[0/3] wifi: rt2x00: various improvements for MT7620 initialization

Message

Shiji Yang Oct. 13, 2023, 11:21 p.m. UTC

Hi!
This series patches aim to improve the initialization of MT7620. All
changes and register values were ported from the Mediatek vendor
driver, and I have been testing them on OpenWrt for several weeks.
Regards,
  Shiji Yang


Shiji Yang (3):
  wifi: rt2x00: improve MT7620 register initialization
  wifi: rt2x00: rework MT7620 channel config function
  wifi: rt2x00: rework MT7620 PA/LNA RF calibration

 drivers/net/wireless/ralink/rt2x00/rt2800.h   |  18 ++
 .../net/wireless/ralink/rt2x00/rt2800lib.c    | 290 +++++++++++-------
 .../net/wireless/ralink/rt2x00/rt2800mmio.c   |   3 +
 drivers/net/wireless/ralink/rt2x00/rt2x00.h   |   6 +
 4 files changed, 210 insertions(+), 107 deletions(-)

Comments

Stanislaw Gruszka Oct. 14, 2023, 5:54 p.m. UTC | #1

On Sat, Oct 14, 2023 at 07:21:27AM +0800, Shiji Yang wrote:
> This series patches aim to improve the initialization of MT7620. All
> changes and register values were ported from the Mediatek vendor
> driver, and I have been testing them on OpenWrt for several weeks.

Hi Shiji!

Thank you for improving MT7620 support in rt2x00!
Your patches generally look ok to me.

I would suggest however consider to move support of MT7620 to
mt76 driver. It's much nicer and better written driver with
better mac80211/linux integration and support of various
features that rt2x00 lacks. Adding mt7620 sub-driver there
should not be that hard, and would be much "cooler" project
for you in my opinion.

This of course depend of how much time do you have to work
on this. Tweaking/improving rt2x00 can be easier and faster
way to achieve your goals and I don't want discourage 
you from doing this, if that's something you prefer.
Keep good work!

Regards
Stanislaw

Shiji Yang Oct. 16, 2023, 12:45 p.m. UTC | #2

On Sat, 14 Oct 2023 19:54:28 +0200, Stanislaw Gruszka wrote:
>On Sat, Oct 14, 2023 at 07:21:27AM +0800, Shiji Yang wrote:
>> This series patches aim to improve the initialization of MT7620. All
>> changes and register values were ported from the Mediatek vendor
>> driver, and I have been testing them on OpenWrt for several weeks.
>
>Hi Shiji!
>
>Thank you for improving MT7620 support in rt2x00!
>Your patches generally look ok to me.
>
>I would suggest however consider to move support of MT7620 to
>mt76 driver. It's much nicer and better written driver with
>better mac80211/linux integration and support of various
>features that rt2x00 lacks. Adding mt7620 sub-driver there
>should not be that hard, and would be much "cooler" project
>for you in my opinion.
>
>This of course depend of how much time do you have to work
>on this. Tweaking/improving rt2x00 can be easier and faster
>way to achieve your goals and I don't want discourage 
>you from doing this, if that's something you prefer.
>Keep good work!
>
>Regards
>Stanislaw
>

Hi!
Thanks for your quick review.

Porting MT7620 support to mt76 is definitely a good idea, but to
be honest I don't have much interesting to do that. Overall, there
are two reasons that hinder me. First I'm not familiar with linux
net driver development. This is the main reason. And then it is
difficult to analyze the MT7620 vendor driver. There are too many
dead code and macro switches.

rt2x00 is already stable enough for me (except watchdog reset, it
still requires some work). Thank you all for your contributions
to the rt2x00. It's really a hard to job to provide an usable open
source driver ten years ago.

Regards,
Shiji Yang