Dynalink DL-WRX36 Askey RT5010W IPQ8072A technical discussion

After a few days of work both Wi-Fi radios stop working... happened 3 times already, over past 2 weeks.
Restart resolves the problem.

Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.444838] ath11k_warn: 49 callbacks suppressed
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.444857] ath11k c000000.wifi: failed to send WMI_PDEV_BSS_CHAN_INFO_REQUEST cmd
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.448571] ath11k c000000.wifi: failed to send pdev bss chan info request
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.456201] ath11k c000000.wifi: failed to send WMI_PDEV_SET_PARAM cmd
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.462966] ath11k c000000.wifi: Failed to set beacon mode for VDEV: 0
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.469570] ath11k c000000.wifi: failed to send WMI_BCN_TMPL_CMDID
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.476145] ath11k c000000.wifi: failed to submit beacon template command: -108
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.482407] ath11k c000000.wifi: failed to update bcn template: -108
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.490046] ath11k c000000.wifi: failed to send WMI_VDEV_SET_PARAM_CMDID
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.496371] ath11k c000000.wifi: failed to set BA BUFFER SIZE 256 for vdev: 0
Sun Oct 15 16:56:25 2023 kern.warn kernel: [924435.503155] ath11k c000000.wifi: failed to send WMI_VDEV_SET_PARAM_CMDID

OpenWrt SNAPSHOT r23930-6cf27094e9 / LuCI Master git-23.236.53281-98e3743

I posed the same question in my thread. How does the dynalink and/or the dynalink + NSS compare to the r7800 with NSS? Theoretically AX should offer more throughput but how much more with speedtest (or iperf)?

Here is the r7800 with NSS, client is an iphone 13 (2x2 client) on a 80Mhz channel. Just walked up to my r7800 and ran the test just now, stock build settings.

Based on my experience, the router wifi is stable if there is just one SSID per radio, but if there are multiple SSIDs, then the radio may crash at come point where a client leaves/disassociates.

In style of

3 Likes

Here's what I put in my startup. @hnyman did some of his own testing and found moving too much around created massive instability, but these have been rock solid for me.

#assign 4 rx interrupts to each core
echo 2 > /proc/irq/66/smp_affinity
echo 4 > /proc/irq/67/smp_affinity
echo 8 > /proc/irq/68/smp_affinity
echo 8 > /proc/irq/69/smp_affinity

#assign 3 tcl completions to 3 CPUs
echo 2 > /proc/irq/47/smp_affinity
echo 4 > /proc/irq/52/smp_affinity
echo 8 > /proc/irq/55/smp_affinity
1 Like

Here's some iperf tests that were run recently when testing various wifi board BDFs. No NSS. Two posts prior was same-room testing.

Great that someone else has come to this conclusion (and not great to experience the issue :slight_smile: )
This is my experience too. More than one SSID per radio and whenever I shutdown my laptop, all radios crash and the router needs to restart (manually, by script or watchcat addon).
Do you know why that happens?

No, as the crash happens inside the closed-source firmware blob.
But I suspect that there is a bug where the SSID instances get mixed inside the firmware. E.g. some function always gets applied to the first SSID while the action should actually be performed on second/third SSID...

2 Likes

How do you determine max txpower?

By reading them with the WMI command: Dynalink DL-WRX36 Askey RT5010W IPQ8072A technical discussion - #1990 by lytr

At short distances, the Dynalink may get up to 1.4 times the WIFI speeds of R7800 with NSS. However, R7800 has better range and is very stable.

1 Like

Your crash theory makes a lot of sense. My router has many SSIDs on each of the radios, and stable it is not. (Currently trying WLAN.HK.2.9.0.1-01890-QCAHKSWPL_SILICONZ-1.)

Is there a recommended way to feed this back to the Qualcomm Innovation Center (quic) providing the firmware? You'd figure they'd want to know :man_shrugging: Their GitHub has the typical Issues page, but I'm not sure that's being monitored. Just wondering...

Why is the "release" of 23.05.0 showing as "snapshot" for this device in the firmware selector when the RC was showing the normal build? I am on rc4 but if I'm correct and load the "upgrade" to the snapshot I will lose Luci and have to screw around to get it back.

If by screw around you mean simply typing "opkg update && opkg install luci" then yea :slight_smile:

Anyway, it's not showing as a snapshot on firmware selector. Maybe just on the attendedupdate page.

https://firmware-selector.openwrt.org/?version=23.05.0&target=ipq807x%2Fgeneric&id=dynalink_dl-wrx36

Ah, ok, I was looking in the wrong place... silly me. :slight_smile:

And yeah, I know its not that hard to put Luci back, but I would prefer not to have to reload all the configuration too, and when things aren't what I expect that would mean I might "get to"....

Thanks.

Then use LuCI Attended Sysupgrade or cli auc to do your upgrades. They keep all your config and installed packages...

In my experience, I feel I get better milage if I stick with 5ghz only for multiple ssids and just 1 ssid on 2.4Ghz.

Current openwrt builds have an issue where wireless radio (mostly 2.4GHz) crashes, requiring a reboot. Therefore I would not recommend getting WRX36 at this time, unless you're ok rebooting every day at midnight.

1 Like

I bought a cheap ipq4019 based AP for 2.4GHz and I am using 5GHz on WRX36 exclusively.

1 Like

So I'm doing something kind of similar, I'm using the 2.4Ghz exclusively for IoT ssid but I also have DIR-505 (x5) around the home also broadcasting IoT ssid on 2.4Ghz as a redundancy.

It's working well.

Anyone try a new stable build recently?

…what? Nothing posted makes any sense. The IRQs I listed and manually assigned are the only ones being assigned. If you ssh and run cat/proc/interrupts - you will see these IRQs and it may make more sense.

Based on your reply I’m assuming you may not fully understand IRQs and how they work?

1 Like