[PATCH] drm/panel/boe-tv101wum-nl6: set MIPI_DSI_MODE_LPM after sending panel disable cmds

Brady Norander posted 1 patch 1 month, 4 weeks ago
[PATCH] drm/panel/boe-tv101wum-nl6: set MIPI_DSI_MODE_LPM after sending panel disable cmds
Posted by Brady Norander 1 month, 4 weeks ago
A recent change to Mediatek drm driver enabled HS mode support. With that
change, my MT8183-based krane sku176 chromebook display's colors would look darker
and desaturated after turning the display off and back on. Looking at other panel
drivers, it seems common to enable HS mode to send the disable commands and to
disable HS mode afterwards. However, the boe-tv101wum-nl6 driver would never
disable HS mode, leading to this issue. Update the behavior of boe_panel_disable
to match those other panel drivers. As a side note, I did also test removing the
line which enables HS mode during panel_disable. This worked fine for my device,
but just in case that doesn't work for other devices, I chose to keep that and
instead disable HS mode after sending the panel disable commands.

Signed-off-by: Brady Norander <brady.norander@mainlining.org>

diff --git a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
index d5fe105bdbdd..f69b5bd776c0 100644
--- a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
+++ b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
@@ -1326,6 +1326,8 @@ static int boe_panel_disable(struct drm_panel *panel)
 
 	mipi_dsi_msleep(&ctx, 150);
 
+	boe->dsi->mode_flags |= MIPI_DSI_MODE_LPM;
+
 	return ctx.accum_err;
 }
 
-- 
2.52.0
Re: [PATCH] drm/panel/boe-tv101wum-nl6: set MIPI_DSI_MODE_LPM after sending panel disable cmds
Posted by Neil Armstrong 1 month, 3 weeks ago
On 4/21/26 17:31, Brady Norander wrote:
> A recent change to Mediatek drm driver enabled HS mode support. With that
> change, my MT8183-based krane sku176 chromebook display's colors would look darker
> and desaturated after turning the display off and back on. Looking at other panel
> drivers, it seems common to enable HS mode to send the disable commands and to
> disable HS mode afterwards. However, the boe-tv101wum-nl6 driver would never
> disable HS mode, leading to this issue. Update the behavior of boe_panel_disable
> to match those other panel drivers. As a side note, I did also test removing the
> line which enables HS mode during panel_disable. This worked fine for my device,
> but just in case that doesn't work for other devices, I chose to keep that and
> instead disable HS mode after sending the panel disable commands.
> 
> Signed-off-by: Brady Norander <brady.norander@mainlining.org>
> 
> diff --git a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
> index d5fe105bdbdd..f69b5bd776c0 100644
> --- a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
> +++ b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
> @@ -1326,6 +1326,8 @@ static int boe_panel_disable(struct drm_panel *panel)
>   
>   	mipi_dsi_msleep(&ctx, 150);
>   
> +	boe->dsi->mode_flags |= MIPI_DSI_MODE_LPM;
> +
>   	return ctx.accum_err;
>   }
>   

Could you add a Fixes tag ?

Neil
Re: [PATCH] drm/panel/boe-tv101wum-nl6: set MIPI_DSI_MODE_LPM after sending panel disable cmds
Posted by Brady Norander 1 month, 3 weeks ago
On 4/24/26 05:30, Neil Armstrong wrote:
> On 4/21/26 17:31, Brady Norander wrote:
>> A recent change to Mediatek drm driver enabled HS mode support. With that
>> change, my MT8183-based krane sku176 chromebook display's colors would 
>> look darker
>> and desaturated after turning the display off and back on. Looking at 
>> other panel
>> drivers, it seems common to enable HS mode to send the disable 
>> commands and to
>> disable HS mode afterwards. However, the boe-tv101wum-nl6 driver would 
>> never
>> disable HS mode, leading to this issue. Update the behavior of 
>> boe_panel_disable
>> to match those other panel drivers. As a side note, I did also test 
>> removing the
>> line which enables HS mode during panel_disable. This worked fine for 
>> my device,
>> but just in case that doesn't work for other devices, I chose to keep 
>> that and
>> instead disable HS mode after sending the panel disable commands.
>>
>> Signed-off-by: Brady Norander <brady.norander@mainlining.org>
>>
>> diff --git a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c b/drivers/ 
>> gpu/drm/panel/panel-boe-tv101wum-nl6.c
>> index d5fe105bdbdd..f69b5bd776c0 100644
>> --- a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
>> +++ b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c
>> @@ -1326,6 +1326,8 @@ static int boe_panel_disable(struct drm_panel 
>> *panel)
>>       mipi_dsi_msleep(&ctx, 150);
>> +    boe->dsi->mode_flags |= MIPI_DSI_MODE_LPM;
>> +
>>       return ctx.accum_err;
>>   }
> 
> Could you add a Fixes tag ?
> 
> Neil

I can add a fixes tag and send a v2. Just one thing I'm not sure about, 
do I tag the mediatek drm commit which git bisect found, or do I tag the 
commit which added the panel driver since this faulty logic has existed 
since the driver was added?