Add support for MDSS on Milos.

Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
---
 drivers/gpu/drm/msm/msm_mdss.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/msm/msm_mdss.c b/drivers/gpu/drm/msm/msm_mdss.c
index 2d0e3e784c04..aa63c079d730 100644
--- a/drivers/gpu/drm/msm/msm_mdss.c
+++ b/drivers/gpu/drm/msm/msm_mdss.c
@@ -535,6 +535,10 @@ static void mdss_remove(struct platform_device *pdev)
 	msm_mdss_destroy(mdss);
 }
 
+static const struct msm_mdss_data data_14k = {
+	.reg_bus_bw = 14000,
+};
+
 static const struct msm_mdss_data data_57k = {
 	.reg_bus_bw = 57000,
 };
@@ -553,6 +557,7 @@ static const struct msm_mdss_data data_153k6 = {
 
 static const struct of_device_id mdss_dt_match[] = {
 	{ .compatible = "qcom,mdss", .data = &data_153k6 },
+	{ .compatible = "qcom,milos-mdss", .data = &data_14k },
 	{ .compatible = "qcom,msm8998-mdss", .data = &data_76k8 },
 	{ .compatible = "qcom,qcm2290-mdss", .data = &data_76k8 },
 	{ .compatible = "qcom,sa8775p-mdss", .data = &data_74k },

-- 
2.52.0

On Fri, Dec 19, 2025 at 05:41:09PM +0100, Luca Weiss wrote:
> Add support for MDSS on Milos.
> 
> Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
> ---
>  drivers/gpu/drm/msm/msm_mdss.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>

Though 14000 seems lower than the value for all other platforms.

-- 
With best wishes
Dmitry

On 12/20/25 5:52 PM, Dmitry Baryshkov wrote:
> On Fri, Dec 19, 2025 at 05:41:09PM +0100, Luca Weiss wrote:
>> Add support for MDSS on Milos.
>>
>> Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
>> ---
>>  drivers/gpu/drm/msm/msm_mdss.c | 5 +++++
>>  1 file changed, 5 insertions(+)
>>
> 
> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> 
> Though 14000 seems lower than the value for all other platforms.

IIUC it's essentially "nonzero" or "first-level-above-just-nonzero"

The downstream DT also defines a 140_000 and a 310_000 point.. though
the middle one is never(?) used and the latter one is only used during
UEFI handoff and upon the first commit after (runtime) resume

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

On 22/12/2025 11:09, Konrad Dybcio wrote:
> On 12/20/25 5:52 PM, Dmitry Baryshkov wrote:
>> On Fri, Dec 19, 2025 at 05:41:09PM +0100, Luca Weiss wrote:
>>> Add support for MDSS on Milos.
>>>
>>> Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
>>> ---
>>>   drivers/gpu/drm/msm/msm_mdss.c | 5 +++++
>>>   1 file changed, 5 insertions(+)
>>>
>>
>> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
>>
>> Though 14000 seems lower than the value for all other platforms.
> 
> IIUC it's essentially "nonzero" or "first-level-above-just-nonzero"
> 
> The downstream DT also defines a 140_000 and a 310_000 point.. though
> the middle one is never(?) used and the latter one is only used during
> UEFI handoff and upon the first commit after (runtime) resume

If I remember correctly, those are also used for programming the LUTs or 
other big tables into the hardware.

> 
> Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> Konrad


-- 
With best wishes
Dmitry

On 12/22/25 10:12 AM, Dmitry Baryshkov wrote:
> On 22/12/2025 11:09, Konrad Dybcio wrote:
>> On 12/20/25 5:52 PM, Dmitry Baryshkov wrote:
>>> On Fri, Dec 19, 2025 at 05:41:09PM +0100, Luca Weiss wrote:
>>>> Add support for MDSS on Milos.
>>>>
>>>> Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
>>>> ---
>>>>   drivers/gpu/drm/msm/msm_mdss.c | 5 +++++
>>>>   1 file changed, 5 insertions(+)
>>>>
>>>
>>> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
>>>
>>> Though 14000 seems lower than the value for all other platforms.
>>
>> IIUC it's essentially "nonzero" or "first-level-above-just-nonzero"
>>
>> The downstream DT also defines a 140_000 and a 310_000 point.. though
>> the middle one is never(?) used and the latter one is only used during
>> UEFI handoff and upon the first commit after (runtime) resume
> 
> If I remember correctly, those are also used for programming the LUTs or other big tables into the hardware.

FWIW this is an interconnect path for non-DMA accesses so that only
makes sense - huge writes could overwhelm the bus

Konrad