Message-ID: <CAD=FV=XfvD1OniNBrCrA8C6XjOB15fye8EdnniNmgpu4DnpH6w@mail.gmail.com>
Date: Mon, 22 Jul 2024 08:49:22 -0700
From: Doug Anderson <dianders@...omium.org>
To: neil.armstrong@...aro.org
Cc: Stephan Gerhold <stephan.gerhold@...aro.org>, Bjorn Andersson <andersson@...nel.org>, 
	Konrad Dybcio <konrad.dybcio@...aro.org>, Jessica Zhang <quic_jesszhan@...cinc.com>, 
	Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>, 
	dri-devel@...ts.freedesktop.org, devicetree@...r.kernel.org, 
	linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org, 
	Abel Vesa <abel.vesa@...aro.org>, Johan Hovold <johan@...nel.org>
Subject: Re: [PATCH v2 2/4] Revert "drm/panel-edp: Add SDC ATNA45AF01"

Hi,

On Mon, Jul 15, 2024 at 6:51 AM Doug Anderson <dianders@...omium.org> wrote:
>
> Hi,
>
> On Mon, Jul 15, 2024 at 6:02 AM Neil Armstrong
> <neil.armstrong@...aro.org> wrote:
> >
> > On 15/07/2024 14:54, Stephan Gerhold wrote:
> > > On Mon, Jul 15, 2024 at 02:42:12PM +0200, Neil Armstrong wrote:
> > >> On 15/07/2024 14:15, Stephan Gerhold wrote:
> > >>> This reverts commit 8ebb1fc2e69ab8b89a425e402c7bd85e053b7b01.
> > >>>
> > >>> The panel should be handled through the samsung-atna33xc20 driver for
> > >>> correct power up timings. Otherwise the backlight does not work correctly.
> > >>>
> > >>> We have existing users of this panel through the generic "edp-panel"
> > >>> compatible (e.g. the Qualcomm X1E80100 CRD), but the screen works only
> > >>> partially in that configuration: It works after boot but once the screen
> > >>> gets disabled it does not turn on again until after reboot. It behaves the
> > >>> same way with the default "conservative" timings, so we might as well drop
> > >>> the configuration from the panel-edp driver. That way, users with old DTBs
> > >>> will get a warning and can move to the new driver.
> > >>>
> > >>> Reviewed-by: Douglas Anderson <dianders@...omium.org>
> > >>> Signed-off-by: Stephan Gerhold <stephan.gerhold@...aro.org>
> > >>> ---
> > >>>    drivers/gpu/drm/panel/panel-edp.c | 2 --
> > >>>    1 file changed, 2 deletions(-)
> > >>>
> > >>> diff --git a/drivers/gpu/drm/panel/panel-edp.c b/drivers/gpu/drm/panel/panel-edp.c
> > >>> index 3a574a9b46e7..d2d682385e89 100644
> > >>> --- a/drivers/gpu/drm/panel/panel-edp.c
> > >>> +++ b/drivers/gpu/drm/panel/panel-edp.c
> > >>> @@ -1960,8 +1960,6 @@ static const struct edp_panel_entry edp_panels[] = {
> > >>>     EDP_PANEL_ENTRY('L', 'G', 'D', 0x05af, &delay_200_500_e200_d200, "Unknown"),
> > >>>     EDP_PANEL_ENTRY('L', 'G', 'D', 0x05f1, &delay_200_500_e200_d200, "Unknown"),
> > >>> -   EDP_PANEL_ENTRY('S', 'D', 'C', 0x416d, &delay_100_500_e200, "ATNA45AF01"),
> > >>> -
> > >>>     EDP_PANEL_ENTRY('S', 'H', 'P', 0x1511, &delay_200_500_e50, "LQ140M1JW48"),
> > >>>     EDP_PANEL_ENTRY('S', 'H', 'P', 0x1523, &delay_80_500_e50, "LQ140M1JW46"),
> > >>>     EDP_PANEL_ENTRY('S', 'H', 'P', 0x153a, &delay_200_500_e50, "LQ140T1JH01"),
> > >>>
> > >>
> > >> How will we handle current/old crd DT with new kernels ?
> > >>
> > >
> > > I think this is answered in the commit message:
> > >
> > >>> We have existing users of this panel through the generic "edp-panel"
> > >>> compatible (e.g. the Qualcomm X1E80100 CRD), but the screen works only
> > >>> partially in that configuration: It works after boot but once the screen
> > >>> gets disabled it does not turn on again until after reboot. It behaves the
> > >>> same way with the default "conservative" timings, so we might as well drop
> > >>> the configuration from the panel-edp driver. That way, users with old DTBs
> > >>> will get a warning and can move to the new driver.
> > >
> > > Basically with the entry removed, the panel-edp driver will fallback to
> > > default "conservative" timings when using old DTBs. There will be a
> > > warning in dmesg, but otherwise the panel will somewhat work just as
> > > before. I think this is a good way to remind users to upgrade.
> >
> > I consider this as a regression
> >
> > >
> > >> Same question for patch 3, thie serie introduces a bindings that won't be valid
> > >> if we backport patch 3. I don't think patch should be backported, and this patch
> > >> should be dropped.
> > >
> > > There would be a dtbs_check warning, yeah. Functionally, it would work
> > > just fine. Is that reason enough to keep display partially broken for
> > > 6.11? We could also apply the minor binding change for 6.11 if needed.
> >
> > I don't know how to answer this, I'll let the DT maintainer comment this.
> >
> > The problem is I do not think we can pass the whole patchset as fixes
> > for v6.11, patches 2 & 3 could, patches 1 & 4 definitely can't.
> >
> > Neil
>
> IMO: patch #3 (dts) and #4 (CONFIG) go through the Qualcomm tree
> whenever those folks agree to it. If we're worried about the
> dtbs_check breakage I personally wouldn't mind "Ack"ing patch #1 to go
> through the Qualcomm tree as long as it made it into 6.11-rc1. I have
> a hunch that there are going to be more Samsung OLED panels in the
> future that will need to touch the same file, but if the change is in
> -rc1 it should make it back into drm-misc quickly, right?
>
> Personally I think patch #2 could go in anytime since, as people have
> said, things are pretty broken today and the worst that happens is
> that someone gets an extra warning. That would be my preference. That
> being said, we could also snooze that patch for a month or two and
> land it later. There's no real hurry.

For now I'm going to snooze this patch for a month just to avoid any
controversy. I'll plan to apply it (to drm-misc-next) when I see the
device tree patch land. Since the device tree patch should land as a
fix that should keep things landing in the correct order. ...and, as
per above, the worst case is that if someone has an old DTS and a new
kernel then a panel that was already not working well will print a fat
warning and startup a bit slower.

If somehow I mess up and forget about this patch, feel free to send me
a poke when the device tree patch is landed.

-Doug

Powered by blists - more mailing lists