Re: No support for Realtek 8821CE RFE Type 6

[Stuart Longland <stuartl@longlandclan.id.au>]

On Mon, 8 Nov 2021 03:28:07 +0000
Hi Ping Ke, Bogdan et all…
Pkshih <pkshih@realtek.com> wrote:

> Thanks for your information. 
> Since we don't have this module, I'll have internal discussion further.
> 
> I have a quick look of RFE type 0, 2, 4 and 6. I think type 0 and 6
> are similar. So, could you try following patch to see if it works?
> 
> diff --git a/rtw8821c.c b/rtw8821c.c
> index 80a6f4da..1b86a21b 100644
> --- a/rtw8821c.c
> +++ b/rtw8821c.c
> @@ -1513,6 +1513,7 @@ static const struct rtw_rfe_def rtw8821c_rfe_defs[] = {
>         [0] = RTW_DEF_RFE(8821c, 0, 0),
>         [2] = RTW_DEF_RFE_EXT(8821c, 0, 0, 2),
>         [4] = RTW_DEF_RFE_EXT(8821c, 0, 0, 2),
> +       [6] = RTW_DEF_RFE(8821c, 0, 0),
>  };

I'll give this patch a try but I stumbled on another machine with this
particular WiFi chip… the Leader SN4PROv3.

I have 14 of them sitting here… all victims of the 2022 Brisbane flood.
(My workplace at Milton bought a lot of these to load with Linux and
run NodeRED flows on … then the Brisbane River decided to get out of
bed and check them out.)

Despite this, they all power on, and I'm going through and assessing
the condition of the survivors.  If there's any corrosion, so far it's
on mounting hardware like screws, with little on the PCBs itself.  The
machines were in their factory packaging at the time.

I've got three powered on now, doing a test install of Gentoo Linux as
a stress-test.  Two of the lot are happily talking on the WiFi network
here with no ill effects… barring the odd complaint:

[    4.076641] rtw_8821ce 0000:03:00.0: enabling device (0000 -> 0003)
[    4.084721] rtw_8821ce 0000:03:00.0: Firmware version 24.8.0, H2C version 12
[    4.108215] rtw_8821ce 0000:03:00.0 wlp3s0: renamed from wlan0
[    7.196982] rtw_8821ce 0000:03:00.0: start vif 24:18:c6:cd:e2:28 on port 0
[   16.273500] rtw_8821ce 0000:03:00.0: sta 7a:45:58:c8:b4:33 joined with macid 0
[   16.801990] rtw_8821ce 0000:03:00.0: failed to get tx report from firmware
[ 3104.402095] rtw_8821ce 0000:03:00.0: firmware failed to leave lps state
[ 3844.402114] rtw_8821ce 0000:03:00.0: firmware failed to leave lps state
[11020.402339] rtw_8821ce 0000:03:00.0: firmware failed to leave lps state
[11404.402353] rtw_8821ce 0000:03:00.0: firmware failed to leave lps state

A third one reports the error mentioned in this thread.  I'll grab more
details when I get it onto a wired network (at the moment it's
isolated).  Sadly, I've just nuked its Windows 10 installation.  (Not
that I expect such an OS to run well on a Celeron N3350 with 4GB RAM.)

The WiFi module appears to be soldered to the main PCB: not removable
without a hot air gun.  Seems strange that the same model PC would have
two different versions of the same family of modules.

Questions:

1. is it possible to read what "RFE" type the two working units are?
2. is this RFE value derived from reading GPIO pins?  (i.e. is possible
   my type 6 is actually a type 2 or type 4; but with some Brisbane River
   crud shorting pins out?)

-- 
Stuart Longland (aka Redhatter, VK4MSL)

I haven't lost my mind...
  ...it's backed up on a tape somewhere.