Do you have any suggestions how to replace a chipset, soldered onto the motherboard?

huh? There are 2x minipcie ports inside.

Basically everything is possible, it's like a PC, even put a graphics card.

Can say but AC v1 with r3901 is stable for 4 hours now. More stable then the last release if i talk about wifi.

Lets see if it stays on without rebooting.

(Last edited by pat30724 on 11 Apr 2017, 06:16)

I know. Im asking if anyone tried that, because I dont own any card for replacement and Im not sure if boot loader doesnt have some whitelist, etc. I would like to not lose money for card that wont be working.

Hello David.  First off thank you for the nice builds.  I have a wrt3200 and just installed your latest build.  The 2.4 wifi seems to work pretty well but i have to mess around with the 5ghz to get it to work.  Also why is there two 5 ghz radios?   Ive never had a 3200.  I have a couple of 1900's but never had these issues or extra radios. Could you tell me the best way to configure these radios?

config wifi-device 'radio0'
    option type 'mac80211'
    option hwmode '11a'
    option path 'soc/soc:pcie-controller/pci0000:00/0000:00:01.0/0000:01:00.0'
    option country 'US'
    option htmode 'VHT40'
    option channel 'auto'

config wifi-iface 'default_radio0'
    option device 'radio0'
    option network 'lan'
    option mode 'ap'
    option encryption 'psk2'
    option key '12341234'
    option ssid 'M&D '
    option disabled '1'

config wifi-device 'radio1'
    option type 'mac80211'
    option hwmode '11g'
    option path 'soc/soc:pcie-controller/pci0000:00/0000:00:02.0/0000:02:00.0'
    option country 'US'
    option channel '6'
    option htmode 'HT40'

config wifi-iface 'default_radio1'
    option device 'radio1'
    option network 'lan'
    option mode 'ap'
    option encryption 'psk2'
    option key '12341234'
    option ssid 'M&D '

config wifi-device 'radio2'
    option type 'mac80211'
    option hwmode '11a'
    option path 'platform/soc/soc:internal-regs/f10d8000.sdhci/mmc_host/mmc0/mmc0:0001/mmc0:0001:1'
    option htmode 'VHT80'
    option country 'US'
    option channel '36'

config wifi-iface 'default_radio2'
    option device 'radio2'
    option network 'lan'
    option mode 'ap'
    option ssid 'M&D 5'
    option encryption 'psk2'
    option key '12341234'

@belliash

As proof of concept I have replaced the wifi cards on Mamba with ath9k ones and it works. Too bad we can't do the same on Rango, as wifi chips are soldered onto the board. But this already offtopic here. Open a new thread please.

nitroshift

You are right, I opened new thread.

Just had a reboot 5 minutes ago after being up for almost 16 hours.

I think it should just require proper kernel modules.

Uptime of almost 24 hours now. Wifi seems more stable then the official release.

(Last edited by pat30724 on 12 Apr 2017, 06:07)

@davidc502 does this commit mean you won't be able to build 4.4?

https://github.com/lede-project/source/ … a21e6f5921

Good find.  Looks that way. They better have the new kernel fixed on v1 or there's gonna be a lot of pissed people

*EDIT*

I've asked Felix if V1 has been fixed with the new kernel.

(Last edited by davidc502 on 13 Apr 2017, 00:07)

Downloaded new sources, and can confirm 4.4 is no longer there.

Compiling 4.9, and will post the images for everyone to test. Hoping we have some V1 users to confirm the problem is either fixed or not.

I'm happy to try it again, its been a week or so since I did last.

Are we talking about the wifi problem on 3200acm? Sorry I'm new to these forums and I'm trying to follow. Just started to work with OpenWRT/LEDE so I'm still trying to understand the flow of conversation

The V1 problem they are talking about is with the wrt1900ac v1 rebooting for no apparent reason with david's latest posted build.

There are also wireless issues with the wrt3200acm which we are waiting for a new driver from Marvell to address the issues. Hopefully released soon. Not sure why, but my wrt3200acm has been most stable on wireless with dd-wrt Brainsylayer r31971. Uses the same drivers, but is using an older 3.18 kernel.

Before you ask, David, yes I tried your beta that you posted a link to. Failed on 5GHz wireless connections after about 12 hours of up time.

Kernel 4.9 has been deployed to the server.  Here is the link, and download the appropriate image for the hardware.

https://davidc502sis.dynamic-dns.net/sn … u/generic/

I'm interested to hear from V1 users as to the reboot problem on this kernel.

3200acm Wifi drivers are still going to be wonky as the driver has not been changed/updated yet. We are still waiting on a fix.

(Last edited by davidc502 on 13 Apr 2017, 17:49)

Is it confirmed that we don't need to do uart recovery for WRT3200ACM build anylonger? Quite eager to replace the WRT1900ACS as main router. I can live with wonky 5GHz for a while.

Happy Easter!!! (for those lucky on holiday)

Thanks for that I just started last week with my project of getting rid of my ISP-provided modem (Bell Fibe in Canada, FTTH) and I'm almost 100% complete. Just waiting on that wifi bug to be fixed so I can switch to my own router permanently

Those using the test image ----   Please manually input the distribution feeds. Currently, it is pointed to lede trunk snapshots, but after a few days will stop working due to mismatching.

For some new folks it may not be clear, so just post in the thread, and I'll be happy to help.

Thanks,

I got an uptime of around 1h 35m before reboot this time. The longest yet I think.

Same thing as usual, eSATA light on bright + all the other lights freeze, then reboot.

No email responses yet. I've also posted a question on the LEDE board.

https://forum.lede-project.org/t/mvebu- … h-4-9/3046

On my AC v1, I used the old 4.9 snapshot that was in the snapshot rebooted yesterday afternoon at 5:30 (GMT+2) after being up for 2 days and 10 hours.

Flashed the new one you posted before.

With the new one its online for 6 hours now.

(Last edited by pat30724 on 14 Apr 2017, 12:30)

I appreciate the testing of V1 on this kernel<  Good information.

Though no direct answers, the distinct feeling is that the 1900ac V1 will be just left behind, and unknown reason why.  The only way around would be to fix the bug, or have someone compile a image specifically for the V1 on kernel 4.4.