Optimized build for the D-Link DIR-860L

Thanks to your tests I tried to restore the original antennas and the problem disappeared.
Do not install poor external antennas!

Good to hear that the router is working well for you :slight_smile: $18,- is a steal for such a device and if you want to test a master branch build, I just uploaded a new one.

Good to hear that you have sorted your problems. To bad it is the antenna mod. I was just looking into doing a antenna mod as a fun little project. Which antenna's did you use?

@All, if you update, please check the build number after flashing! There was a bug in OpenWRT which prevented the system updating (hostapd did not get killed). If you run into this problem either flash using bootloader recovery or flash over ssh terminal and first kill both hostapd instances before running the flashing command (eg. kill xxxx xxxx && sysupgrade /path/to/image (I would use /tmp)).

@Bartvz I'm running r6796 is r7161 safe to flash in terms of the mentioned bug

These but I do not recommend it.
Instead these are working very well and can be installed inside the router.

1 Like

I believe everything below 7xxx should be unaffected but I wanted to mention it. So if it happens people are not confused what is going on.

So some higher dBm antenna's then stock. Good to know :slight_smile: I was thinking about some external antenna and make a nice routing "totem pole". Internal antenna is cleaner though...

@All, @Mushoz this commit in NBD's staging tree hopefully will drop soon in the master branch. Considerably reduces CPU utilization for routing, yum!

I'll try to install these on in my second Dir-860L and then compare them vs the stock antennas, a bit more range wouldn't be bad... :wink:

R7161 is working just fine in my 860L, uptime over 2 days now

I got a second-hand DIR-860L from Amazon recently, and found it to be A1 version, not the B1 the thread owner offers. Is there an optimized build for A1, and more importantly, does A1 suppports HW offloading?

A1 isn't supported by OpenWRT/Lede unfortunately.

Completely different SoC.

Different SoC with closed source drivers. So if someone doesn't take op the work to work on an alternative driver it is not going to change soon unfortunately :confused:

@All, NBD dropped some flow offload patches for connection handling in the master branch. So, if my new build behaves, there is a new build tomorrow!

2 Likes

yay, cant wait! =)

Ye me too !!!
My productive unit is running over 7 days now without any problem @LEDE Reboot 17.01-SNAPSHOT r3909.
Hopefully your next build will be as stable as the one which i'm using right now, but i'm quite confident. :slight_smile:

And latest build loses internet connectivity after a while using flow offload and hw flow offload :confused:
However, NBD, dropped a fifth commit considering flow offload so my server is building once again.

So, it should be fixed in the new build!

2 Likes

@Bartvz do you know if the problem with the update has been solved?
Thanks

No problems running my new build (r7188) with flow offload and hw flow offload for around ±16 hours with ± 10 devices connected. No buildup of a huge amount of connections so that seems to be fixed. However, keep in mind that flow offload is a new beast so it is bound to have some bugs not worked out yet.

@All, new build (r7188) available in the OP!

2 Likes

@All, new build tonight because of a new version of mt76 driver and the commit I listed earlier which should improve routing throughput dropped in the master branch.

2 Likes

Thanks @Bartvz :slight_smile:

<3 @Bartvz

And r7274 is up, enjoy!

Just noticed this so a new build tomorrow. Good times to own a mt7621 device!

3 Likes