Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 17 Nov 2022 21:00:07 +0000
Message-ID: <86tu2xnvoo.wl-maz@kernel.org>
From:   Marc Zyngier <maz@kernel.org>
To:     Nicolas Frayer <nfrayer@baylibre.com>
Cc:     tglx@linutronix.de, linux-kernel@vger.kernel.org,
        khilman@baylibre.com, glaroque@baylibre.com
Subject: Re: [PATCH 0/2] irqchip: Kconfig: Add module support for TI inta/intr
In-Reply-To: <CANyCTtTfhhORL32EweGN2Gtx9=kpA+PAxTUKE84C5sGmjTfPyw@mail.gmail.com>
References: <20221116185500.40431-1-nfrayer@baylibre.com>
        <864juypoax.wl-maz@kernel.org>
        <CANyCTtTfhhORL32EweGN2Gtx9=kpA+PAxTUKE84C5sGmjTfPyw@mail.gmail.com>
User-Agent: Wanderlust/2.15.9 (Almost Unreal) SEMI-EPG/1.14.7 (Harue)
 FLIM-LB/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL-LB/10.8 EasyPG/1.0.0 Emacs/27.1
 (aarch64-unknown-linux-gnu) MULE/6.0 (HANACHIRUSATO)
MIME-Version: 1.0 (generated by SEMI-EPG 1.14.7 - "Harue")
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, 17 Nov 2022 20:49:58 +0000,
Nicolas Frayer <nfrayer@baylibre.com> wrote:
>=20
> Le mer. 16 nov. 2022 =C3=A0 22:44, Marc Zyngier <maz@kernel.org> a =C3=A9=
crit :
> >
> > On Wed, 16 Nov 2022 18:54:58 +0000,
> > Nicolas Frayer <nfrayer@baylibre.com> wrote:
> > >
> > > Added module support for TI interrupt aggregator and interrupt router
> > >
> > > Nicolas Frayer (2):
> > >   irqchip: Kconfig: module build support for the TI interrupt router
> > >     driver
> > >   irqchip: Kconfig: Added module build support for the TI interrupt
> > >     aggregator
> > >
> > >  arch/arm64/Kconfig.platforms |  2 --
> > >  drivers/irqchip/Kconfig      | 12 +++++++-----
> > >  2 files changed, 7 insertions(+), 7 deletions(-)
> >
> > How did you test that it doesn't introduce any regression due to
> > implicit ordering requirements that are now voided?
> Hi Marc,
> I tested on a TI AM62x SK board.
> By default the drivers are set to ARCH_K3 (built-in).

That wasn't exactly my question: even if they are by sefault set as
built-in, you still run the risk of exposing some breakage for
endpoint drivers that expect the irqchip to be present before they get
initialised. Every time an interrupt controller gets modularised, we
end-up with a ton of failures.

A good way to test this is to run a distro kernel with all the drivers
build as modules.

> >
> > Please Cc the TI folks so that they can at least check this.
> I'll resend the series with more description and CC Ti engineers.

Please also see the kernel robot result: this stuff now gets compiled
on unsuspecting architectures, and fails.

	M.

--=20
Without deviation from the norm, progress is not possible.