Re: [PATCH wireless-drivers] mt76x0: eeprom: fix chan_vs_power map in mt76x0_get_power_info

From: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
To: Kalle Valo <kvalo@codeaurora.org>
Cc: linux-wireless <linux-wireless@vger.kernel.org>,
	Felix Fietkau <nbd@nbd.name>,
	Stanislaw Gruszka <sgruszka@redhat.com>
Subject: Re: [PATCH wireless-drivers] mt76x0: eeprom: fix chan_vs_power map in mt76x0_get_power_info
Date: Tue, 29 Jan 2019 08:28:11 +0100	[thread overview]
Message-ID: <CAJ0CqmVPK6R45F511Bnc4EXGFkEnwRsYhPTGpfvy59v8RbsOBg@mail.gmail.com> (raw)
In-Reply-To: <87bm40ym88.fsf@codeaurora.org>

>
> Lorenzo Bianconi <lorenzo.bianconi@redhat.com> writes:
>
> >> Lorenzo Bianconi <lorenzo.bianconi@redhat.com> writes:
> >>
> >> > Report correct eeprom per channel power value.
> >> > Fix chan_vs_power map in mt76x0_get_power_info routine
> >> >
> >> > Fixes: f2a2e819d672 ("mt76x0: remove eeprom dependency from mt76x0_get_power_info")
> >> > Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
> >>
> >> So what's the actual bug this fixes? The commit log is not really clear
> >> on that. Wondering because you marked this for wireless-drivers.
> >
> > Hi Kalle,
> >
> > You are right, I have been not so clear in the commit log.
> > Offsets in chan_map were wrong since mt76x02_eeprom_get() reads at even addresses.
> > Moreover 'if' condition in the for loop was wrong (chan->hw_value and
> > chan_map[i].chan were swapped)
>
> But how does this affect from user's point of view? I mean what feature
> is broken and how does this fix that?
>
> --
> Kalle Valo

mt76x0_get_power_info() is used to read from eeprom the global max tx
power according to the operating channel.
mt76x02_eeprom_get() reads just at even addresses and it reports -1
for odd ones (so offset must be a even).
This issue will affect the tx power that userspace is able to configure.
Do you think it is better to merge it in wireless-drivers-next?

Regards,
Lorenzo