You can also install the attended-sysupgrade package and from there you can upgrade snapshots with all your currently installed packages.
Does this work reliably and safely now to be a recommendation? I could be wrong, but I think there was some commotion about using this at some point in the past.
Save for a breaking commit that was fixed later (Dynalink DL-WRX36 Askey RT5010W IPQ8072A technical discussion - #2812 by NinjaBreaker), it has been working fine.
If I'm running 23.05.2, but having issues with the wifi periodically dropping, should I install this wifi device firmware (https://github.com/quic/upstream-wifi-fw/tree/main/ath11k-firmware/IPQ8074/hw2.0/2.9.0.1/WLAN.HK.2.9.0.1-01977-QCAHKSWPL_SILICONZ-1) or just go to the newest snapshot? The random wifi device issues are playing havoc with my IOT devices.
The snapshots are significantly more up-to-date with many more commits and overall fixes, so I generally always recommend them. They are remarkably stable as well overall. Just remember that snapshots do not include Luci by default, so be prepared to use command line to add it... or use something like the Firmware Selector to assist with adding in your preferred packages to a custom build.
I am running OpenWrt 23.05.2 r23630-842932a63d and I have absolutely zero problems. Well, unless there is something I haven't realized.
The issue has been intermittent, which is frustrating. From reading it seems to be related to clients disconnecting, but it can affect all clients. Even more rarely the wifi will just crash needing a router reboot.
I've upgraded to snapshot. I'll monitor if the issue comes back. The upgrade was easier than expected. I had no apparent issue with the ipq807x target rename.
To be clear, I'm trying to solve the error:
ath11k c000000.wifi: failed to flush transmit queue, data pkts pending 9, which causes pauses for all clients.
I get multiple of these when a client disconnects, usually increasing numbers. The newest snapshot doesn't fix it. I'm continuing to research, but it appears that this is an issue with the IPQ807x firmware.
Is it related to this?
I don't think so. The ath11k failed to flush error happens when a client (like a phone) leaves the area of the wifi. Since I have two APs in the house, this can happen roaming, but most often happens when someone leaves the house. I've noticed some wi-fi impacts to other devices when this happens. The system log gives errors like: kern.warn kernel: [ 4910.498400] ath11k c000000.wifi: failed to flush transmit queue, data pkts pending 3 Occationally, this causes the wifi to crash.
a related post is here, but there doesn't seem to be a solution yet: Ath11k c000000.wifi: failed to flush transmit queue, data pkts pending
I know there's a lot of chatter around this with the Mediatek based gl.inet mt6000. This is a thread regarding their issues. Maybe it carries over?
You are not alone. I have the same problem on this unit. I put some thoughts of what is happening here: Dynalink DL-WRX36 Askey RT5010W IPQ8072A technical discussion - #2746 by NinjaBreaker
I only see this in ath11k along with the warning you describe in the kernel.
Are you guys running latest Ath11k firmware?
I'm running: ath11k c000000.wifi: fw_version 0x290604a5 fw_build_timestamp 2023-10-12 02:06 fw_build_id WLAN.HK.2.9.0.1-01977-QCAHKSWPL_SILICONZ-1
Regarding device development, I finally have a weekend uninterrupted.
Please answer me, before I invest a significant amount of time reverse engineering again, is there any point to looking into the uboot's OEM dual partition system? I'm not sure if "bootipq" (which is the command that actually uses the original dual partition mechanism) can be made to work with OpenWRT images.
What was the reason that didn't work originally?
I am not a developer, but would it have any similarity in reason to this post?
https://forum.openwrt.org/t/askey-rac2v1k-rt4230w-rev6-support/15830/81?u=justanotherenduser
That would look highly similar, although that's not quite the same device.
Can a dev weigh in on whether the behavior is different here? If bootipq works (for the most part barring some kind of extreme kernel boot corruption discrepancy) perhaps there could be a guide to set it up in this manner rather than bootm. I guess for that, we don't set bootcmd, but what other lines would be wrong?
If bootipq will work I think I can sit down with this thing and figure out a way that a partition swap can be triggered (for sysupgrade purposes, advanced reboot, etc.) assuming someone's willing to merge (pseudo?)code that'll do that.
We already have a way earlier in the thread to detect what partition it's on, though I think there may have been a typo in that code? I posted an addendum on it a while back but don't recall the details.
Are the updated code fixes etc in snapshots rolled into the service updates? For instance, I just updated to 23.05.3 yesterday.
Wondering if this update would include everything in the snapshots up until that was released or not.
Mainly not.
The service release 23.05.3 is still mainly the may 2023, plus some fixes and backports from master.
Only really minor part of stuff in master is backported to stable branches like 23.05
But development master is one year ahead regarding many aspects.
Oh man. Now I really want to try the snapshot. 
luci is the only package not included by default in the snapshot vs a release correct?
I click HTTPS redirect for my luci config, but i dont care about certs and just ignore the warning.
Do I need luci-ssl in this scenario?
