19:07 <f_> I'm comparing U-Boot's MMC_TRACE output to SPL's, hm..

19:07 <f_> so there's an additional CMD6 being sent in SPL...

19:08 <f_> with args 0x03af0100

19:08 <f_> it seems to try touching the ERASE_GROUP_DEF register ... which, uh, not sure why it does that

19:10 <f_> Then obviously there's also an additional CMD13 being sent, as CMD6's response is of type R1b

19:14 <f_> I don't think it's what's causing an issue

19:14 <f_> but still might be worth noting..

19:16 <xdarklight> minute: https://paste.debian.net/1326110/ brings the fix from https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/net/wireless/realtek/rtw88?id=28818b4d871bc93cc4f5c7c7d7c526a6a096c09c to the sdio HCI. I barely made sure it "works" in the sense of me not seeing any significant breakage. maybe you can give it a try and see if it fixes/improve your connection issues?

20:03 <minute> xdarklight: ohh nice, will try it in a few minutes

20:13 <minute> so, coincidentally after turning on the device the wifi was in a broken stage at first

20:13 <minute> so it connected to AP but packets were not flowing, and i had multiple of the following in dmesg:

20:13 <minute> [ 39.551307] rtw_8822cs mmc2:0001:1: failed to get tx report from firmware

20:13 <minute> but doing ifconfig wlan0 down and then up again made it work

20:14 <minute> i didn't expect that to work because "failed to get tx report from firmware" sounds like it's in some broken state, and i wonder how reconnecting gets it out of there

20:15 <minute> anyway, will try the patch now

20:19 <minute> already getting a hang just when trying to load speedtest.net

20:20 <minute> (complete stall of all wifi)

20:20 <minute> works after down/upping the interface

20:21 <minute> i think the hardware is just borked in some deeper way, maybe on the sdio connection

20:21 <minute> or power glitches with the wifi module or sth

20:23 <minute> my current plan is to give up on the a311d+8822cs combo and ship an alternative wifi card for the wwan port (from asiarf)

20:24 <minute> i guess usb might be the more reliable connection solution for wifi on a311d...

20:39 <minute> interesting https://lore.kernel.org/linux-wireless/87zg5mjeu4.fsf@kernel.org/T/#m65695e06fefb8cc5ae541dadacdd89ff540b875f

20:42 <minute> also https://lore.kernel.org/linux-wireless/20230612133023.321060-1-petter@technux.se/T/#u

20:42 <minute> i wonder how to disable hw offload scan

20:54 <minute> ok, found a way to do that

20:55 <minute> ok, no use on my hardware. speedtest immediately stalls the connection

20:56 <minute> i also tried replacing the create_singlethread_workqueue with alloc_workqueue high pri, but noticed that there's only a tx workqueue for sdio, not for rx