Enable on SM8750 MTP the Iris video codec for accelerated video
encoding/decoding.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

---

Do not merge because firmware (hard-coded in the driver) is not released.

For Rob's bot reports:

qcom,sm8750-videocc bindings and clock headers dependency (will fail
build):
https://lore.kernel.org/all/20241206-sm8750_videocc-v1-0-5da6e7eea2bd@quicinc.com/

qcom,sm8750-iris bindings:
https://lore.kernel.org/r/20250804-sm8750-iris-v2-0-6d78407f8078@linaro.org
---
 arch/arm64/boot/dts/qcom/sm8750-mtp.dts | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm64/boot/dts/qcom/sm8750-mtp.dts b/arch/arm64/boot/dts/qcom/sm8750-mtp.dts
index 75cfbb510be57a1ab8cb3d870b5c34d3baa53c70..4c155b731a68138154f66fdb0d0e6db5e47adf3c 100644
--- a/arch/arm64/boot/dts/qcom/sm8750-mtp.dts
+++ b/arch/arm64/boot/dts/qcom/sm8750-mtp.dts
@@ -812,6 +812,10 @@ vreg_l7n_3p3: ldo7 {
 	};
 };
 
+&iris {
+	status = "okay";
+};
+
 &lpass_vamacro {
 	pinctrl-0 = <&dmic01_default>, <&dmic23_default>;
 	pinctrl-names = "default";

-- 
2.48.1

On Wed, Aug 06, 2025 at 02:38:31PM +0200, Krzysztof Kozlowski wrote:
> Enable on SM8750 MTP the Iris video codec for accelerated video
> encoding/decoding.
> 
> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> 
> ---
> 
> Do not merge because firmware (hard-coded in the driver) is not released.

I don't think we have been delaying enablement of the hardware for these
reasons. The user might have other ways to get the firmware (or to
disable the device) in DT.

-- 
With best wishes
Dmitry

On Sat, Aug 09, 2025 at 12:42:16PM +0300, Dmitry Baryshkov wrote:
> On Wed, Aug 06, 2025 at 02:38:31PM +0200, Krzysztof Kozlowski wrote:
> > Enable on SM8750 MTP the Iris video codec for accelerated video
> > encoding/decoding.
> > 
> > Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> > 
> > ---
> > 
> > Do not merge because firmware (hard-coded in the driver) is not released.
> 
> I don't think we have been delaying enablement of the hardware for these
> reasons. The user might have other ways to get the firmware (or to
> disable the device) in DT.
> 

At least with the venus driver, missing firmware would result in
sync_state never happening. I see there's some changes in that behavior
lately, so perhaps this has changed?

But I agree with you that it would be preferable to deal with this
shortcoming in the implementation, rather than holding back the hardware
description.

Regards,
Bjorn

> -- 
> With best wishes
> Dmitry