[alpha] Early preview builds for Asuswrt-Merlin 380.66

[Blinkyz]

Minimal layout problems with Firefox 52.

Firmware 380.66 alpha 3 works fine with 87U.

https://www.snbforums.com/threads/a...uswrt-merlin-380-66.37920/page-10#post-316790

 

[Gregory Phillips]

I'm going to retry the alpha 3 one more time so I just reloaded it into the router.

 

[Gregory Phillips]

Okay I've been back on alpha 3 since yesterday afternoon and I had to manually unplug the primary wan to force the secondary to get a lease. When I plugged the primary back in the secondary went to standby mode. Why won't it obtain a lease and go to standby on its own? 380.65_2 would do that by itself overnight. Otherwise I don't seem to be having any problems with it and my radios seem to have a constant output as measured by Wi-Fi analyzer on my cell phone.

 

[bradbort]

Anyone running this alpha3 on their 88u and if so are you having any issues with it?

I am. All is fine, as long as I keep airtime fairness disabled.

 

[Gregory Phillips]

I am. All is fine, as long as I keep airtime fairness disabled.

I don't use airtime fairness and let my devices decide for themselves when they need to change connection types. Alpha 3 seems to be doing fine for me with no strange results to any wired/wireless clients.

On a side note asus tech support now wants my Comcast username and password so they can investigate why the router will not obtain an IPv6 connection using their firmware. Since that ain't gonna happen I've snipped every page off the cable modem for them along with the syslogs in my reply. I told them they should implement Merlin's IPv6 method which works great for me on Comcast's dual stack here in Jax.

 

[RMerlin]

I told them they should implement Merlin's IPv6 method which works great for me on Comcast's dual stack here in Jax.

Comcast's IPv6 support is flat out quirky, and reliability actually varies based on which node you are connected to. There are Comcast users who are also unable to get a stable IPv6 connection with my firmware, so there's no constant there.

My IPv6 implementation is exactly the same as Asus, with maybe just a few minor fixes or changes here and there. But the core of it remains unchanged, as I have no way of testing it anyway.

 

[Gregory Phillips]

Well then brother your tweaks are the right ones for my area because it works for me. You have many settings available to the user that are not available in the stock firmware that are prefix designation specific. I've had other Comcast users in different areas tell me that they have to set their routers to prefix designation specifically in order to get an IPv6 connection.

 

[RMerlin]

To people with wifi stability issues, this is why I recommend they first start by looking at their environment before the firmware:

The phone and Nexus 7 have been connected for 17 hours (since I got back home with them), and the Nexus 9 that always remains at home hasn't disconnected in 65 hours.

 

[BlackMage]

what is the difference at "NAT Firewall" between "Deactivate", "Asus" and "Merlin" (in 'Firewall General')? any performance(or technical) issues and what actually happens with the packages that need to be forwarded when it is set to 'Deactivated'?

 

[stambeccuccio]

what is the difference at "NAT Firewall" between "Deactivate", "Asus" and "Merlin" (in 'Firewall General')? any performance(or technical) issues and what actually happens with the packages that need to be forwarded when it is set to 'Deactivated'?

https://www.snbforums.com/threads/asus-nat-loopback-and-merlin-nat-loopback-differences.33310/

 

[Naftali Oziel]

Since I finally took care of the most important issues with 7266, I decided to upload some early builds in the Test Builds folder on Mediafire. As usual, these are provided without any support, and are work-in-progress.

Some major changes coming to my professional life in the coming weeks, so I don't know how much development time I will be able to put into this in the coming weeks. So far, I haven't planned anything regarding 380.66, so this is just a GPL merge (was nice for once to have ALL the binary blobs for all supported models), and the usual random fix and tweaks here and there.

bd7fdd4 Bumped revision to alpha 2
97cb3bd Updated documentation
2292033 webui: cleaned up indentation on Tools->Other Settings page (no code change)
f3c7469 webui: add setting to disable wanduck's DNS probes
d174cd4 Merge pull request #1259 from blackfuel/upstream
fe27c86 Provide separate binary blobs and wireless driver for AC3100 as they're not entirely identical to the AC88 ones
3b22668 Merged binary blobs for RT-AC88U from 7266 GPL
dcd47a1 wl: revert to older version for AC88U, the AC3100 version from 7266 causes AdaptiveQoS to stop marking packets after a while
b9c3079 samba: add gitignore to Samba 3.0.33
5a2dd53 openvpn: add support to the ovpn importer for the "proto" keyword; extend the "remote" keyword to also support the proto parameter if present
6f125de Merge remote-tracking branch 'upstream/master' into upstream
8f8fcf9 shared: Add the physical notify_rc.o binary blob, referenced through symlinks
56b9ed7 Updated documentation
3f86a28 firewall: log accepted new outbound connections (if logging is enabled)
ec59c40 curl: fix warning at build time, moving flags to CPPFLAGS and LIBS
0acc483 rc: Revert "rc: Kludge to support GPL 3831 with binary blobs from 32xx-34xx" (no longer required)
ac1fdea Merge remote-tracking branch 'upstream/master' into upstream
79d5a27 webui: escape quotes in client names when building clickable client list, fixes DHCP page (and possibly others)
c4277cc webui: add missing variable to state.js, fixing the Apps page
8769145 webui: fix broken WOL page if client name contains single quotes
d5c1104 rc: move Merlin-specific nvram conversion/upgrade code to the new format.c source, and run it at the same time as Asus's own version conversion code
1cddc24 rc,shared: Add new prebuilt binary blobs
949c67a build: add new closed source blobs for AiCloud's sync clients
5a46c52 Tor firewall: use new Asus function addr_type_parse()
fc95554 Updated documentation
05b9ace Revert "kernel: ipv6: honor IPV6_PKTINFO with v4 mapped addresses on sendmsg (backported c8e6ad0829a723a74cd2fea9996a3392d2579a18)"
ac3773a kernel: ipv6: honor IPV6_PKTINFO with v4 mapped addresses on sendmsg (backported c8e6ad0829a723a74cd2fea9996a3392d2579a18)
d71498d openvpn: allow disable DH usage by entering none in the DH field
cced648 kernel: fix reuse-after-free in DCCP (backport from 5edabca9d4cff7f1f2b68f0bac55ef99d9798ba4)
9e665fc rc: merge binary blobs from RT-AC3200
bf26819 shared: add missing symlink to notify_rc.o for RT-AC66U
761f9aa build: rename IPSEC option to IPSECMOD, to avoid clashing with Asus's own IPSEC option
09a4791 openssl now requires makedepend to be installed - document it
5b65381 qos: Fix GPL 7266 breakage of Bandwidth Limiter (patch from Asus)
961f278 httpd: fix potential buffer overrun introduced in previous commit, as src buffer was too small by 3 bytes
55a7fd0 webui: clarify the label of the new source field on the port forward page
4718100 webui: support user-defined source IP on the Forwarded Port list
8988177 openssl: Fix building, following GPL 7266's removal of rc4
1b567d7 Merge with GPL 7266
50ad9fc Updated documentation
2f4be88 Bumped revision to 380.66 alpha 1
251a846 Re-apply GPL 4180

So am contemplating in upgrading my AC68U AP to this alpha version. With that in mind, I understand that I may need to disable airtime fairness. Couple of questions w/no airtime fairness on 2.4

1. If am only using all N devices, is there a purpose for airtime fairness? thought I have left my wireless mode to Auto (found performance to be better), not sure why?
2. Will this impede on my network if everyone is streaming (netflix, facetime, sykype etc.) at the same time?

I ask as I found my older N devices respond better without the airtime fairness enabled, I do not encounter any wireless disconnects. I do notice when running the speedtest that my bandwidth testing starts very low and stays that way for awhile until I browse. If I retest my speedtest results are much higher going from 25Mbps to 90+Mbps. Without airtime fairness speedtest results are consistently higher all the time.

Your thoughts?

BTW I also have my beamforming disabled both universal and explicit on 2.4 since I use apple devices. Any thoughts around this?

 

[bradbort]

So am contemplating in upgrading my AC68U AP to this alpha version. With that in mind, I understand that I may need to disable airtime fairness. Couple of questions w/no airtime fairness on 2.4

1. If am only using all N devices, is there a purpose for airtime fairness? thought I have left my wireless mode to Auto (found performance to be better), not sure why?
2. Will this impede on my network if everyone is streaming (netflix, facetime, sykype etc.) at the same time?

I ask as I found my older N devices respond better without the airtime fairness enabled, I do not encounter any wireless disconnects. I do notice when running the speedtest that my bandwidth testing starts very low and stays that way for awhile until I browse. If I retest my speedtest results are much higher going from 25Mbps to 90+Mbps. Without airtime fairness speedtest results are consistently higher all the time.

Your thoughts?

BTW I also have my beamforming disabled both universal and explicit on 2.4 since I use apple devices. Any thoughts around this?

I have a mixed environmnet, that includes apple, N, etc. I leave airtime fairness off, and have seen no streaming issues with it off. I actually leave beamforming on (implicit and explicity)

Brad

 

[Naftali Oziel]

I have a mixed environmnet, that includes apple, N, etc. I leave airtime fairness off, and have seen no streaming issues with it off. I actually leave beamforming on (implicit and explicity)

Brad

Thanks Brad, appreciate the feedback. Do you find any difference with beamforming 2.4? I've heard so many issues around since there was no true standard around it as it was designed for AC. So I left it off.

 

[bradbort]

Thanks Brad, appreciate the feedback. Do you find any difference with beamforming 2.4? I've heard so many issues around since there was no true standard around it as it was designed for AC. So I left it off.

Minimal. But I have not done rigorous testing.

 

[eddiez]

Just a quick observation: WiFi calling (Vodafone NL) seems to be working again with 380.66a3.

Verstuurd vanaf mijn SM-G935F met Tapatalk

 

[Gregory Phillips]

I started having some very strange issues with my RT-AC88U using 380.66_alpha3 after two days uptime that finally culminated on day 3 with the router not issuing IPv6 addresses to wired clients. It started with some strange spikes on the wireless network which had airtime fairness enabled by default. Then wired clients started slowing down and eventually stopped receiving IPv6 addresses.

I just rolled back to 380.65_2 because it worked so well for me before and I never detected any anomalies while using it. I will continue to monitor the behavior of this router and if I see anything else I'll probably just RMA the unit to rule out hardware issues.

Edit: After flashing back the router failed to issue IPv6 addresses to clients even though it had established an IPv6 connection with the modem. I had to power cycle it in order for it to begin issuing IPv6 addresses. Another issue I witnessed before the flashback was some severe signal spikes on the 2.4ghz band using acrylic Wi-Fi on my tablet. The signal to that tablet's Broadcom radio fell down to the almost gone range of the spectrum.

I failed to break out my laptop with an intel ac 7260 Wi-Fi adaptor before the flashback but since I power cycled the router I also started using it to monitor the 2.4ghz signal and it shows a much flatter signal pattern in the very good range at the same distance and timeframes that the Broadcom unit reports severe spiking. I also reinstalled the driver to the tablet and am actively searching for an updated version since this one is the latest on the asus site for my transformer tam100.

On my laptop with the intel adaptor 2.4ghz has a saw tooth signal pattern while 5ghz is super flat with very little variance. On my transformer 2.4ghz is saw tooth with very large cyclical drops in the signal while 5ghz is very flat resembling the pattern shown on the intel adaptor. I've also noticed several drops with the intel adaptor that show clear loss of signal gaps

 

[MacG32]

380.66 alpha 3 is working very well and without any problems here.

 

[Deleted member 22229]

Same here installed on the release date and have had not one single drop or hiccup in any devices. AC3100

 

[Gregory Phillips]

The whole time that was going on my Samsung galaxy note 2 running Wi-Fi analyzer didn't show any loss of signal but it did reflect signal fluctuations.

 

[Deleted member 22229]

Wifi analyzer shows signal flux with all signals it picks up even all my neighbors always up and down. Not a good way to test. I check my laptop here that has a Netgear wifi adapter using( netsh wlan show interfaces ) And its always rock solid at 90% signal.