharing mode on my RT-N16.

I chose LAN1 for secondary WAN, would have been better to select LAN4 as it is physically closer to WAN port - and I was using LAN1 for LAN...

I can report it is basically working, status of secondary WAN display on web gui could be improved, and interface names are no longer familiar, but sensible:-

admin@RT-N16:/tmp/home/root# nvram show | grep vlan
size: 25492 bytes (7276 left)
vlan3hwname=et0
vlan2ports=0 8
wan0_ifname=vlan2
wl0_vlan_prio_mode=off
wan1_ifname=vlan3
vlan2hwname=et0
wl_vlan_prio_mode=off
lan_ifnames=vlan1 eth1 wl0.1
wan_ifnames=vlan2 vlan3
vlan3ports=4 8
vlan1hwname=et0
vlan1ports=1 2 3 8*
wandevs=vlan2 vlan3
landevs=vlan1 wl0

If I traceroute to say google I can see routing automatically switching between ISPs, I will continue to use/check. Already working better than Chinese tomato dualwan when I last tried - that didn't like my adsl half-bridge modem with short lease times.

However load-balancing mode seems to get broken if either WAN goes down, which makes it useless for me (2 unreliable WAN connections!) - so I am now looking at the failover mode and wondering about how recovery etc works.

Note Asus uses the 'wanduck' routine to determine state of wan connections, Linux in my experience has always had issues with media state and dead network detection - maybe there's something that can be back-ported from later official kernel versions - surely the kernel knows when it can't communicate down a certain path without an external ping script?

Another specific issue - where ISP DNS servers are supplied by dhcp, specific routes to them via the specific interface should be set automatically - it looks like web interface can be used to do this manually.

Thanks again for your compiles - and big plus for Asus for making such new features available on the now mature RT-N16!

 

[RMerlin]

Yeah, I shipped it out earlier than expected, because the two issues I had on my "to check" list ended up being impossible to reproduce on my side.

Thanks you for testing the dual wan support I must say I'm not entirely surprised that error handling in load balancing mode is quite limited. That's usually the trickiest part to handle, and should usually be handled at the app level rather than at the kernel level. Some systems use a heartbeat monitoring system and will remove dead routes as they are shown to be down. I'm not sure how Asus handles it as I never looked at the details, but since this is a feature that's still in development (AFAIK), I'm sure it will improve over the next releases.

I forgot to mention that I also enabled power control on the RT-N16 build. Can someone confirm whether it's working or not? There should be new entries on the Wireless Professional page to handle this.

 

[mstombs]

Re TX power - the web page works, changing power seems to require a soft reboot and the new value sticks in the web page BUT does not seem to have any effect on actual power, or in wl vars such as those seen with "wl curpower".

What does work from the console, immediately without a reboot is "wl txpwr"

txpwr Set tx power in milliwatts. Range [1, 84].

txpwr1 Set tx power in in various units. Choose one of (default: dbm):
-d dbm units
-q quarter dbm units
-m milliwatt units
Can be combined with:
-o turn on override to disable regulatory and other limitations
Use wl txpwr -1 to restore defaults

I can see the effect clearly on android phone with "Wifi Analyzer"

admin@RT-N16:/tmp/home/root# wl txpwr 1
admin@RT-N16:/tmp/home/root# wl txpwr
0.0 dBm = 1 mw.
admin@RT-N16:/tmp/home/root# wl txpwr 80
admin@RT-N16:/tmp/home/root# wl txpwr
19.0 dBm = 79 mw.
admin@RT-N16:/tmp/home/root# wl curpower
Power Control: On, HW
Current channel: (null)
BSS channel: (null)
BSS Local Max: 20.0 dBm
BSS Local Constraint: 0.0 dB
User Target: 19.0 dBm
SROM antgain: 2G: 2.0 dB, 5G: 0.0 dB

Regulatory Limits:
CCK : 17.0
CCK CDD 1x2 : 0.0
CCK CDD 1x3 : 0.0
Legacy OFDM 20MHz SISO : 17.0
Legacy OFDM 20MHz CDD : 17.0
MCS 0-7 20MHz SISO : 16.0
MCS 0-7 20MHz CDD : 13.0
MCS 0-7 20MHz STBC : 13.0
MCS 8-15 20MHz SDM : 16.0
Legacy OFDM 40MHz SISO : 16.0
Legacy OFDM 40MHz CDD : 13.0
MCS 0-7 40MHz SISO : 16.0
MCS 0-7 40MHz CDD : 13.0
MCS 0-7 40MHz STBC : 13.0
MCS 8-15 40MHz SDM : 16.0
MCS 32 : 16.0

Board Limits:
CCK : 25.50
CCK CDD 1x2 : 0.0
CCK CDD 1x3 : 0.0
Legacy OFDM 20MHz SISO : 22.50
Legacy OFDM 20MHz CDD : 22.0
MCS 0-7 20MHz SISO : 22.50
MCS 0-7 20MHz CDD : 22.0
MCS 0-7 20MHz STBC : 22.0
MCS 8-15 20MHz SDM : 20.50
Legacy OFDM 40MHz SISO : 22.0
Legacy OFDM 40MHz CDD : 22.0
MCS 0-7 40MHz SISO : 22.0
MCS 0-7 40MHz CDD : 22.0
MCS 0-7 40MHz STBC : 22.0
MCS 8-15 40MHz SDM : 20.50
MCS 32 : 22.0

Power Target:
CCK : 15.50
CCK CDD 1x2 : 0.0
CCK CDD 1x3 : 0.0
Legacy OFDM 20MHz SISO : 14.50
Legacy OFDM 20MHz CDD : 11.50
MCS 0-7 20MHz SISO : 14.50
MCS 0-7 20MHz CDD : 11.50
MCS 0-7 20MHz STBC : 11.50
MCS 8-15 20MHz SDM : 14.50
Legacy OFDM 40MHz SISO : 0.0
Legacy OFDM 40MHz CDD : 0.0
MCS 0-7 40MHz SISO : 0.0
MCS 0-7 40MHz CDD : 0.0
MCS 0-7 40MHz STBC : 0.0
MCS 8-15 40MHz SDM : 0.0
MCS 32 : 0.0

Maximum Power Target among all rates: 15.50 15.50
Rate index with Maximum Power Target: 0 0
Last adjusted est. power : 16.0 0.0

So I'm sure it can be fixed. By the way I'd prefer use of dB, not mW, tx power effect is better expressed logarithmically, to make a significant change you have to change 3dB which is double the mW!

 

[RMerlin]

Thanks for the feedback.

Asus has an odd way of handling power output. For some reason I don't know, they convert the requested mw output into quarter db using a chart, and apply the new value as a quarter db instead of a straight mw. Not sure why, could be to make the code more device-agnostic perhaps?

In any case, changing this might trigger quite a few other issues, so I'd rather leave that code alone, at least for the time being.