Netgear R7800 wifi internet connectivity drops

Hi,

Running 19.07.0-rc2 r10775-db8345d8e4 on my R7800.

On this firmware - and all of the previous OpenWRT versions I installed since I purchased the R7800 during summer - I get an occasional loss of internet connectivity on some wifi clients. Perhaps 3-4 times a day. Wifi connection symbol in windows stays on, but loss of internet connectivity is detected by Windows. I also confirmed I cannot ping any other devices on the LAN. The loss of connectivity lasts about a minute or two - and then it is working again. If I disconnect/reconnect Wifi it is resolved (until it hits again)...

Wired Ethernet seems OK, so I do not believe this is a WAN issue - but rather some Wifi/routing issue.

Few weeks back I tried switching the wifi firmware to the "-ct" version:

[   11.994391] ath10k_pci 0000:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[   11.994431] ath10k_pci 0000:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 0
[   12.005519] ath10k_pci 0000:01:00.0: firmware ver 10.4b-ct-9984-fW-012-6acc9b999 api 5 features mfp,peer-flow-ctrl,txstatus-noack,wmi-10.x-CT,ratemask-CT,regdump-CT,txrate-CT,flush-all-CT,pingpong-CT,ch-regs-CT,nop-CT,set-special-CT,tx-rc-CT,cust-stats-CT,txrate2-CT,beacon-cb-CT,wmi-block-ack-CT,wmi-bcn-rc-CT crc32 2f261949

...but this had no effect on this problem (on the other hand I now witnessed a crash of this firmware in the kernel log; but only one).

Does anyone have a clue here? Seems there are a lot of happy R7800 users here, so I am puzzled why this hardware is acting up for me...

Recommend to upgrade to 19.07.0 first, probably without keeping the settings and only core packages to run.

stick with current 18 or run a hnyman master build ( if you don't need wireguard / kmods )... 19 is not stable enough...

Version 18 exhibited the same problem for me, hence I jumped on the new version with the hopes it would be resolved.

Did hear about the hnyman builds; could try that. What is the difference between mainline and these builds? And do I have to wipe settings to update to this? I'd hate to config everything again....

Checkout his thread...

suggested because it gives you "snapshot" with luci out of the box... + it's been vetted a little so as to reduce any teething issues due to any (unlikely) bugs in "todays" snapshot...

this device/target has seen active development since 19... so there is a big chance a snapshot release will resolve / alter problematic symptoms...

Thanks for your suggestions.

Switched to hnymans build from beginning of January (took a while to find the Dropbox link in fact). Cool set of features baked in this firmware :slight_smile:

So this ran since yesterday evening, but already today two times the problem happened. One time may correlate with a Wifi firmware crash:

[38628.604912] ath10k_pci 0001:01:00.0: firmware crashed! (guid n/a)
[38628.605031] ath10k_pci 0001:01:00.0: qca9984/qca9994 hw1.0 target 0x01000000 chip_id 0x00000000 sub 168c:cafe
[38628.610033] ath10k_pci 0001:01:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 1 testmode 0
[38628.624414] ath10k_pci 0001:01:00.0: firmware ver 10.4b-ct-9984-fW-012-17ba98334 api 5 features mfp,peer-flow-ctrl,txstatus-noack,wmi-10.x-CT,ratemask-CT,regdump-CT,txrate-CT,flush-all-CT,pingpong-CT,ch-regs-CT,nop-CT,set-special-CT,tx-rc-CT,cust-stats-CT,txrate2-CT,beacon-cb-CT,wmi-block-ack-CT,wmi-bcn-rc-CT crc32 877928bc
[38628.635688] ath10k_pci 0001:01:00.0: board_file api 2 bmi_id 0:2 crc32 85498734
[38628.656724] ath10k_pci 0001:01:00.0: htt-ver 2.2 wmi-op 6 htt-op 4 cal pre-cal-file max-sta 32 raw 0 hwcrypto 1
[38628.665880] ath10k_pci 0001:01:00.0: firmware register dump:
<snip>

I also see a good deal of these in the system logs:

Thu Jan 23 14:13:24 2020 daemon.info hostapd: wlan1: STA 5c:f6:dc:3e:62:1b IEEE 802.11: disassociated due to inactivity

Is this normal? Will it kill the connection?

Will try to keep an eye on the logs and see if anything else correlate with the problems...

The thing still runs intermittent. I'd argue 5g is close to unusable. Several VolLTE calls goofed up with garbled audio/call drops.

Logs don't produce anything of value except for the occasional:

Fri Jan 31 17:20:47 2020 daemon.info hostapd: wlan1: STA 80:b0:3d:c0:c2:4d IEEE 802.11: deauthenticated due to inactivity (timer DEAUTH/REMOVE)

...which I am unsure is a problem? I also see these popping up:

Fri Jan 31 17:27:23 2020 daemon.info hostapd: wlan1: STA 5c:f6:dc:3e:62:1b RADIUS: starting accounting session 3148089D5ADC5547

...why is RADIUS popping up here? I have not configured a RADIUS system. But I also don't know if it can play a useful role without an authentication server.

try different wireless firmwares:

I did try a few. The original one in 19.07. Then switched to a -ct version i pulled . And now I see another one in the hnyman build. All give problems. I could try a fourth - but not sure which one to pick and by which argument. Just looking for a stable connection.

Perhaps someone here with a rock solid performing 7800 could share which version they are using?

Any other ideas? Considering if I should purchase some other hardware; wife and kids are complaining :slight_smile:

I suspect these issues can come from multitple sources (many not related to your choice of wifi firmware or settings).

You likely will have to spend some time searching the forums and learn from others experience.

There are some adjustments to the /etc/config/wireless you can make (for both 2.4 and 5 GHz radios). For example an excerpt from my /etc/config wireless:

config wifi-iface 'default_radio1'
...
	option wpa_group_rekey '86400'
	option max_inactivity '3600'
	option disassoc_low_ack '0'
...

search the forum for disassoc_low_ack to start.

EDIT 0: apologies for the botch link... should be fixed now.
EDIT 1: ugh, i just can't get seem to get my links right today

Regarding audio/video, there is a recent thread by an experienced R7800 user having issues that turned out to be related to router location (a dead spot). In that thread there are some suggestions about tracking down wifi interference from neighbors...

FWIW I use the ath10k-ct driver and ath10k-ct-htt firmware on the r7500v2 (an older and slightly different cousin of the r7800) running a recent build from "master". The device is an AP only and will run for 30+ days without reset.

I see the log entries that you reported earlier but can't attribute them to wifi drops that are the fault of the wifi firmware. That said, I do have wifi drops (somes multiple times per day) on both iphones and android phones. I suspect there are multple causes not least of which is these devices going into deep sleep states. For my use case, these drops are tolerable enough that I don't log packets in an effort to see if this can be improved at the router.

My laptop 5 GHz connection is rock solid.

HTH

OpenWrt 18.06.6 on R7800 user here. Still rock stable. Will update to 18.06.7 soon to have recent security updates. classic ath10k. No one complains about stability here.

19.07-SNAPSHOT. Everyone stopped complaining once I switched back to the non-ct firmware and driver.

In master the non-ct driver still semi-constantly produces errors in kernel log and crashes router sooner or later:

[   71.542629] ath10k_pci 0001:01:00.0: SWBA overrun on vdev 0, skipped old beacon
[   71.549676] ath10k_pci 0001:01:00.0: SWBA overrun on vdev 0, skipped old beacon
[   71.557110] ath10k_pci 0001:01:00.0: SWBA overrun on vdev 0, skipped old beacon
[   71.564184] ath10k_pci 0001:01:00.0: SWBA overrun on vdev 0, skipped old beacon

The non-ct got broken in master in July 2019, after 19.07 was branched off, so 19.07 branch may still have the "old" and working non-ct mainline driver.

1 Like

19.07.x has kernel v4.19 based ath10k backports, the newer v5.4 based ones in master are unusable. ath10k-ct is functional in both branches.

Thanks for the pointers all.

Based on the feedback going in different directions here, I am unsure what is the right choice. I now used 19.07.1 r10911-c155900f66 for the last month - no improvement. Few days back I did a "Reset to defaults" and manually configured it up again to rule out some stale configuration change messing up things. No improvement.

Is it worth attempting to downgrade to 18.06 as suggested by some? Security updates are important to me (in fact the reason I use 3rd party firmware).

I am now seriously considering buying another device to get a stable setup. I really do not care about traffic shaping or insane bandwidths. But availability and responsiveness is everything. My old ASUS N56-U chugging along with Padavan firmware would run solid for 6-12 months and NEVER exhibited this flakyness; and that is dated hardware to say the least.

Would the Archer C7 be a sensible alternative to R7800? It is wonderfully cheap :slight_smile:

Would try the upcoming 18.06.8...it's free ;- )

  • try both 2.4 / 5Ghz (if possible) and choose none DFS channels
  • KRACK mitigation / Disassociate On Low Acknowledgement disabled?

You wrote "on some wifi clients" ....perhaps client dependend?

I have also intermitting DNS resolution / ping problems on 19.07.x in conjuction with adblock and unbound....18.06 is working fine with other minor problem (perhaps LineageOs Android client dependent)

Did you try the non-ct firmware/driver? Either those in 19.07 or the latest? I am currently running firmware-5.bin_10.4-3.9.0.2-00086 in 19.07 and it is awesome.

1 Like

I would like to try this. Could you tell me how to install it manually?

Welcome back to the forum!

This should help you with a step-by-step:

Hello from a double newbie - to OpenWRT and to this forum.

Experience the same problem with my Netgear Centria WNDR4700 - WiFi connection to all devices drops every few days and has to be restarted to get working again.

The firmware for the Atheros devices from the last post that has to be removed is not installed.

Will post the dmesg output on the next crash.

Thanks in advance,

Max