Re: [PATCH v2] drm/probe-helper: Default to 640x480 if no EDID

From: Doug Anderson <dianders@chromium.org>
To: Thomas Zimmermann <tzimmermann@suse.de>
Cc: dri-devel <dri-devel@lists.freedesktop.org>,
	Sankeerth Billakanti <quic_sbillaka@quicinc.com>,
	LKML <linux-kernel@vger.kernel.org>,
	David Airlie <airlied@linux.ie>,
	linux-arm-msm <linux-arm-msm@vger.kernel.org>,
	"Abhinav Kumar (QUIC)" <quic_abhinavk@quicinc.com>,
	Stephen Boyd <swboyd@chromium.org>,
	Dmitry Baryshkov <dmitry.baryshkov@linaro.org>,
	"Aravind Venkateswaran (QUIC)" <quic_aravindh@quicinc.com>,
	"Kuogee Hsieh (QUIC)" <quic_khsieh@quicinc.com>,
	freedreno <freedreno@lists.freedesktop.org>
Subject: Re: [PATCH v2] drm/probe-helper: Default to 640x480 if no EDID
Date: Fri, 13 May 2022 13:07:35 -0700	[thread overview]
Message-ID: <CAD=FV=VbwkK-z8T-98aPSiybd2c94n8p46oBxY_MtPjV608YRQ@mail.gmail.com> (raw)
In-Reply-To: <CAD=FV=WoSTcSOB_reDbayNb=q7w00rd7p-zHUDt+evTkSjQ=2g@mail.gmail.com>

Hi,

On Wed, May 11, 2022 at 2:32 PM Doug Anderson <dianders@chromium.org> wrote:
>
> Hi,
>
> On Wed, May 11, 2022 at 12:14 AM Thomas Zimmermann <tzimmermann@suse.de> wrote:
> >
> > Hi
> >
> > Am 10.05.22 um 22:51 schrieb Douglas Anderson:
> > > If we're unable to read the EDID for a display because it's corrupt /
> > > bogus / invalid then we'll add a set of standard modes for the
> > > display. When userspace looks at these modes it doesn't really have a
> > > good concept for which mode to pick and it'll likely pick the highest
> > > resolution one by default. That's probably not ideal because the modes
> > > were purely guesses on the part of the Linux kernel.
> >
> > I'm skeptical. Why does the kernel do a better job than userspace here?
> > Only the graphics driver could possibly make such a decision.
> >
> > Not setting any preferred mode at least gives a clear message to userspace.
>
> OK, that's a fair point. So I tried to find out what our userspace is
> doing. I believe it's:
>
> https://source.chromium.org/chromium/chromium/src/+/main:ui/ozone/platform/drm/common/drm_util.cc;l=529
>
> Specifically this bit of code:
>
>   // If we still have no preferred mode, then use the first one since it should
>   // be the best mode.
>   if (!*out_native_mode && !modes.empty())
>     *out_native_mode = modes.front().get();
>
> Do you agree with what our userspace is doing here, or is it wrong?
>
> If our userspace is doing the right thing, then I guess the problem is
> the call to "drm_mode_sort(&connector->modes);" at the end of
> drm_helper_probe_single_connector_modes(). Would you be OK with me
> _not_ sorting the modes in the "bad EDID" case? That also seems to fix
> my problem...

I've implemented the "don't mark preferred, but don't sort" as a v3.
Hopefully it looks good.

https://lore.kernel.org/r/20220513130533.v3.1.I31ec454f8d4ffce51a7708a8092f8a6f9c929092@changeid

-Doug