Re: MT7601U Kernel Driver Issue

Re: MT7601U Kernel Driver Issue

[Jakub Kicinski]

On Thu, 27 Jul 2023 16:00:45 -0400 Adam Elzinga wrote:
> Hi Jakub,
> 
> I've been experimenting with your mt7601u kernel driver (both building from
> source and built-in to kernel 4.14.98 and 5.19.0) and think there may be a
> bug when reading EEPROM values from a couple different mt7601u-based
> devices. Full issue report is posted here:
> https://github.com/kuba-moo/mt7601u/issues/94
> 
> In short, are these EEPROM values unexpected, they seem to be the cause of
> repeated kernel warnings?
> 
> RF freq offset: 5d
> RSSI offset: 0 0
> Reference temp: f9
> LNA gain: 0
> Reg channels: 1-14
> Per rate power:
>          raw:00 bw20:00 bw40:02
>          raw:00 bw20:00 bw40:02
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
>          raw:34 bw20:f4 bw40:f6
> Per channel power:
>          tx_power  ch1:04 ch2:04
>          tx_power  ch3:04 ch4:04
>          tx_power  ch5:04 ch6:04
>          tx_power  ch7:04 ch8:04
>          tx_power  ch9:05 ch10:05
>          tx_power  ch11:05 ch12:05
>          tx_power  ch13:05 ch14:05
> 
> Emailing you directly as I realize your GitHub repo is no longer maintained
> now that the driver is included in the Linux kernel. Apologies if this is
> not the proper way to report an issue.

I saw the GH issue, just didn't get to it yet :) I'm adding
the linux-wireless@ mailing list to the CC as is customary
(if you reply please make sure not to top post and not to send 
HTML emails, only plain text).

I think that as you mentioned on GH the warning is harmless.
The vendor driver just caps the value to 6 bits (it's in 
the MT7601_ReadTxPwrPerRate() function AFAICT in the driver 
from here:
https://www.mediatek.com/products/broadband-wifi/mt7601u )

If you are able to compile the vendor driver it may be a good idea
to check that my calculation gives the same result. At which point
we can confidently remove the warning. Would you be willing to send
a patch yourself?