I have this in dmesg:
nf_conntrack: default automatic helper assignment has been turned off for security reasons and CT-based firewall rule not found. Use the iptables CT target to attach helpers instead.
Probably not firmware related? Unless I have some old config somewhere. Does master need updated iptables settings?
There are some info about it if you google, does not make me any wiser other than "it's nothing to worry about". Probably some conntrack module triggering it, irc, ftp whatever.
In all honesty, writing a few lines about it certainly wouldn't hurt and it would have been a lengthy discussion if someone else would've made a patch about the change. It also seems to have a lot of regressions...
I've not been involved in the decision, nor the patches, at all (ath10k is fine for me on QCA9984, very personally I'd prefer to stick with mainline ath10k) and just tried to fairly summarize the discussions that happened on IRC. Following the discussions, there apparently are substantial issues with IEEE 802.11w and ath10k/ ath10k-firmware (mostly the later) on QCA9880-BR4A, including frequent firmware crashes - which have been debugged with- and fixed in ath10k-ct.
So far the experience from -ct has not been that great. I am seeing disconnections due to "excessive missing ACKs" and inactivity with my Android tablet.
I haven't really been able to test it (besides a couple of 5 minute tests) at all, as the mac80211 update breaks WDS/4addr functionality (WDS-AP) for me on QCA9984 (with both ath10k and ath10k-ct 4.13 && 4.16); it's apparently fine with ath9k.
Noticed that some units really wants to switch between 802.11nac and 802.11bgn quite often.... is there any way to prevent that other than using different SSID's for the networks?
How are your setup? Never really thought about it much but I'm wondering if CT could be the cause, because I have a feeling using the iPhone sometimes causes delays when using Safari that I never experienced before.
EDIT:
Or maybe I've misunderstood the concept, even if I disable the N radio and only have AC on units are connecting at 24.0 Mbit/s, 20MHz
6.0 Mbit/s, 20MHz
EDIT2.5:
Now iPhone usally connects at this rate, this is after changing Firmware back to non-CT, don't think I did anything else.
24.0 Mbit/s, 20MHz
325.0 Mbit/s, 80MHz, VHT-MCS 7, VHT-NSS 1, Short GI
I am having an issue where Wi-Fi becomes unavailable; my Samsung Galaxy S8 (as well as my wife's, and my MacBook Pro) all lose Wi-Fi connectivity despite appearing to be still connected. I can't ping the R7800's IP at the time it happens. The issue happens on the main OpenWrt release, as well as this one 18.06 (master had the driver crashing).
I don't see any kind of error in dmesg or in the logs. Is there anything I can do to pull more info to try to track this down?
Edit:
Just to clarify, other devices on the network that are wired are fine and I can ping the "AP" from the wired devices.
I can confirm issues with R7800-master-r8221-3b53d6fdbc-20181004-1842-sqfs-sysupgrade.bin firmware:
— Android randomly disconnects from Wi-Fi (Samsung Note 8, 5 GHz).
— Some errors in dmesg:
Yeah I've always had this issue with both stock and ct firmware on BCM43xx clients (pretty much all Samsung hardware)
It appears to be fixed by QCA's 3.6.0.1 firmware, see PR https://github.com/openwrt/openwrt/pull/1374
I think the fix will eventually end up in ct firmware as well but for now I'm using 3.6.0.1 qca firmware with much success.
I don't know if that dmesg stacktrace is related to it though.
hnyman. I am trying to recompile/update my firmware, but i believe there is something wrong with my build environment.
I have never had this problem until now.
I cannot compile past 8125: R7800-master-r8125-d3b43f49ac-20181007-1913-sqfs-sysupgrade.bin
It does not update any packages or kernel/kmods.
Also getting an error now: WARNING: Makefile 'package/feeds/packages/keepalived/Makefile' has a dependency on 'kmod-nf-ipvs', which does not exist.
What could be wrong?
Sounds like you may have got conflicts. If you get conflicts when updating sources, you may need to stash your changes, pull update and then pop the stash and fix conflicts.
git stash
git pull
git stash pop
(and then editing the possible change conflicts)
That was mainly a guess, as you did not provide any error messages. But sounds like normal git work.