My 5ghz ssid stops working after some time.
Kernel log says:
1075926.819433] ath10k_pci 0000:01:00.0: Invalid peer id 169 peer stats buffer
[1354576.170673] ath10k_pci 0000:01:00.0: Invalid peer id 204 peer stats buffer
[1423408.228237] ath10k_pci 0000:01:00.0: Invalid peer id 200 peer stats buffer
[1466978.074911] ath10k_pci 0000:01:00.0: Invalid peer id 133 peer stats buffer
[1541328.370156] ath10k_pci 0000:01:00.0: rx ring became corrupted: -5
After a reboot it is fixed.
I've installed build r17217 and applied ath10k mainline using hnyman instructions in this topic. I tried mainline because the default ath10k-ct gave the same issue. The same issue occurred also in previous builds.
Is there a fix or workaround available? Or anything else I can do?
I've tried some other things that you can try if you haven't already. I'm not an expert and no clue if it's directly related to your problem, but if you haven't tried swapping out the firmware blob in addition to the driver it might help. I had 5ghz stability issues that were fixed by moving to mainline non-ct driver and the mainline non-ct firmware blob.
It's not directly related to hnyman's build so out of scope for this thread, but I'll point out a few things that have already been said in the thread that I have tried -- your mileage may vary and if anyone else has a different view feel free to chime in.
In this build hnyman is providing the ath10k mainline driver as an IPK as a favor to us (thanks hnyman!!!). The other variable is the firmware blob which he does not provide. Like hnyman says above, you can try a combination that works for you and your hardware.
In Post 3016 hnyman shows one of many different ways to do this.
The key steps are going to be the same: (1) remove the -ct firmware blob and (2) install the mainline firmware blob.
You should read the respective posts but to summarize you can do one of the following, remember to reboot after this.
Like in hnyman's post you can use opkg list-installed | grep ath10k before and after the above to see the differences.
Read the posts above and the ones after them since they will provide some more context and information.
Again, the ath10k driver and firmware blob aren't really related to hnyman's specific build so if you can't get by with the above I would recommend starting another thread or feel free to DM me.
FWIW I've been running master-r17341-5c13177c55-20210815 for a day or so with no issues using the mainline driver and mainline firmware blob.
Thank you!
Just updated the firmware to master-r17341 and followed lv4nl and your summary. Hopefully the wifi will keep stable.
One last question, after every firmware update I need to repeat these steps.
Should I keep my firmware updated with the latest releases from hnyman or if for instance master-r17341 is good just keep it running for several months/years?
I usually update my firmware to the latest whenever I get the change and the wife is not at home to complain the wifi isn't working
Yes, that's correct. hnyman has done a fantastic job giving us a build that works for him, and a majority of us. If you're finding that the adjustments to the ath10k driver and firmware blob are better for your situation, then every time you update the router firmware you're starting with the hnyman build as your base, and then add and adjust from there which would include updating the ath10k driver and firmware blob as needed.
That's totally up to you. Sometimes I've done what you are saying, upgrading for the sake of upgrading just so I can be on the latest version. You do have to be mindful that the latest isn't always the greatest. With every new release there could be problems introduced that are totally out of hnyman's control, in addition to new features or fixes.
Another view is upgrading only when you need to upgrade. If everything is working and the router firmware is doing everything you need, there could be no reason for you to upgrade. If there is a specific problem you're having or a feature you think is missing, you can keep an eye on what is changed in any given release and then decide if it makes sense to upgrade.
Just to add to the conversation...
I've been using hnyman's build since I got my R7800.... it's been over a year - but not sure exactly when. Anyway, a few months ago I noticed my wife and my own laptop both dropped internet. They were still connected to 5Ghz, but I couldn't ping the router or anything - not even on LAN. A which disconnect-reconnect fixed it. This has been happening now for a few months... effects devices on 5Ghz only, laptops and phones.
I've been trying to find out what is causing it, but haven't figured it out yet. I updated to the latest firmware 2 or 3 times thinking that would fix it - but still happens. Today it happened... system log didn't show anything... and I can't tell if what is in the kernel log is related or not:
[1993612.984977] ath10k_pci 0000:01:00.0: mac flush null vif, drop 0 queues 0xffff
[1993612.986322] ath10k_pci 0000:01:00.0: peer-unmap-event: unknown peer id 0
[1993612.991179] ath10k_pci 0000:01:00.0: peer-unmap-event: unknown peer id 0
[1993619.510934] ath10k_pci 0000:01:00.0: 10.4 wmi init: vdevs: 16 peers: 48 tid: 96
[1993619.510965] ath10k_pci 0000:01:00.0: msdu-desc: 2500 skid: 32
[1993619.593585] ath10k_pci 0000:01:00.0: wmi print 'P 48/48 V 16 K 144 PH 176 T 186 msdu-desc: 2500 sw-crypt: 0 ct-sta: 0'
[1993619.594190] ath10k_pci 0000:01:00.0: wmi print 'free: 84920 iram: 13156 sram: 11224'
[1993619.977212] ath10k_pci 0000:01:00.0: rts threshold -1
[1993619.977832] ath10k_pci 0000:01:00.0: Firmware lacks feature flag indicating a retry limit of > 2 is OK, requested limit: 4
Anyway, what you are seeing sounds similar - but might be a little different. Not sure if anyone else on here has experience this "still on network but can't reach anything" behavior. I still am not sure if the above is the cause or not - because I really don't know what I'm reading...
Just for reference currently running on R7800-master-r17216-8c2509dc5f-20210725-2040-sysupgrade.bin. So yeah, if anyone has seen this or knows what or where I need to look for the correct error log, please let me know .
The times I see behavior like this, it's usually the interface to the outside world that's causing it. I'm on cable internet, and when this happens I look at the cable modem's built-in web interface, and it will be full of errors and lost connections. Rebooting/resetting the cable modem will usually fix it.
Of course, you might be on some other kind of provider altogether, in which case, please ignore my silly suggestion. Others who still use cable might find it useful, though.
Nothing wrong with a suggestion!
Unfortunately, I don't think this is my problem. 2.4Ghz and wired clients are fine, but 5Ghz is effected. I have updated the firmware one or two times since then - so multiple reboots have happened.
Hi sorry bit of a noob here have flashed dd-wrt before but never openwrt but have done plenty of reading on the r7800 page
I have just purchased a used 7800 for £50 which is due in the next few days so it will be running stock firmware, did a bit of research and this build comes highly recommended
A little bit confused by what I need to flash from the dropbox link - obviously I see the latest builds do I just select the factory img and thats it? confused because I see some people mentioning the master builds
Also what are the ipk's? what I am hoping to do in the next few days is set this up with 2 AP's (Archer C7 and a Mi Router 3) and ulitmately set up wifi fast roaming
Now that 21.02.0 is out, and apologies if this has already been answered, but has the R7800 target already been converted to the new dsa framework , or is that going to come later?
Notre that I have nothing to with the downloadable packages. Check the buildbot faillogs etc. Theer has been lots of major toolchain changes in the last 1-2 days: musl upgrade, gcc upgrade, binutils upgrade, libtool update so quite possibly buildbot may take 1-2 days before everything is again coherent.
EDIT: likely a symptom of the change reverted a few minutes ago:
I have built Test DSA images and seems wireless clients are isolated despite client isolation disabled. Can anybody with Test DSA build verify if it is possible to ping one wireless client from another?
BTW problem with VLANs enabled I was experiencing earlier is gone.
.
