OpenWrt Forum Archive

Topic: Update on Linksys WRT1900AC support

The content of this topic has been archived between 16 Sep 2014 and 7 May 2018. Unfortunately there are posts – most likely complete pages – missing.

guitarman wrote:

Hi nyt...when you've got your aa build fixed could you kindly post a link to the *.img.
Also, does the openvpn feature appear to work ok?
Thanks

I'm running OpenVPN successfully on BB (latest image). Works great.

guitarman wrote:

Hi nyt...when you've got your aa build fixed could you kindly post a link to the *.img.
Also, does the openvpn feature appear to work ok?
Thanks

works fine, using it now.  My image will not be suitable for all, I do not use luci.  I'll post the source tree and the .config i use so others can build.

Has anyone been able to get vlans working on eth0 switch? I might be doing something incorrectly (first time attempting this), but no matter what I tried I could not get it to work.

Guides I referenced include:

http://wiki.openwrt.org/doc/uci/network/switch
http://wiki.openwrt.org/doc/techref/swconfig

I saw this patch, but not sure if it's relevant:

http://patchwork.openwrt.org/patch/5367/

I could get vlans (eth0.1, eth0.2, which ports for which, tagged/untagged) to be recognized, but not the switch itself.

The laptop I used for testing didn't even register a cable being plugged in with those configurations.

If anyone has gotten this working, I would very much like to hear details of your config.

building AA, as well as using the aa image they pushed, qos is broken =[

tc qdisc add dev eth1 parent 1:10 handle 100: fq_codel
generates error

root@ZOMGWTFBBQWIFI:/etc/config# /etc/init.d/qos start
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory

Still.... latest AA, everything working except qos...  There's no fq_codel patch in the 3.2 kernel.... works with sfq... but codel is much more efficient... still working on it...

ATTITUDE ADJUSTMENT (Attitude Adjustment, r40835)

root@ZOMGWTFBBQWIFI:/etc/config# uname -a
Linux ZOMGWTFBBQWIFI 3.2.36 #2 SMP Thu May 29 03:43:02 EDT 2014 armv7l GNU/Linux

root@ZOMGWTFBBQWIFI:/etc/config# /etc/init.d/qos start
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory


I modified qos-scripts to use sfq instead for this... this seems to be the best functionality I'm going to get until we get some better BB support.. which worked great minus the wifi driver.  I'm at least happy now until then... maybe I'll put this router back together again =]

(Last edited by nyt on 29 May 2014, 09:46)

Ok, if you want to build latest AA... you can try cloning this..  I left my .config in there.

https://github.com/notnyt/wrt1900ac-aa

this has working qos and stable wifi

After installing these images from serial console, I have to clear the ubi overlay.


root@(none):/# ubiformat /dev/mtd9
ubiformat: mtd9 (nand), size 39845888 bytes (38.0 MiB), 304 eraseblocks of 131072 bytes (128.0 KiB), min. I/O size 2048 bytes
libscan: scanning eraseblock 303 -- 100 % complete
ubiformat: 296 eraseblocks have valid erase counter, mean value is 2
ubiformat: 8 bad eraseblocks found, numbers: 296, 297, 298, 299, 300, 301, 302, 303
ubiformat: formatting eraseblock 303 -- 100 % complete

(Last edited by nyt on 29 May 2014, 10:58)

jklap wrote:

So here is what I've been seeing after running it solid for over a week.

The disconnects don't seem to be directly tied to running out of memory-- I've been running with with tweaked memory settings and swap and they are still occurring.

I do see a lot of cpu stalls-- ALL of them showing stack traces related to network ops-- often Sys_sendto but not always.

I've also seen several kernel panics-- hadn't turned on minicom logging and so have missed the full details before it would reboot itself.

The migration/0 process is consuming a massive amount of CPU time-- migration/1 almost nothing.

At times the route's load will jump to 4-5 when nothing seems to be going on-- feels like it's due to the migration/0 process. This often happens around the time of disconnects-- certainly see cpu stalls during this time.

Several of my machines are using the WRT1900AC as their router-- caveat that any machine that needs a solid wifi is connecting to an Airport Express smile

Other devices such as phones are using the WRT1900AC's wifi and rarely see issues... but they are likely 1) always re-connecting and 2) not moving huge amounts of data.

I am still running the factory software, I will occasionally see for no reason the connection will just "crash" and even causes the laptops driver to "crash"... so this explains a lot! thanks for all your hard work.

DaedalusX wrote:

Has anyone been able to get vlans working on eth0 switch? I might be doing something incorrectly (first time attempting this), but no matter what I tried I could not get it to work.

I'd been trying to figure out the same... even spent an hour one night trying to make sense of their code and plugging different values into /sys/devices/platform/neta/switch/reg_w.

As I really only want to filter on wifi traffic, until they document the switch config I've started looking into ebtables.

nyt wrote:

Ok, if you want to build latest AA... you can try cloning this..  I left my .config in there.

https://github.com/notnyt/wrt1900ac-aa

this has working qos and stable wifi

After installing these images from serial console, I have to clear the ubi overlay.


root@(none):/# ubiformat /dev/mtd9
ubiformat: mtd9 (nand), size 39845888 bytes (38.0 MiB), 304 eraseblocks of 131072 bytes (128.0 KiB), min. I/O size 2048 bytes
libscan: scanning eraseblock 303 -- 100 % complete
ubiformat: 296 eraseblocks have valid erase counter, mean value is 2
ubiformat: 8 bad eraseblocks found, numbers: 296, 297, 298, 299, 300, 301, 302, 303
ubiformat: formatting eraseblock 303 -- 100 % complete

That's awesome, thanks man! I'll give this a try this weekend.

sysupgrade seems to work, just need to relink /etc/fw_env.config to /tmp/fw_env.config which does not seem to be building proper.. will look into it more soon..

edit: sysupgrade fw_env stuff sorted in latest stuff, but something still seems amiss with it.. at least it doesn't blow out the ubifs now.  I don't like the double partition stuff...

all seems well with the latest build, pushed updates to github... no more 6scripts breaking firewalls...

https://github.com/notnyt/wrt1900ac-aa should be good to roll once you enable what you need.

(Last edited by nyt on 30 May 2014, 02:00)

Now that I have a proper build I think I'm just going to let this thing run until we get proper bb support.  It's quite fast.

(Last edited by nyt on 30 May 2014, 04:02)

Thanks nyt.  BTW how long does it take to build this image? I started the make process at the ubuntu command prompt over 2 hours ago and its still compiling and building?

guitarman wrote:

Thanks nyt.  BTW how long does it take to build this image? I started the make process at the ubuntu command prompt over 2 hours ago and its still compiling and building?

Depends on the system... make -jX where X is number of cores will speed it up.... compiling the toolchain and then cross compiling openwrt is not a quick  thing.

Ok buddy..I'm a virgin at this build stuff but so far so good. I appreciate all your efforts on this project to date smile

TBH, I don't really understand the point of all this until they're supporting us better on the driver side...
There's only a tiny clique of people showing interest, & that's largely because of how they've cooperated (i.e. not) with the community.
Retarded they are....

(Last edited by jalyst on 31 May 2014, 02:21)

jalyst wrote:

TBH, I don't really understand the point of all this until they're supporting us better on the driver side...
There's only a tiny clique of people showing interest, & that's largely because of how they've cooperated (i.e. not) with the community.
Retarded they are....

The point is to have a working image running on this router for those who have purchased it.

Belkin has sent routers to the openwrt devs and submitted patches, though not so good, to the mailing list.  They are working, so saying they're not is really misleading.  Also, you should see some good news in 2-3 weeks, so sit tight.

I'm not saying they're not working, the criticism is focused on the way they've gone about interfacing with the community.
It's mostly been very inept, so far...
I really do hope that pattern of behaviour is changing, but their actions so far have made it hard to believe that, lets see.

nyt wrote:

Also, you should see some good news in 2-3 weeks, so sit tight.

Now this is just cruel, if you know something is about to fundamentally change then why not share!?! ;-P

(Last edited by jalyst on 31 May 2014, 04:20)

Hi nyt - the "make" appears to have gone fine, took a few hours smile  it's created a 5.2GB directory in total. please advise what I do now to get an image into the router etc.  I have the serial hardware etc.  Cheers!

guitarman wrote:

Hi nyt - the "make" appears to have gone fine, took a few hours smile  it's created a 5.2GB directory in total. please advise what I do now to get an image into the router etc.  I have the serial hardware etc.  Cheers!

There should be an image in the bin/armadaxp directory...

openwrt-armadaxp--jffs2-128k.img is what you want

instructions to flash have been posted in this thread

guitarman wrote:

Hi nyt - the "make" appears to have gone fine, took a few hours smile  it's created a 5.2GB directory in total. please advise what I do now to get an image into the router etc.  I have the serial hardware etc.  Cheers!

How to update from Belkin/Linksys UI
  Login to WRT1900AC local UI
  Navigate to the Connectivity tab
  Select Manual firmware update
  Select image to load (e.g., openwrt-armadaxp--jffs2-128k.img)
  Select Update firmware
  After the firmware is updated, the unit will reboot, and the default ip address will be 192.168.200.1
  The default SSID's will be MAMBA_2G4, and MAMBA_5G2.
  The default username for OpwnWrt is 'root', and there is no default password set. 
  The unit will ask you to set a default password after you login to the UI.
 
 
How to ipdate from OpenWrt to Belkin/Linksys
  Login to the WRT1900AC web UI
  Select the 'System' Tab, and then 'Backup / Flash Firmware' tab
  In the 'Flash new firmware image' section click the 'choose file' button and select your firmware
  Click 'flash image'

gufus wrote:

https://github.com/wrt1900ac/opensource

Are they using?

BARRIER BREAKER (Bleeding Edge, r39624)

... it says right on that page

OpenWRT svn base revision: 40006

jalyst wrote:

I'm not saying they're not working, the criticism is focused on the way they've gone about interfacing with the community.
It's mostly been very inept, so far...
I really do hope that pattern of behaviour is changing, but their actions so far have made it hard to believe that, lets see.

nyt wrote:

Also, you should see some good news in 2-3 weeks, so sit tight.

Now this is just cruel, if you know something is about to fundamentally change then why not share!?! ;-P

propaganda ?

@gufus Just FYI, that Veriksystem build is older than the current Belkin release. Commit numbers are sequential, bigger number = newer base code.

Also their git repo doesn't exist/is private.

Sorry, posts 326 to 325 are missing from our archive.