Received-SPF: pass (google.com: best guess record for domain of linux-wireless-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20200110132142.7737-1-denis281089@gmail.com> <aab66c75-49ae-0955-03a8-f817685b0925@nbd.name>
In-Reply-To: <aab66c75-49ae-0955-03a8-f817685b0925@nbd.name>
From:   Denis Kalashnikov <denis281089@gmail.com>
Date:   Mon, 13 Jan 2020 17:13:12 +0300
Message-ID: <CAKoLU8P3MyauCUevcRHtzMj8HSZQLD-hYyEaZyRH9OfLye6knQ@mail.gmail.com>
Subject: Re: [PATCH v3] ath9k: fix firmware bug in some Mikrotik R11e-2HPnD cards
To:     Felix Fietkau <nbd@nbd.name>
Cc:     linux-wireless@vger.kernel.org, Kalle Valo <kvalo@codeaurora.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

I read value from the SREV register of AR9582 -- got 0x001c04ff. I
looked through the code of ath9k and found out that in this case the
revision is 4. So, since AR_SREV_REVISION_9580_10 is defined as 4, I
logically concluded that AR_SREV_9580_10 would do the job. But you
suspicions turned to be right. I tested AR9580 now and it has the same
value as AR9582 in the SREV register -- 0x001c04ff. PCI Subsystem ID
is zero on those Mikrotik cards, so it is not useful. Felix, guys, may
be somebody have any thoughts how we can distinguish AR9582? Or that a
chip has only 2 chains? May be values from other registers?

=D1=81=D0=B1, 11 =D1=8F=D0=BD=D0=B2. 2020 =D0=B3. =D0=B2 13:37, Felix Fietk=
au <nbd@nbd.name>:
>
> On 2020-01-10 14:21, Denis Kalashnikov wrote:
> > Some of the Mikrotik R11e-2HPnD cards have EEPROM where is
> > flashed that a card has 3 chains, but actually all this cards
> > have only 2. This leads ath9k to write into the logs:
> > 'ath: phy0: Unable to reset channel, reset status -5' and
> > stations don't see that AP.
> >
> > Mikrotik R11e-2HPnD is based on AR9582 chip.
> >
> > Signed-off-by: Denis Kalashnikov <denis281089@gmail.com>
> > ---
> >
> > Changelog:
> >
> > Changes since v2:
> >   Use macros AR_SREV_9580_10 to select only AR9582 chips (they all
> >   have only 2 chains) instead of more general macros AR_SREV_9580
> >   that selects not only AR9582 but also AR9580 chips with 3 chains.
> I don't think you can detect AR9582 based on SREV. Where did you get
> this information from?
>
> - Felix