Re: wifimgr with iwlwifi

On Sat, 26 Mar 2022, J.R. Oldroyd wrote:

Hi,

> On Sat, 26 Mar 2022 09:58:12 -0700 Kevin Oberman <rkoberman@gmail.com> wrote:
>>
>> On Sat, Mar 26, 2022 at 3:28 AM Takashi Inoue <inoue.takashi@nihon-u.ac.jp>
>> wrote:
>>
>>> Do you use "wifimgr" which is a GUI tool to select AP.
>>> When I use "if_iwlfiwi.ko", wave strength displayed
>>> in wifimgr is 0 for all access point.
>>> Do you see this phenomenon? Or, does someone see this?
>>>
>>> Best,
>>> Takashi
>>>
>>
>> I'll have to leave the wifimge to others.
>
> Wifimgr parses the output of "ifconfig -v wlanN list scan" in order to
> compute the number of bars to display for the signal strength.
>
> With the iwm driver, I see s/n values like -49:-96 for my network.

I am actually not sure if the values in iwm are correct;  there is a
factor-2 calculation in there which I believe is only done for max tx
power of older cards based on eeprom values but not for the rssi
anymore in iwlwifi.  I haven't started digging where that originated,
so hard to say.


> But with the iwlwifi driver, the values for the same network are -121:-96 .
>
> This large negative value for the signal strength is causing wifimgr
> to compute the number of bars as 0.

Quick shot.  I just pushed two changes to main for anyone who wants to
try.

If you are on stable/13 and want to test them early you can (unitl
they'll be merged):

git cherry-pick -x 4a22cd6c4e5f4fc4a38aa7400742d4005c5ae3de
git cherry-pick -x 60970a328e280b25b05f1d9a9de1ef91af573c4a

Let me know if values start to look more reasonable.

Happy weekend,
Bjoern

-- 
Bjoern A. Zeeb                                                     r15:7