@DavidMcWRT

Cheers, I didn't think to look in luci for a place to change the package lists, far simpler than I expected!

I'm always watching the timeline, McWRT git (THANKS CHADSTER!!) and this thread ever since Belkin's first submissions and the release of their own openwrt fork

Thank you for being so helpful to me and the rest of the community It's difficult being new to this, trying to stitch all this information together to form a complete picture. Your guide is very useful at this point in time, maybe the post could be linked in the wiki?

(Last edited by listerwrt on 13 Feb 2015, 04:39)

@tusc

Did you compile your own image? Because latest trunk image still uses kernel 3.18.6. Also, the fan_ctrl.sh has to be modified to read

/sys/devices/platform/pwm_fan/hwmon/hwmon0/pwm1

Note the insertion of /platform.

nitroshift

(Last edited by nitroshift on 13 Feb 2015, 10:00)

Can anyone tell me which would be considered the best firmware for now?
I have tried lifehacksback's and Kaloz's firmware(followed by DavidMcWRT's instruction, what I have done were first flash Kaloz's Jan 28th image, do sysupgrade to the snapshot and then install luci, done, no other packages or mods where installed) . Both can give me a fast wireless speed, however, unstable wireless connection. Since most of my wireless devices are from Apple, I was experiencing lock downs a lot.
Then I tried the linksys stock firmware. It is more stable, but less features, and MUCH slower wireless. I can only get 6 Mbps on both 2.4GHz and 5GHz wireless
network. Also, wireless network coverage is much weaker on the stock firmware. With OpenWRT, I can get 2 bars of signal in my bedroom, but none with the stock.
I still prefer OpenWRT. Can anyone tell me if there is something that I missed or did wrong?

@haoster

You didn't miss or do anything wrong. This is the actual situation with Apple devices as reported by different members in this thread. I can NOT talk for myself as I don't own any Apple devices (not a fan of them at all, I prefer Blackberry's).

nitroshift

(Last edited by nitroshift on 13 Feb 2015, 10:56)

Not to mention the wider population:
http://www.bing.com/news/search?q=apple+wifi+problems

@haoster
Your only other recourse is to try McWRT:
https://github.com/Chadster766/McWRT/re … 2-128k.img
(flash back to Linksys first; DON'T keep/load old settings)

However, it's old and unmaintained code, both by OpenWRT (being a fork of the AA release) and now Chadster himself.

(Last edited by DavidMcWRT on 13 Feb 2015, 13:34)

3.18.7 kernel now out.

The maintainer stated "All users of the 3.18 kernel series must upgrade."

http://lkml.iu.edu/hypermail/linux/kern … 01739.html

Yes

Right, and with 3.19 it's changed again. That does not exist in 3.19:

root@wrt1900ac:/# find /sys -name pwm?
/sys/firmware/devicetree/base/pwm_fan/pwms
root@wrt1900ac:/#

@tusc

Yes, it does exist. I'm running my self-built firmware based on 3.19 and that's the way I'm controlling the fan.

nitroshift

That's odd. I must be missing a definition in .config. Using your path it's not found.

root@wrt1900ac:~# ls -la /sys/devices/platform/pwm_fan/hwmon/hwmon0/pwm1
ls: /sys/devices/platform/pwm_fan/hwmon/hwmon0/pwm1: No such file or directory
root@wrt1900ac:~#

I suspect it will be updated here  https://downloads.openwrt.org/snapshots … u/generic/
Tomorrow?

so I moved my usb to a different port on the router and now I get fstab errors being not found.

did an ls on the /etc and get lrwxrwxrwx    1 root     root            10 Jan 28 12:35 fstab -> /tmp/fstab
so I ls the /tmp and there is no fstab

any thoughts that might be causing this?  on Kaloz Jan 28th image

@tusc

Make sure you don't have gpio fan module selected in your config.

nitroshift

@chadster766,
Thank you for all the hard work you put into the McWRT project. I hope the WRT community with the help of a few handy men here, continues to support this router. Thanks!

Nothing in the timeline committing changes to trunk, so I've emailed Luka.

If it gets patched today then it might show up in tomorrow's snapshot, else wait another 24 hours.

(Last edited by DavidMcWRT on 13 Feb 2015, 22:33)

Hmmm... I am running the trunk image from the  Feb 10th (r44324), but I was running the  Kaloz's Jan 28th image before this one and I had no problem.

Looking at my /etc directory I have the same link to /tmp but there isn't a fstab in /tmp

Here's a thought - Do you have block-mount installed? If not, try installing it and see if it helps.

I needed that for the Kaloz's Jan 28th image, but on the Feb10th's trunk - it gave me errors and would not load.

Is 3.18.6 broken? All my images experience kernel panic

Do you have a boot log?

I'll flash again and get a log

------------------------------------------------\

Also check out my videos (for those who want to flash through tty)

https://www.youtube.com/watch?v=kGRCg8t … 29vCbWvj7d

OK. If I understand your example correctly I should be able to tag the remaining port as well. (I grasp the vlan 'principles', just trying to get my head around the nitty gritty details ...)

The use case is that my internet provider provides an utp PPPoE connection with internet on vlan 6 (both IPv4 and IPv6) and TV on vlan 4 (currently not used by me). The current device is a TP-Link TL-WDR4900 v1 with port 0 being the CPU and port 1 WAN. Vlan 1 and 2 were the device defaults. I added 4 and 6 (and changed some tagging on 1 and 2) according to some local instructions found. And the wan interface uses eth0.6.

(Last edited by bvhoesel on 14 Feb 2015, 13:13)

Kaloz has literally just switched trunk to 3.18.7:
https://dev.openwrt.org/changeset/44443

so if you're still rolling your own pull and try that.

As for me, I'll wait for tomorrow's snapshot .

(Last edited by DavidMcWRT on 14 Feb 2015, 13:36)

Well, to further add confusion, I was actually confused about mixing tagged and untagged VLANs on the same port. I was thinking of mixing port-based and 802.1q VLANs, which still wouldn't work. But the config I provided before is not using port-based VLANs, so even if you mix tagged and non-tagged VLANs on the same port it will work just fine.

For example, this works fine (config is just an excerpt):

config interface 'lan'
        option ifname 'eth0'
        option force_link '1'
        option proto 'static'
        option ipaddr '192.168.1.23'
        option netmask '255.255.255.0'
        option ip6assign '60'

config interface 'lan2'
        option ifname 'eth0.4'
        option proto 'static'
        option ipaddr '192.168.4.1'
        option netmask '255.255.255.0'

config switch_vlan
        option device 'switch0'
        option vlan '1'
        option ports '0 1 2 3 5'

config switch_vlan
        option device 'switch0'
        option vlan '4'
        option ports '0t 5t'

My newest image 3.18.7 failed again. Something about VFS not mounting no root file system?

Tried Kaloz image... same problem

(Last edited by lifehacksback on 14 Feb 2015, 14:56)

2015-02-13 trunk image seem fail to mount root:

logs:
http://pastebin.com/0VpvC1MD

Thanks, kallam. It's (presumably) nothing to do with the kernel, but rather one of the mvebu-specific changes recently. It looks like the ubi image is generated with incorrect parameters at the moment:

[    1.029728] UBI: auto-attach mtd7
[    1.033154] UBI: attaching mtd7 to ubi0
[    1.037257] UBI error: validate_ec_hdr: bad VID header offset 512, expected 2048
[    1.044738] UBI error: validate_ec_hdr: bad EC header
[    1.049810] Erase counter header dump:
[    1.053634]  magic          0x55424923
[    1.057400]  version        1
[    1.060378]  ec             0
[    1.063428]  vid_hdr_offset 512
[    1.066587]  data_offset    2048
[    1.069828]  image_seq      494569088
[    1.073574]  hdr_crc        0x9dfa6aa8

What I don't know is why you're seeing this now when there haven't been any relevant changes in the past few days, and others here (I think) have tested the snapshot images since the changes went in.

(Last edited by leitec on 14 Feb 2015, 16:02)

Oh, hmm, might be the difference between sysupgrading (where the UBI data is kept intact, changing only the filesystem contained within) and the factory image.