RE: [PATCH] rtw88: add TX power limit support to 114 and 130 channels

[Ping-Ke Shih <pkshih@realtek.com>]

Thadeu Lima de Souza Cascardo <cascardo@igalia.com> wrote:
> On Mon, Apr 13, 2026 at 05:56:17AM +0000, Ping-Ke Shih wrote:
> > Thadeu Lima de Souza Cascardo <cascardo@igalia.com> wrote:
> > > On Fri, Apr 10, 2026 at 03:56:11AM +0000, Ping-Ke Shih wrote:
> > > > Thadeu Lima de Souza Cascardo <cascardo@igalia.com> wrote:
> > > > > Though 114 and 130 are not usual channels, they are found in the wild with
> > > > > setups using 5350MHz as the center frequency of a 80MHz setup.
> > > >
> > > > What did the AP setup? channel 114 160MHz?
> > > > I wonder why rtw88 can select a not usual channel 114 80MHz.
> > > >
> > > > Please share your environment setup.
> > > >
> > >
> > > This is a Mikrotik that uses channel 130 at 80MHz.
> >
> > I'm surprised that an AP can work on this not usual channel/bandwidth.
> > Can you change the setting to usual channel/bandwidth? We'd avoid using
> > this unsupported channel/bandwidth by [1].
> >
> 
> It seems to be "well-known" that some APs do it and it has caused other
> issues in other drivers. But it works just fine with a lot of other
> drivers. So, I would rather make rtw88 work with that setting, even if it
> falls back to 40MHz, for example.
> 
> Let me check what happens when using the proposed patch and I will report
> back to you.
> 
> > >
> > > > >
> > > > > rtw88 supports that, but issues a WARNING because it cannot find the TX
> > > > > power limit for those channels.
> > > >
> > > > Actually, rtw88 hardware can't support that, so we are working on patch
> > > > to avoid selecting unusual channels. Can it work properly with
> > > > the AP after this patch?
> > > >
> > >
> > > It does work just fine even without the patch. The only issue is the
> > > WARNING that is triggered.
> > >
> >
> > As internal discussion, hardware doesn't work on channel 130 80M,
> > which means connection might be well, but it can't yield expected performance.
> > More, the power limit is not really verified on ch130 80M, so we wonder
> > that the signal might not in expectation.
> >
> 
> The rationale behing my patch was exactly to try to set the power limit to
> the lesser one of the adjacent channels. 

The hardware setting of channel 130 80M is unknown, since we have never
verified that, so I can't say use the lesser one is always correct. 
(Personally I'd make this choice if I have no more information from internal
hardware experts.)

> I am not sure that the expected
> performance was not achieved. How would you suggest that I verify?

I think you can check if RX rate is 80MHz or not. If it really works
properly on 80MHz, you can yield TX/RX UDP throughput about 0.7 *
PHY rates.

Ping-Ke