Message-ID: <5dcc3dc3-ba12-4d66-88e9-5e06bb707135@linaro.org>
Date: Fri, 27 Jun 2025 17:31:54 +0100
From: Bryan O'Donoghue <bryan.odonoghue@...aro.org>
To: Vikash Garodia <quic_vgarodia@...cinc.com>,
 Dikshita Agarwal <quic_dikshita@...cinc.com>,
 Abhinav Kumar <abhinav.kumar@...ux.dev>,
 Mauro Carvalho Chehab <mchehab@...nel.org>, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>
Cc: linux-media@...r.kernel.org, linux-arm-msm@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 1/5] media: dt-bindings: add non-pixel property in iris
 schema

On 27/06/2025 16:48, Vikash Garodia wrote:
> Existing definition limits the IOVA to an addressable range of 4GiB, and
> even within that range, some of the space is used by IO registers,
> thereby limiting the available IOVA to even lesser. Video hardware is
> designed to emit different stream-ID for pixel and non-pixel buffers,
> thereby introduce a non-pixel sub node to handle non-pixel stream-ID.
> 
> With this, both iris and non-pixel device can have IOVA range of 0-4GiB
> individually. Certain video usecases like higher video concurrency needs
> IOVA higher than 4GiB.
> 
> Add reference to the reserve-memory schema, which defines reserved IOVA
> regions that are *excluded* from addressable range. Video hardware
> generates different stream IDs based on the predefined range of IOVA
> addresses. Thereby IOVA addresses for firmware and data buffers need to
> be non overlapping. For ex. 0x0-0x25800000 address range is reserved for
> firmware stream-ID, while non-pixel (bitstream) stream-ID can be
> generated by hardware only when bitstream buffers IOVA address is from
> 0x25800000-0xe0000000.
> Non-pixel stream-ID can now be part of the new sub-node, hence iommus in
> iris node can have either 1 entry for pixel stream-id or 2 entries for
> pixel and non-pixel stream-ids.
> 
> Signed-off-by: Vikash Garodia <quic_vgarodia@...cinc.com>
> ---
>   .../bindings/media/qcom,sm8550-iris.yaml           | 40 ++++++++++++++++++++--
>   1 file changed, 38 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/media/qcom,sm8550-iris.yaml b/Documentation/devicetree/bindings/media/qcom,sm8550-iris.yaml
> index c79bf2101812d83b99704f38b7348a9f728dff44..4dda2c9ca1293baa7aee3b9ee10aff38d280fe05 100644
> --- a/Documentation/devicetree/bindings/media/qcom,sm8550-iris.yaml
> +++ b/Documentation/devicetree/bindings/media/qcom,sm8550-iris.yaml
> @@ -65,10 +65,31 @@ properties:
>         - const: core
>   
>     iommus:
> +    minItems: 1
>       maxItems: 2
>   
>     dma-coherent: true
>   
> +  non-pixel:
> +    type: object
> +    additionalProperties: false
> +
> +    description:
> +      Non pixel context bank is needed when video hardware have distinct iommus
> +      for non pixel buffers. Non pixel buffers are mainly compressed and
> +      internal buffers.

You do a better job in the cover letter of describing what this is, why 
its needed etc.

Not asking for this verbatim but its clearer:

"All non pixel buffers, i.e bitstream, HFI queues
and internal buffers related to bitstream processing, would be managed
by this non_pixel device."

Where does the term "non-pixel" come from if its a meaningful name wrt 
to the firmware then non-pixel is fine but, consider a name such as 
"out-of-band" or "oob"

out-of-band is a common term as is "sideband" but sideband I think has a 
different meaning, really this non-data/non-pixel data stuff is out-of-band.

At least for the way the language pack I have installed in my brain 
right now, "oob" or "out-of-band" is a more intuitive name. Its really 
up to you though the main point would be to enumerate the description 
here with some of the detail you've put into the cover letter.

> +
> +    properties:
> +      iommus:
> +        maxItems: 1
> +
> +      memory-region:
> +        maxItems: 1
> +
> +    required:
> +      - iommus
> +      - memory-region
> +
>     operating-points-v2: true

>     opp-table:
> @@ -86,6 +107,7 @@ required:
>   
>   allOf:
>     - $ref: qcom,venus-common.yaml#
> +  - $ref: /schemas/reserved-memory/reserved-memory.yaml
>     - if:
>         properties:
>           compatible:
> @@ -117,6 +139,16 @@ examples:
>       #include <dt-bindings/power/qcom-rpmpd.h>
>       #include <dt-bindings/power/qcom,rpmhpd.h>
>   
> +    reserved-memory {
> +      #address-cells = <2>;
> +      #size-cells = <2>;
> +
> +      iris_resv: reservation-iris {
> +        iommu-addresses = <&iris_non_pixel 0x0 0x0 0x0 0x25800000>,
> +                          <&iris_non_pixel 0x0 0xe0000000 0x0 0x20000000>;
> +      };
> +    };

iris_oob would be less text in the end.

> +
>       video-codec@...0000 {
>           compatible = "qcom,sm8550-iris";
>           reg = <0x0aa00000 0xf0000>;
> @@ -144,12 +176,16 @@ examples:
>           resets = <&gcc GCC_VIDEO_AXI0_CLK_ARES>;
>           reset-names = "bus";
>   
> -        iommus = <&apps_smmu 0x1940 0x0000>,
> -                 <&apps_smmu 0x1947 0x0000>;
> +        iommus = <&apps_smmu 0x1947 0x0000>;
>           dma-coherent;
>   
>           operating-points-v2 = <&iris_opp_table>;
>   
> +        iris_non_pixel: non-pixel {
> +            iommus = <&apps_smmu 0x1940 0x0000>;
> +            memory-region = <&iris_resv>;
> +        };
> +
>           iris_opp_table: opp-table {
>               compatible = "operating-points-v2";
>   
> 

So I was trying to think of a way to catch you out with an ABI break 
but, I don't see how you add minItems: 1 to the iommus declaration above 
so dts prior to this change should still be valid.

I think this adds up but, consider oob instead of non-pixel.

Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@...aro.org>

---
bod

Powered by blists - more mailing lists