Thanks!

Are the led lights working? for some this might not be necessary but it is the easiest way for me to find out if our ISP is down.

Thanks, it is working fine on my v1.1
Reboot is working, too.

I just upgraded to 17.01.1 and it's working fine.

The LEDs work all except for the wifi leds. Those are controlled via the ath10k rather than IO pins. No one has put together and submitted a patch to get them working (although some people have recently tested code that seems to at least activate them).

(Last edited by TeutonJon78 on 19 Apr 2017, 17:45)

Looks like the reboot fix for 1.1 should be in trunk now.

https://github.com/lede-project/source/ … 4f38786e45

thanks

Nevermind, Figured this one out.

(Last edited by weemo on 22 Apr 2017, 00:30)

Indeed, I compiled a version from the trunk (r4018) and reboot is working with v1.1 now

Also, the bug affecting the wireless I reported earlier using kernel 4.9 seems to be fixed as well. I did a few tests and no kernel/wireless errors.

Stable so far. I'll report issues, if any.

I spoke too soon, after a few more intensive tests, it crashed again.

I'll rebuild another one but with the CT firmware to see if it works. I never was able to make it work properly but it's worth a try... I'll let you know how it goes...

[ 1240.529816] ath10k_pci 0000:01:00.0: firmware crashed! (uuid b93a999a-3bc8-4f4a-9314-f0636cfcb243)
[ 1240.529853] ath10k_pci 0000:01:00.0: qca99x0 hw2.0 target 0x01000000 chip_id 0x003b01ff sub 168c:0002
[ 1240.537765] ath10k_pci 0000:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 1
[ 1240.549259] ath10k_pci 0000:01:00.0: firmware ver 10.4.1.00030-1 api 5 features no-p2p crc32 d2901e01
[ 1240.555240] ath10k_pci 0000:01:00.0: board_file api 1 bmi_id N/A crc32 7e56fd07
[ 1240.564370] ath10k_pci 0000:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal file max-sta 512 raw 0 hwcrypto 1
[ 1240.573511] ath10k_pci 0000:01:00.0: firmware register dump:
[ 1240.581206] ath10k_pci 0000:01:00.0: [00]: 0x01000000 0x00000000 0x00000000 0x00000000
[ 1240.586796] ath10k_pci 0000:01:00.0: [04]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.594575] ath10k_pci 0000:01:00.0: [08]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.602479] ath10k_pci 0000:01:00.0: [12]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.610320] ath10k_pci 0000:01:00.0: [16]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.618287] ath10k_pci 0000:01:00.0: [20]: 0x00000000 0x00401B60 0x00000000 0x00000000
[ 1240.626190] ath10k_pci 0000:01:00.0: [24]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.634094] ath10k_pci 0000:01:00.0: [28]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.641966] ath10k_pci 0000:01:00.0: [32]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.649817] ath10k_pci 0000:01:00.0: [36]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.657787] ath10k_pci 0000:01:00.0: [40]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.665677] ath10k_pci 0000:01:00.0: [44]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.673581] ath10k_pci 0000:01:00.0: [48]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.681485] ath10k_pci 0000:01:00.0: [52]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.689312] ath10k_pci 0000:01:00.0: [56]: 0x00000000 0x00000000 0x00000000 0x00000000
[ 1240.697262] ath10k_pci 0000:01:00.0: Copy Engine register dump:
[ 1240.705170] ath10k_pci 0000:01:00.0: [00]: 0x0004a000   6   6   3   3
[ 1240.710981] ath10k_pci 0000:01:00.0: [01]: 0x0004a400  26  26  28  29
[ 1240.717528] ath10k_pci 0000:01:00.0: [02]: 0x0004a800  47  47  46  47
[ 1240.724006] ath10k_pci 0000:01:00.0: [03]: 0x0004ac00   3   3   5   3
[ 1240.730374] ath10k_pci 0000:01:00.0: [04]: 0x0004b000 746 746  65  25
[ 1240.736846] ath10k_pci 0000:01:00.0: [05]: 0x0004b400  17  17  48  49
[ 1240.743281] ath10k_pci 0000:01:00.0: [06]: 0x0004b800   0   0   0   0
[ 1240.749645] ath10k_pci 0000:01:00.0: [07]: 0x0004bc00   1   1   1   1
[ 1240.756123] ath10k_pci 0000:01:00.0: [08]: 0x0004c000   0   0 127   0
[ 1240.762556] ath10k_pci 0000:01:00.0: [09]: 0x0004c400   0   0   0   0
[ 1240.768916] ath10k_pci 0000:01:00.0: [10]: 0x0004c800   0   0   0   0
[ 1240.775396] ath10k_pci 0000:01:00.0: [11]: 0x0004cc00   0   0   0   0
[ 1240.921152] ieee80211 phy0: Hardware restart was requested
[ 1242.529568] ath10k_pci 0000:01:00.0: device successfully recovered

(Last edited by RedVortex on 25 Apr 2017, 00:10)

The problem is worse with the CT firmware, the router reboots completely after a few seconds of high traffic on the wireless.

I don't have a serial console so I can't say what happens exactly before reboot though, I suspect a kernel panic or something similar...

At least I know that the CT firmware now works, but it has a stability issue worse than the original firmware for the 2600...

Unfortunately I can confirm this, when wifi is placed under load, the compat-wireless driver crashes and throws a trace..
eg:
------------[ cut here ]------------
[  578.981090] WARNING: CPU: 0 PID: 0 at compat-wireless-2017-01-31/drivers/net/wireless/ath/ath10k/htt_rx.c:1006 ath10k_htt_t2h_msg_handler+0x1064/0x1578 [ath10k_core]
...
.
. etc

very annoying. Been here since ages, and was always hoping this will go away one day.. but that is yet to happen, but the box under any AC-class load and poof.

Those of you testing the trunk and finding errors should report them to LEDE. There might not be anyone working on them.

One of the reasons ath9k devices are very stable now is because competent developers worked directly on the driver..

after qca trolled the community with "ath10k" which strips all important functionality from driver and puts it in "firmware", said developers abandoned qca entirely. Its actually primarily a single person, nbd

Now at the mercy of the community and charity patches from qca..

Commercially and for their partners, they use another closed source driver so I don't reckon they care much

TLDR; I reported that problem ages ago.. but hey, pls report if you come across it. more noise cant hurt.

(Last edited by james04 on 2 May 2017, 03:05)

So, we're on the "wrong" platform? Unfortunately, there doesn't seem to be an alternative with real MU-MIMO, does it?

No alternative. The marvell armada stuff found in the WRT series is much worse from what I hear..
Mu-mimo? AFAIK that stuff doesn't work in the current implementation, you need  to use stock FW for that..

Can anyone confirm i reboot is working on 17.01.1 for v1.1? Is there a way to upgrade without having to install everything back again?

(Last edited by danieldefreitasleite on 8 May 2017, 23:40)

Reboot should only be working in trunk or on special builds of 17.x.

Tx for the info.
Is there any added value of a ath10k device compared to a ath9k one if no ath10k specific functionality is supported in Lede?

I wanted to revert back to stock firmware and bricked my router. Has anyone managed to debrick it via tftpd server?

I connected my router directly to my PC via ethernet cable (used a LAN port on the router), set my PCs IP to 192.168.0.66, launched tftpd server and put a stock bin file. I hold on a reset button, power on the router and after 15s release it. The problem is that the router never tries to connect to my tftpd server...

The tftpd server works fine because I tried connecting from my windows console.

PS after holding reset button for like 10 seconds, the router is pingable at 192.168.0.1 for like 4 seconds. According to the debricking manual, the tftpd client should use 192.168.0.86 Ip? I never managed to ping that IP address during the process.

The IP needs to be 192.168.0.66 (you typed 66 and 86, so not sure which you used). It will actually try grabbing the file after about 7-8 seconds in my experience.

The file also MUST be named ArcherC2600_1.0_tp_recovery.bin . That is the only file it will try to grab.

I've used TFTP to recover several times. After it takes the file (which it does super fast), it takes a few minutes t (> 5) o flash and reboot. If you interrupt the process, it will be stalled in a weird state.

I used Wireshark and noticed that the router was always trying to access ArcherC2600_1.0_tp_recovery.bin at 192.168.0.66 but all packets were timed out. I tried multiple TFTP servers but in the end, it was my firewall blocking 69 port! I disabled it and succesfully managed to recover.

Thinking about buying this to replace an Archer C7 to have more CPU horsepower for VPN applications. But if the wireless is unusable, then I guess I'd just wait and see.

Or what would you recommend? Does anything more powerful than C7 have nice support?

For significant VPN speed improvements you need a CPU which supports hw encyrption. Routers don't have such CPU's. If I recall correctly an Odroid C2 does support it, but I haven't read anything on anyone using it for encrypted VPN connections.

(Last edited by Videopac on 19 May 2017, 22:24)

Wireless is not bad on the C2600, its just unstable only at full load.
Eg I have one in my place, all my devices are connected with wireless and it works fine without crashes.

So someone one day asked about a router for office, so I suggested the C2600. Here's the thing, this guy has over 30 devices concurrently active, mixed 2.4 and 5Ghz. The wireless crashes non-stop for him and he keeps calling me..
The ISP provided box, which has a far less powerful CPU and much less RAM, works stable with the same workload (stock FW though, not OpenWrt/Lede).

The CPU on the C2600 can do about 500mbps of VPN traffic IIRC, the stats can actually be found here in this thread if you look through it..
And thats on the CPU, the SoC has HW acceleration but its not used.
The WRT series (eg WRT1900ACS) also has HW acceleration and a driver is available, people benched those at even higher speeds.. but wireless from what I hear is quite bad on those.

In conclusion, if you dont have a gazillion wireless devices, and your VPN speed is <= 500mbps, then the C2600 should serve you well.

HTH

The Archer C7 can do 5 Mbps through a VPN connection. I'd expect C2600 to do maybe twice as much, but not much more than that. Are you sure you know where you get that 500 Mbps from?

How difficult would it be to get HW support for encryption? Could someone write it if someone else paid enough? Or is it about binary modules or the chip maker not releasing the specs?