MIME-Version: 1.0
References: <CAJ+vNU0TuNA26F+hFwTRGc2pVvW34-7Sc7oQ9EW8V+2cVFgcag@mail.gmail.com>
 <CAD8Lp44c1k-xP1U9X8PR_kw=8joMY8RUwcKVBvuWaV7faNhWZg@mail.gmail.com>
In-Reply-To: <CAD8Lp44c1k-xP1U9X8PR_kw=8joMY8RUwcKVBvuWaV7faNhWZg@mail.gmail.com>
From:   Andy Shevchenko <andy.shevchenko@gmail.com>
Date:   Mon, 20 Mar 2023 14:18:17 +0200
Message-ID: <CAHp75VeHxZqqvuJXYet=Nt1GA2n_Up23Uf+p7LidY8=fvbvxxQ@mail.gmail.com>
Subject: Re: ath9k legacy vs MSI interrupt issue on imx8mm soc
To:     Daniel Drake <drake@endlessos.org>
Cc:     Tim Harvey <tharvey@gateworks.com>,
        linux-wireless <linux-wireless@vger.kernel.org>,
        rhu@qti.qualcomm.com, ryanhsu@qti.qualcomm.com,
        rwchang@qti.qualcomm.com, aeolus@qti.qualcomm.com,
        "Valo, Kalle" <kvalo@qca.qualcomm.com>,
        ath9k-devel <ath9k-devel@qca.qualcomm.com>, linux@endlessm.com,
        rafael.j.wysocki@intel.com, andy@infradead.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Mar 20, 2023 at 1:29=E2=80=AFPM Daniel Drake <drake@endlessos.org> =
wrote:
>
> On Sat, Mar 18, 2023 at 8:17=E2=80=AFAM Tim Harvey <tharvey@gateworks.com=
> wrote:
> > It seems that Daniel may have been running into something similar back
> > when commit 7368160f0ab0 ("ath9k: add MSI support") was merged
> > [1][2][3]
> >
> > Does anyone have any advice here about what can be going on?
>
> The issue we found earlier was that ath9k MSI is broken unless the MSI
> IRQ number coincidentally is aligned to 4.
>
> Worked around with:
> https://github.com/endlessm/linux/commit/0a2d86a2b9da8f936d873addf0a97e34=
1177fe8a
> https://github.com/endlessm/linux/commit/970d2b4b59465cd926e5213e0feb633f=
6b623447
>
> IIRC the PCI patch wasn't accepted upstream at the time

Briefly looking into it, it won't be accepted in the future either (in
the current form).
The solution there as far as I (possibly mistakenly) can see is that
it's only for x86 cases. For ARM or others there is no explanation on
what would be the effect of the MSI enabling. Hence, it's not widely
tested, nor root caused: May it be the certain x86 PCI bridge / switch
issue? May it be the APIC related issue on that x86 platform? Is it a
bug in Atheros FW/HW?

> and I never
> got any response from Atheros/Qualcomm about this device
> hardware/firmware bug.

--=20
With Best Regards,
Andy Shevchenko