[Fork] 374.43 LTS Beta including new rev AC68U routers (V27BI)

[MarCoMLXXV]

Two remarks:

This build reboots the router when toggling QoS. Is it supposed to do that?

The logged packets type setting reverts to 'Dropped' after a reboot, when I've set it to 'None' (prior to rebooting).

 

[Adamm]

The logged packets type setting reverts to 'Dropped' after a reboot, when I've set it to 'None' (prior to rebooting).

This is because of Skynet. Because this specific setting changes the way the IPTables rules are implemented I force the dropped setting if set to none (it also allows you to set it to both).

 

[MarCoMLXXV]

This is because of Skynet. Because this specific setting changes the way the IPTables rules are implemented I force the dropped setting if set to none (it also allows you to set it to both).

Ah, I see. Thanks for clarifying. If it's done by Skynet, it's fine by me

 

[cybrnook]

This build reboots the router when toggling QoS. Is it supposed to do that?

That's because things like ctf are being enabled / disabled when turning qos on / off.

 

[jrmwvu04]

What do you want to change?

Why? I dont need and i dont want
Already stock at 1.4ghz cpu inside,and Wifi seem to be stable.
Just want an "Install and Forget " like my old ac68u(A1,800mhz cpu)

The answer, for those curious, is that you can.

 

[john9527]

This build reboots the router when toggling QoS. Is it supposed to do that?

As everyone has said, yes, if you are set to use CTF (HW Acceleration) as the default on the LAN>Switch Control page. If you disable QoS, then set CTF to disabled in the LAN/Switch Control dropdown, then the QoS On/Off won't reboot anymore.

 

[cvx01]

LOL. Yeah... I agree. Too many flavors.

Just as an FYI if helps , I've been running your LTS fork on this model since June 2016 when I picked it up at MicroCenter. (currently on 3.0.0.4.374.43_2-26E4j9527). All have been flawless without issue.

Looks like there was a review done on V3 variant. 1GHz out of box. OCed to 1.2GHz without issue.

https://www.smallnetbuilder.com/wir...s-ac1900-gigabit-router-reviewed?limitstart=0

I will roll the dice on this build later tonight and see how it goes.

Thanks again

Updated to 27BDSDKj9527 on AC68U V3 without issue.

Some NVRAM changes that some might find of interest...

The new driver seems to have added Airtime Fairness and Bluetooth Coexistence, which is expected based on John's original post noting these items. What was aunexpected is new "txpower" flags have been added. Details below

ATF appears to be related to = Airtime Fairness
BTC appears related to Bluetooth Coexistence

• (https://www.google.com/patents/US8265017)

_txpower - Looks like tadditional TX power flags have been added

• Existing set in mixed case (TxPower) remains
• A NEW set all in lower case (txpower) has been added

NVRAM variables ADDED between 2-26E4j9527 and 27BDSDKj9527

• wl0_atf=0
• wl0_atf_mode=0
• wl0_atf_ssid=
• wl0_atf_sta=
• wl0_btc_mode=0
• wl0_txpower=250
• wl1_atf=0
• wl1_atf_mode=0
• wl1_atf_ssid=
• wl1_atf_sta=
• wl1_btc_mode=0
• wl1_txpower=250
• wl_atf=0
• wl_atf_mode=0
• wl_atf_ssid=
• wl_atf_sta=
• wl_btc_mode=0
• wl_txpower=250

 

[john9527]

_txpower - Looks like tadditional TX power flags have been added

• Existing set in mixed case (TxPower) remains
• A NEW set all in lower case (txpower) has been added

Very observant! The LTS fork actually maintains both the 'old' TxPower settings and the 'new' txpower. Under the hood so to speak, it takes the old values entered in the gui and translates them into the new values required by the new wireless driver. It's set to still allow some increase in the power settings just like the original LTS release, but frankly it's been a bit tough for me to verify that is actually working, or I'm running into a 'lock down' in the driver.

 

[cvx01]

Very observant! The LTS fork actually maintains both the 'old' TxPower settings and the 'new' txpower. Under the hood so to speak, it takes the old values entered in the gui and translates them into the new values required by the new wireless driver. It's set to still allow some increase in the power settings just like the original LTS release, but frankly it's been a bit tough for me to verify that is actually working, or I'm running into a 'lock down' in the driver.

That's awesome thinking!!

If TxPower is supposed to update txpower automatically, I didn't see that after the update or even after reboot. txpower remained 250 regardless of what was listed in TxPower. Manually updating txpower took w/commit + reboot, however after reboot it reverts back to 250.

 

[john9527]

That's awesome thinking!!

If TxPower is supposed to update txpower automatically, I didn't see that after the update or even after reboot. txpower remained 250 regardless of what was listed in TxPower. Manually updating txpower took w/commit + reboot, however after reboot it reverts back to 250.

That shouldn't be.....can you append the output of
nvram show | grep -i power

 

[cybrnook]

If it helps. Only changes I have made are upping to 120 in the professional tab:

/tmp/home/root# nvram show | grep -i power
Ate_power_on_off_ver=2.4
wl_txpower=150
wl0_txpower=150
wl_TxPower=120
Ate_power_on_off_enable=0
wl0_TxPower=120
wl1_txpower=150
size: 42701 bytes (22835 left)
wl1_TxPower=120

 

[jrmwvu04]

Mine is a bit different.

admin@RT-AC1900P:/tmp/home/root# nvram show | grep -i power
Ate_power_on_off_ver=2.4
size: 43475 bytes (22061 left)
wl_txpower=250
wl0_txpower=250
wl_TxPower=200
Ate_power_on_off_enable=0
wl0_TxPower=200
wl1_txpower=250
wl1_TxPower=200

Edit: worth noting mine may differ because of my cfe edits

 

[john9527]

If it helps. Only changes I have made are upping to 120 in the professional tab:

Yep....that looks right. 120mw is 150% of the reference (80mw).....txpower = 150

Mine is a bit different.

Your is 'max'd' out, but consistent....I cap things at 200mw (250% of 80mw)

 

[jrmwvu04]

Your is 'max'd' out, but consistent....I cap things at 200mw (250% of 80mw)

That's in line with my intent. I have always run with the regular maximums imposed by your firmware, and really don't have any desire to mess with it further, as it performs at or above my expectations.

 

[cvx01]

That shouldn't be.....can you append the output of
nvram show | grep -i power

Here you go...

/tmp/home/root# nvram show | grep -i power
size: 45620 bytes (19916 left)
Ate_power_on_off_ver=2.4
wl_txpower=250
wl0_txpower=250
wl_TxPower=500
Ate_power_on_off_enable=0
wl0_TxPower=500
wl1_txpower=250
wl1_TxPower=500

 

[cybrnook]

Here you go...

/tmp/home/root# nvram show | grep -i power
size: 45620 bytes (19916 left)
Ate_power_on_off_ver=2.4
wl_txpower=250
wl0_txpower=250
wl_TxPower=500
Ate_power_on_off_enable=0
wl0_TxPower=500
wl1_txpower=250
wl1_TxPower=500

Your is 'max'd' out, but consistent....I cap things at 200mw (250% of 80mw)

 

[cvx01]

Yes, if applied via the GUI. All values sync when using the GUI (just tested this using max value), however if set via NVRAM they do not sync.

 

[john9527]

Yes, if applied via the GUI. All values sync when using the GUI (just tested this using max value), however if set via NVRAM they do not sync.

-TxPower is the master, and txpower is the slave (changing TxPower will change txpower, changing txpower will be reverted to the value calculated from TxPower)
- if TxPower (the old gui value) >= 200 (ma), then txpower will be capped at 250
- txpower is recalculated when you change the power in the gui or on a reboot
- if you change TxPower with nvram set/commit, you need to reboot for it to take effect (I'm going to make a change so it also gets recalculated on a restart_wireless)

 

[jrmwvu04]

On the 1900P(v3) as compared to the AC68W...

QoS is working as before, both anecdotally and a synthetic measurement.

UPnP with the Xbox Ones working well.

iPerf3 throughput and general wireless range is a shade better, but negligibly so.

(68W in pink, 1900P in orange)

entware-ng installed no problem, pixelserv-tls running no problem. The 1900P boots about 5 seconds faster, generally. Temperatures are basically the same. nvram usage is the same.

Pretty much coming up roses so far.

 

[tracker.ca]

I just got back from vacation to this fantastic news! I'm picking up a new router to do some tests this week.

thank you John!