Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 14 Sep 2021 17:03:50 +0200
From:   Maxime Ripard <maxime@cerno.tech>
To:     dri-devel@lists.freedesktop.org,
        Daniel Vetter <daniel.vetter@intel.com>,
        David Airlie <airlied@linux.ie>,
        Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
        Thomas Zimmermann <tzimmermann@suse.de>,
        Emma Anholt <emma@anholt.net>,
        linux-rpi-kernel@lists.infradead.org,
        Dave Stevenson <dave.stevenson@raspberrypi.com>,
        Phil Elwell <phil@raspberrypi.com>,
        Tim Gover <tim.gover@raspberrypi.com>,
        Dom Cobley <dom@raspberrypi.com>,
        Sam Ravnborg <sam@ravnborg.org>,
        Nicolas Saenz Julienne <nsaenz@kernel.org>,
        bcm-kernel-feedback-list@broadcom.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 3/3] drm/vc4: hdmi: Actually check for the connector
 status in hotplug
Message-ID: <20210914150350.b6357jq2azalme4w@gilmour>
References: <20210914101724.266570-1-maxime@cerno.tech>
 <20210914101724.266570-3-maxime@cerno.tech>
 <YUCy4AmYDFD2jtG6@phenom.ffwll.local>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha256;
        protocol="application/pgp-signature"; boundary="omqsbnjfiweckfg7"
Content-Disposition: inline
In-Reply-To: <YUCy4AmYDFD2jtG6@phenom.ffwll.local>
Precedence: bulk


--omqsbnjfiweckfg7
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Hi Daniel,

On Tue, Sep 14, 2021 at 04:34:08PM +0200, Daniel Vetter wrote:
> On Tue, Sep 14, 2021 at 12:17:24PM +0200, Maxime Ripard wrote:
> > The drm_helper_hpd_irq_event() documentation states that this function
> > is "useful for drivers which can't or don't track hotplug interrupts for
> > each connector." and that "Drivers which support hotplug interrupts for
> > each connector individually and which have a more fine-grained detect
> > logic should bypass this code and directly call
> > drm_kms_helper_hotplug_event()". This is thus what we ended-up doing.
> >=20
> > However, what this actually means, and is further explained in the
> > drm_kms_helper_hotplug_event() documentation, is that
> > drm_kms_helper_hotplug_event() should be called by drivers that can
> > track the connection status change, and if it has changed we should call
> > that function.
> >=20
> > This underlying expectation we failed to provide is that the caller of
> > drm_kms_helper_hotplug_event() should call drm_helper_probe_detect() to
> > probe the new status of the connector.
> >=20
> > Since we didn't do it, it meant that even though we were sending the
> > notification to user-space and the DRM clients that something changed we
> > never probed or updated our internal connector status ourselves.
> >=20
> > This went mostly unnoticed since the detect callback usually doesn't
> > have any side-effect. Also, if we were using the DRM fbdev emulation
> > (which is a DRM client), or any user-space application that can deal
> > with hotplug events, chances are they would react to the hotplug event
> > by probing the connector status eventually.
> >=20
> > However, now that we have to enable the scrambler in detect() if it was
> > enabled it has a side effect, and an application such as Kodi or
> > modetest doesn't deal with hotplug events. This resulted with a black
> > screen when Kodi or modetest was running when a screen was disconnected
> > and then reconnected, or switched off and on.
>=20
> Uh, why are you running this scrambler restore in your probe function? I
> guess it works, but most drivers that do expensive hotplug restore to
> handle the "no black screen for replug" use-case handle that in their own
> dedicated code.

That's what I got from our discussion back in June (I think?). The
discussion was about an issue we were having back then where one would
disconnect / reconnect the display while it had been setup through SCDC
to use the scrambler and higher bit ratio.

During that power cycle (that also happens when you turn a TV off and
on), the display would obviously reset its SCDC status. But if there's
nothing to handle the uevent, then the same mode remains applied
resulting in a blank screen.

If we're not supposed to set the SCDC status again in detect(), how
would we deal with this?

> But those also tend to have per-output hpd interrupt sources, so maybe
> that's why?

We do have an per-output HPD interrupt here

Maximey

--omqsbnjfiweckfg7
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----

iHUEABYIAB0WIQRcEzekXsqa64kGDp7j7w1vZxhRxQUCYUC51gAKCRDj7w1vZxhR
xcwiAPwOelbHH/acgSu/HJHnxZj2wwWMcLHTWegRB50IMyBHDwD/ZfXzZGwXujvG
F98415jDZnBjFchPKB9vJkeHXlCeegk=
=LDfj
-----END PGP SIGNATURE-----

--omqsbnjfiweckfg7--