[PATCH v1 0/2] Add EDP_PANEL_ENTRY3 to override bpc

[PATCH v1 0/2] Add EDP_PANEL_ENTRY3 to override bpc

[Ajye Huang]

v1:
- EDP_PANEL_ENTRY3: 
  - "override_bpc" is used for override the value of panel's bpc.

Thanks for the review!

Ajye Huang (2):
  drm/panel-edp: Add override bpc quirk for generic edp
  drm/panel-edp: Modify LQ116M1JW10 panel's bpc to 6

 drivers/gpu/drm/panel/panel-edp.c | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

-- 
2.25.1

[PATCH v1 1/2] drm/panel-edp: Add override bpc quirk for generic edp

[Ajye Huang]

Adding override bpc to EDP_PANEL_ENTRY3 quirk.

Signed-off-by: Ajye Huang <ajye_huang@compal.corp-partner.google.com>
---
 drivers/gpu/drm/panel/panel-edp.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/gpu/drm/panel/panel-edp.c b/drivers/gpu/drm/panel/panel-edp.c
index 944c7c70de55..da3e8f223ec3 100644
--- a/drivers/gpu/drm/panel/panel-edp.c
+++ b/drivers/gpu/drm/panel/panel-edp.c
@@ -218,6 +218,9 @@ struct edp_panel_entry {
 
 	/** @override_edid_mode: Override the mode obtained by edid. */
 	const struct drm_display_mode *override_edid_mode;
+
+	/** @override_bpc: Override the Bits per color obtained by edid. */
+	unsigned int override_bpc;
 };
 
 struct panel_edp {
@@ -586,6 +589,9 @@ static int panel_edp_get_modes(struct drm_panel *panel,
 	bool has_override_edid_mode = p->detected_panel &&
 				      p->detected_panel != ERR_PTR(-EINVAL) &&
 				      p->detected_panel->override_edid_mode;
+	bool has_override_bpc = p->detected_panel &&
+				      p->detected_panel != ERR_PTR(-EINVAL) &&
+				      p->detected_panel->override_bpc;
 
 	/* probe EDID if a DDC bus is available */
 	if (p->ddc) {
@@ -611,6 +617,9 @@ static int panel_edp_get_modes(struct drm_panel *panel,
 			} else {
 				num += drm_edid_connector_add_modes(connector);
 			}
+
+			if (has_override_bpc)
+				connector->display_info.bpc = p->detected_panel->override_bpc;
 		}
 
 		pm_runtime_mark_last_busy(panel->dev);
@@ -1871,6 +1880,17 @@ static const struct panel_delay delay_80_500_e50_d50 = {
 	.override_edid_mode = _mode \
 }
 
+#define EDP_PANEL_ENTRY3(vend_chr_0, vend_chr_1, vend_chr_2, product_id, _delay, _name, _bpc) \
+{ \
+	.ident = { \
+		.name = _name, \
+		.panel_id = drm_edid_encode_panel_id(vend_chr_0, vend_chr_1, vend_chr_2, \
+						     product_id), \
+	}, \
+	.delay = _delay, \
+	.override_bpc = _bpc \
+}
+
 /*
  * This table is used to figure out power sequencing delays for panels that
  * are detected by EDID. Entries here may point to entries in the
-- 
2.25.1

[PATCH v1 2/2] drm/panel-edp: Modify LQ116M1JW10 panel's bpc to 6

[Ajye Huang]

The link training is failed when bpc value is 8.
It sure seems like the panel simply doesn't like 8bpp,
Changing the bpc to 6 allows link training to succeed.

The 8bpc log shows that link training failed.
6bpc
----
rate_mhz: 1620
valid rates: 30
bit_rate_khz: 2399760, dp_rate_mhz: 1500, ti_sn_bridge_calc_min_dp_rate_idx return: 1

8bpc
----
rate_mhz: 2160
valid rates: 30
bit_rate_khz: 3199680, dp_rate_mhz: 2000, ti_sn_bridge_calc_min_dp_rate_idx return: 2
Link training failed, link is off.
Disable the PLL if we failed.

Signed-off-by: Ajye Huang <ajye_huang@compal.corp-partner.google.com>
---
 drivers/gpu/drm/panel/panel-edp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/panel/panel-edp.c b/drivers/gpu/drm/panel/panel-edp.c
index da3e8f223ec3..13755168cd75 100644
--- a/drivers/gpu/drm/panel/panel-edp.c
+++ b/drivers/gpu/drm/panel/panel-edp.c
@@ -2094,7 +2094,7 @@ static const struct edp_panel_entry edp_panels[] = {
 	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"),
-	EDP_PANEL_ENTRY('S', 'H', 'P', 0x154c, &delay_200_500_p2e100, "LQ116M1JW10"),
+	EDP_PANEL_ENTRY3('S', 'H', 'P', 0x154c, &delay_200_500_p2e100, "LQ116M1JW10", 6),
 	EDP_PANEL_ENTRY('S', 'H', 'P', 0x158f, &delay_200_500_p2e100, "LQ134Z1"),
 	EDP_PANEL_ENTRY('S', 'H', 'P', 0x1593, &delay_200_500_p2e100, "LQ134N1"),
 
-- 
2.25.1

Re: [PATCH v1 1/2] drm/panel-edp: Add override bpc quirk for generic edp

[Doug Anderson]

Hi,

On Wed, Oct 29, 2025 at 1:11 AM Ajye Huang
<ajye_huang@compal.corp-partner.google.com> wrote:
>
> Adding override bpc to EDP_PANEL_ENTRY3 quirk.
>
> Signed-off-by: Ajye Huang <ajye_huang@compal.corp-partner.google.com>
> ---
>  drivers/gpu/drm/panel/panel-edp.c | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)

This seems OK to me. I'd be OK with:

Reviewed-by: Douglas Anderson <dianders@chromium.org>


I think you could alternatively add a EDID_QUIRK_FORCE_6BPC quirk for
this panel in "drivers/gpu/drm/drm_edid.c", though I haven't tested
it. That _might_ be a better solution? Maybe Jani or someone else CCed
would have an opinion.

At first I was thinking that the quirks in "drm_edid.c" were probably
just for "DP" display, but then I just realized that they probably
also are for "eDP" panels. Specifically I think Intel hardware doesn't
use panel-edp.c so I think the only place quirks could get applied (if
an eDP panel was also used on Intel hardware) was from "drm_edid.c".

Any chance you could confirm if EDID_QUIRK_FORCE_6BPC works for you?

Does anyone else CCed have an opinion of which they like better?

-Doug

Re: [PATCH v1 2/2] drm/panel-edp: Modify LQ116M1JW10 panel's bpc to 6

[Doug Anderson]

Hi,

On Wed, Oct 29, 2025 at 1:11 AM Ajye Huang
<ajye_huang@compal.corp-partner.google.com> wrote:
>
> The link training is failed when bpc value is 8.
> It sure seems like the panel simply doesn't like 8bpp,
> Changing the bpc to 6 allows link training to succeed.
>
> The 8bpc log shows that link training failed.
> 6bpc
> ----
> rate_mhz: 1620
> valid rates: 30
> bit_rate_khz: 2399760, dp_rate_mhz: 1500, ti_sn_bridge_calc_min_dp_rate_idx return: 1
>
> 8bpc
> ----
> rate_mhz: 2160
> valid rates: 30
> bit_rate_khz: 3199680, dp_rate_mhz: 2000, ti_sn_bridge_calc_min_dp_rate_idx return: 2
> Link training failed, link is off.
> Disable the PLL if we failed.

Though I always appreciate details about the debugging that was done,
I suspect that most people reading this won't really understand unless
you give them the context that you are using the ti-sn65dsi86 bridge
chip and that you are printing out values related to bridge training.

I would also note that, to me, the more important test was confirming
that even when you pick the same "rate_mhz" for 6bpc and 8bpc that
6bpc works and 8bpc doesn't work.


> Signed-off-by: Ajye Huang <ajye_huang@compal.corp-partner.google.com>
> ---
>  drivers/gpu/drm/panel/panel-edp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Unless folks end up preferring EDID_QUIRK_FORCE_6BPC:

Reviewed-by: Douglas Anderson <dianders@chromium.org>

Re: [PATCH v1 1/2] drm/panel-edp: Add override bpc quirk for generic edp

[Ajye Huang]

Hi Doug,

On Thu, Oct 30, 2025 at 7:21 AM Doug Anderson <dianders@chromium.org> wrote:

> At first I was thinking that the quirks in "drm_edid.c" were probably
> just for "DP" display, but then I just realized that they probably
> also are for "eDP" panels. Specifically I think Intel hardware doesn't
> use panel-edp.c so I think the only place quirks could get applied (if
> an eDP panel was also used on Intel hardware) was from "drm_edid.c".
>
> Any chance you could confirm if EDID_QUIRK_FORCE_6BPC works for you?

The following quirks added in drm_edid.c is workable,
+       /* Sharp LQ116M1JW10 reports 8 bpc and it  will display noise,
but display noiseless as 6 bpc  */
+       EDID_QUIRK('S', 'H', 'P', 0x154c, EDID_QUIRK_FORCE_6BPC),

I will send upstream a new patch for modifying the drm_edid.c , thank
you so much

Re: [PATCH v1 1/2] drm/panel-edp: Add override bpc quirk for generic edp

[Jani Nikula]

On Wed, 29 Oct 2025, Ajye Huang <ajye_huang@compal.corp-partner.google.com> wrote:
> Adding override bpc to EDP_PANEL_ENTRY3 quirk.
>
> Signed-off-by: Ajye Huang <ajye_huang@compal.corp-partner.google.com>
> ---
>  drivers/gpu/drm/panel/panel-edp.c | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)
>
> diff --git a/drivers/gpu/drm/panel/panel-edp.c b/drivers/gpu/drm/panel/panel-edp.c
> index 944c7c70de55..da3e8f223ec3 100644
> --- a/drivers/gpu/drm/panel/panel-edp.c
> +++ b/drivers/gpu/drm/panel/panel-edp.c
> @@ -218,6 +218,9 @@ struct edp_panel_entry {
>  
>  	/** @override_edid_mode: Override the mode obtained by edid. */
>  	const struct drm_display_mode *override_edid_mode;
> +
> +	/** @override_bpc: Override the Bits per color obtained by edid. */
> +	unsigned int override_bpc;
>  };
>  
>  struct panel_edp {
> @@ -586,6 +589,9 @@ static int panel_edp_get_modes(struct drm_panel *panel,
>  	bool has_override_edid_mode = p->detected_panel &&
>  				      p->detected_panel != ERR_PTR(-EINVAL) &&
>  				      p->detected_panel->override_edid_mode;

Unrelated to the patch at hand, since the pattern is there already, but
something like this should be more reliable:

bool has_override_edid_mode = !IS_ERR_OR_NULL(p->detected_panel) &&
	p->detected_panel->override_edid_mode;

It does not look like p->detected_panel could have other error pointer
values than -EINVAL, but it looks awkward to check for NULL and one
error pointer value, and then go on to dereference it.

I guess even better would be to always use either NULL *or* error
pointers, not a mix, but I digress.

BR,
Jani.


> +	bool has_override_bpc = p->detected_panel &&
> +				      p->detected_panel != ERR_PTR(-EINVAL) &&
> +				      p->detected_panel->override_bpc;
>  
>  	/* probe EDID if a DDC bus is available */
>  	if (p->ddc) {
> @@ -611,6 +617,9 @@ static int panel_edp_get_modes(struct drm_panel *panel,
>  			} else {
>  				num += drm_edid_connector_add_modes(connector);
>  			}
> +
> +			if (has_override_bpc)
> +				connector->display_info.bpc = p->detected_panel->override_bpc;
>  		}
>  
>  		pm_runtime_mark_last_busy(panel->dev);
> @@ -1871,6 +1880,17 @@ static const struct panel_delay delay_80_500_e50_d50 = {
>  	.override_edid_mode = _mode \
>  }
>  
> +#define EDP_PANEL_ENTRY3(vend_chr_0, vend_chr_1, vend_chr_2, product_id, _delay, _name, _bpc) \
> +{ \
> +	.ident = { \
> +		.name = _name, \
> +		.panel_id = drm_edid_encode_panel_id(vend_chr_0, vend_chr_1, vend_chr_2, \
> +						     product_id), \
> +	}, \
> +	.delay = _delay, \
> +	.override_bpc = _bpc \
> +}
> +
>  /*
>   * This table is used to figure out power sequencing delays for panels that
>   * are detected by EDID. Entries here may point to entries in the

-- 
Jani Nikula, Intel

Re: [PATCH v1 1/2] drm/panel-edp: Add override bpc quirk for generic edp

[Jani Nikula]

On Wed, 29 Oct 2025, Doug Anderson <dianders@chromium.org> wrote:
> At first I was thinking that the quirks in "drm_edid.c" were probably
> just for "DP" display, but then I just realized that they probably
> also are for "eDP" panels. Specifically I think Intel hardware doesn't
> use panel-edp.c so I think the only place quirks could get applied (if
> an eDP panel was also used on Intel hardware) was from "drm_edid.c".
>
> Any chance you could confirm if EDID_QUIRK_FORCE_6BPC works for you?
>
> Does anyone else CCed have an opinion of which they like better?

No strong opinions, but obviously having it in drm_edid.c would be more
generic. Until you encounter panels where you need to force something
other than 6 bpc, when it stops being generic and becomes an
obstacle. ;)

BR,
Jani.

-- 
Jani Nikula, Intel

Re: [PATCH v1 2/2] drm/panel-edp: Modify LQ116M1JW10 panel's bpc to 6

[Ajye Huang]

Hi Doug,

On Thu, Oct 30, 2025 at 7:25 AM Doug Anderson <dianders@chromium.org> wrote:

>
> Unless folks end up preferring EDID_QUIRK_FORCE_6BPC:
>
> Reviewed-by: Douglas Anderson <dianders@chromium.org>

After following your suggestion with the following, the issue goes
away during YouTube playback.
I will send a new patch for drm_edid.c, thank you so much

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index e2e85345aa9a..a73d37fe7ea1 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -250,6 +250,9 @@ static const struct edid_quirk {
     EDID_QUIRK('S', 'V', 'R', 0x1019, BIT(EDID_QUIRK_NON_DESKTOP)),
     EDID_QUIRK('A', 'U', 'O', 0x1111, BIT(EDID_QUIRK_NON_DESKTOP)),

+    /* LQ116M1JW10 displays noise when 8 bpc, but display fine as 6 bpc */
+    EDID_QUIRK('S', 'H', 'P', 0x154c, EDID_QUIRK_FORCE_6BPC),
+
     /*
      * @drm_edid_internal_quirk entries end here, following with the
      * @drm_edid_quirk entries.

Re: [PATCH v1 1/2] drm/panel-edp: Add override bpc quirk for generic edp

[Doug Anderson]

Hi,

On Thu, Oct 30, 2025 at 2:44 AM Jani Nikula <jani.nikula@linux.intel.com> wrote:
>
> On Wed, 29 Oct 2025, Ajye Huang <ajye_huang@compal.corp-partner.google.com> wrote:
> > Adding override bpc to EDP_PANEL_ENTRY3 quirk.
> >
> > Signed-off-by: Ajye Huang <ajye_huang@compal.corp-partner.google.com>
> > ---
> >  drivers/gpu/drm/panel/panel-edp.c | 20 ++++++++++++++++++++
> >  1 file changed, 20 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/panel/panel-edp.c b/drivers/gpu/drm/panel/panel-edp.c
> > index 944c7c70de55..da3e8f223ec3 100644
> > --- a/drivers/gpu/drm/panel/panel-edp.c
> > +++ b/drivers/gpu/drm/panel/panel-edp.c
> > @@ -218,6 +218,9 @@ struct edp_panel_entry {
> >
> >       /** @override_edid_mode: Override the mode obtained by edid. */
> >       const struct drm_display_mode *override_edid_mode;
> > +
> > +     /** @override_bpc: Override the Bits per color obtained by edid. */
> > +     unsigned int override_bpc;
> >  };
> >
> >  struct panel_edp {
> > @@ -586,6 +589,9 @@ static int panel_edp_get_modes(struct drm_panel *panel,
> >       bool has_override_edid_mode = p->detected_panel &&
> >                                     p->detected_panel != ERR_PTR(-EINVAL) &&
> >                                     p->detected_panel->override_edid_mode;
>
> Unrelated to the patch at hand, since the pattern is there already, but
> something like this should be more reliable:
>
> bool has_override_edid_mode = !IS_ERR_OR_NULL(p->detected_panel) &&
>         p->detected_panel->override_edid_mode;
>
> It does not look like p->detected_panel could have other error pointer
> values than -EINVAL, but it looks awkward to check for NULL and one
> error pointer value, and then go on to dereference it.
>
> I guess even better would be to always use either NULL *or* error
> pointers, not a mix, but I digress.

Yeah, that makes sense. If someone wants to send a patch cleaning this
up, I'd be happy to review it! :-)

-Doug

Re: [PATCH v1 2/2] drm/panel-edp: Modify LQ116M1JW10 panel's bpc to 6

[Doug Anderson]

Hi,

On Thu, Oct 30, 2025 at 2:58 AM Ajye Huang
<ajye_huang@compal.corp-partner.google.com> wrote:
>
> Hi Doug,
>
> On Thu, Oct 30, 2025 at 7:25 AM Doug Anderson <dianders@chromium.org> wrote:
>
> >
> > Unless folks end up preferring EDID_QUIRK_FORCE_6BPC:
> >
> > Reviewed-by: Douglas Anderson <dianders@chromium.org>
>
> After following your suggestion with the following, the issue goes
> away during YouTube playback.
> I will send a new patch for drm_edid.c, thank you so much

FWIW, it is a bit baffling to me that you report link training seems
to be failing yet then talk about the symptom of noise during youtube
playback. If link training is failing I'd expect nothing to ever show
up on the screen...

-Doug