On 27/06/2022 15:50, Pekka Paalanen wrote:

On Mon, 27 Jun 2022 13:40:04 +0000 Dennis Tsiang Dennis.Tsiang@arm.com wrote:

...

This patch is an early RFC to discuss the viable options and alternatives for inclusion of unsigned integer formats for the DRM API.

This patch adds a new single component 16-bit and a two component 32-bit DRM fourcc’s that represent unsigned integer formats. The use case for needing UINT formats, in our case, would be to support using raw buffers for camera ISPs.

For images imported with DRM fourcc + modifier combination, the GPU driver needs a way to determine the datatype of the format which currently the DRM API does not provide explicitly with a notable exception of the floating-point fourccs such as DRM_FORMAT_XRGB16161616F as an example. As the DRM fourccs do not currently define the interpretation of the data, should the information be made explicit in the DRM API similarly to how it is already done in Vulkan?

The reason for introducing datatype to the DRM fourcc's is that the alternative, for any API (e.g., EGL) that lacks the format datatype information for fourcc/modifier combination for dma_buf interop would be to introduce explicit additional metadata/attributes that encode this information which then would be passed to the GPU driver but the drawback of this is that it would require extending multiple graphics APIs to support every single platform.

By having the DRM API expose the datatype information for formats saves a lot of integration/verification work for all of the different graphics APIs and platforms as this information could be determined by the DRM triple alone for dma_buf interop.

It would be good to gather some opinions on what others think about introducing datatypes to the DRM API.

Hi,

I didn't quite grasp where this information is necessary, and when it is necessary, is it that simple to communicate? Does it even belong with the pixel format at all?

Let us consider the existing problems.

All traditional integer formats in drm_fourcc.h right now are unsigned. They get interpreted as being in the range [0.0, 1.0] for color operations, which means converting to another bit depth works implicitly. That's where the simplicity ends. We assume to have full quantization range unless otherwise noted in some auxiliary data, like KMS properties (I forget if there even was a property to say DRM framebuffer uses limited quantization range). We assume all pixel data is non-linearly encoded. There is no color space information. YUV-RGB conversion matrix coefficients are handled by a KMS property IIRC.

Coming back to the nominal range [0.0, 1.0]. That's an implicit assumption that allows us to apply things like LUTs. It already breaks down if you choose a floating-point format instead of unsigned integer format. Is FP pixel value 1.0 the same as nominal 1.0? Or is the FP pixel value 255.0 the same as nominal 1.0? KMS has no way to know or control that AFAIK.

If you had UINT format, meaning the nominal value range is [0.0, 65535.0] (e.g. for 16 bpc) instead of [0.0, 1.0], then how does that work with a LUT element in the color pipeline? How would it be both meaningful and different to existing 16 bpc integer format?

What's the actual difference between R16 and R16_UINT, what difference does it make to a GPU driver?

Maybe that is intimately dependent on the API where the pixel formats are used?

Maybe for KMS R16 and R16_UINT would be completely equivalent, but not for some other API?

We also need to be very careful to not load pixel format with meaning that does not belong there. Non-linear encoding (transfer function) is obviously something that is completely unrelated to the pixel format, as long as the pixel format defines a conversion to nominal value range. Color space (primaries and white point) are another thing that has nothing to do with pixel format, and so must not be in any way implied by pixel format.

Should a pixel format define how the raw pixel values are converted to nominal values?

No, because we have quantization range as a separate property, currently with "full" and "limited" understood, where "limited" means different things depending on the color model.

Color model is defined by the pixel formats: we have RGB and YUV formats, likewise is chroma sub-sampling.

Hmm.

Thanks, pq

Hi Pekka,

Thanks for your comments. This is not intended to be used for KMS, where indeed there would be no difference. This proposal is for other Graphics APIs such as Vulkan, which requires the application to be explicit upfront about how they will interpret the data, whether that be UNORM, UINT .etc. We want to be able to import dma_bufs which create a VkImage with a "_UINT" VkFormat. However there is currently no explicit mapping between the DRM fourccs + modifiers combos to "_UINT" VkFormats. One solution is to encode that into the fourccs, which is what this RFC is proposing.

Thanks, Dennis IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

On Wed, 29 Jun 2022 14:53:49 +0000 Simon Ser contact@emersion.fr wrote:

On Wednesday, June 29th, 2022 at 16:46, Dennis Tsiang dennis.tsiang@arm.com wrote:

...

Thanks for your comments. This is not intended to be used for KMS, where indeed there would be no difference. This proposal is for other Graphics APIs such as Vulkan, which requires the application to be explicit upfront about how they will interpret the data, whether that be UNORM, UINT .etc. We want to be able to import dma_bufs which create a VkImage with a "_UINT" VkFormat. However there is currently no explicit mapping between the DRM fourccs + modifiers combos to "_UINT" VkFormats. One solution is to encode that into the fourccs, which is what this RFC is proposing.

As a general comment, I don't think it's reasonable to encode all of the VkFormat information inside DRM FourCC. For instance, VkFormat has SRGB/UNORM variants which describe whether pixel values are electrical or optical (IOW, EOTF-encoded or not). Moreover, other APIs may encode different information in their format enums.

Yeah, do not add any of that information to the DRM pixel format codes.

There is *so much* other stuff you also need to define than what's already mentioned, and which bits you need for the API at hand depends totally on the API at hand. After the API has defined some parts of the metadata, the API user has to take care of the remaining parts of the metadata in other ways, like dynamic range or color space.

Besides, when you deal with dmabuf, you already need to pass a lot of metadata explicitly, like the pixel format, width, height, stride, modifier, etc. so it's better to add more of those (like we will be doing in Wayland, and not specific to dmabuf even) than to try make pixel formats a huge mess through combinatorial explosion and sometimes partial and sometimes conflicting image metadata.

You might be able to get a glimpse of what all metadata there could be by reading https://gitlab.freedesktop.org/pq/color-and-hdr/-/blob/main/doc/pixels_color... .

Compare Vulkan formats to e.g. https://docs.microsoft.com/en-us/windows/win32/api/dxgicommon/ne-dxgicommon-... and you'll see that while DXGI color space enumeration is mostly about other stuff, it also has overlap with Vulkan formats I think, at least the SRGB vs. not part.

Btw. practically all buffers you see used, especially if they are 8 bpc, they are almost guaranteed to be "SRGB" non-linearly encoded, but do you ever see that fact being explicitly communicated?

Then there is the question that if you have an SRGB-encoded buffer, do you want to read out SRGB-encoded or linear values? That depends on what you are doing with the buffer, so if you always mapped dmabuf to Vulkan SRGB formats (or always to non-SRGB formats), then you need some other way in Vulkan for the app to say whether to sample encoded or linear (electrical or optical) values. And whether texture filtering is done in encoded or linear space, because that makes a difference too.

IOW, there are cases where the format mapping depends on the user of the buffer and not only on the contents of the buffer.

Therefore you simply cannot create a static mapping table between two format definition systems when the two systems are fundamentally different, like Vulkan and DRM fourcc.

Thanks, pq

On Fri, 2022-07-15 at 11:20 +0100, Dennis Tsiang wrote:

On 30/06/2022 08:47, Pekka Paalanen wrote:

...

On Wed, 29 Jun 2022 14:53:49 +0000 Simon Ser contact@emersion.fr wrote:

...

On Wednesday, June 29th, 2022 at 16:46, Dennis Tsiang dennis.tsiang@arm.com wrote:

...

Thanks for your comments. This is not intended to be used for KMS, where indeed there would be no difference. This proposal is for other Graphics APIs such as Vulkan, which requires the application to be explicit upfront about how they will interpret the data, whether that be UNORM, UINT .etc. We want to be able to import dma_bufs which create a VkImage with a "_UINT" VkFormat. However there is currently no explicit mapping between the DRM fourccs + modifiers combos to "_UINT" VkFormats. One solution is to encode that into the fourccs, which is what this RFC is proposing.

As a general comment, I don't think it's reasonable to encode all of the VkFormat information inside DRM FourCC. For instance, VkFormat has SRGB/UNORM variants which describe whether pixel values are electrical or optical (IOW, EOTF-encoded or not). Moreover, other APIs may encode different information in their format enums.

Yeah, do not add any of that information to the DRM pixel format codes.

There is *so much* other stuff you also need to define than what's already mentioned, and which bits you need for the API at hand depends totally on the API at hand. After the API has defined some parts of the metadata, the API user has to take care of the remaining parts of the metadata in other ways, like dynamic range or color space.

Besides, when you deal with dmabuf, you already need to pass a lot of metadata explicitly, like the pixel format, width, height, stride, modifier, etc. so it's better to add more of those (like we will be doing in Wayland, and not specific to dmabuf even) than to try make pixel formats a huge mess through combinatorial explosion and sometimes partial and sometimes conflicting image metadata.

You might be able to get a glimpse of what all metadata there could be by reading https://gitlab.freedesktop.org/pq/color-and-hdr/-/blob/main/doc/pixels_color... .

Compare Vulkan formats to e.g. https://docs.microsoft.com/en-us/windows/win32/api/dxgicommon/ne-dxgicommon-... and you'll see that while DXGI color space enumeration is mostly about other stuff, it also has overlap with Vulkan formats I think, at least the SRGB vs. not part.

Btw. practically all buffers you see used, especially if they are 8 bpc, they are almost guaranteed to be "SRGB" non-linearly encoded, but do you ever see that fact being explicitly communicated?

Then there is the question that if you have an SRGB-encoded buffer, do you want to read out SRGB-encoded or linear values? That depends on what you are doing with the buffer, so if you always mapped dmabuf to Vulkan SRGB formats (or always to non-SRGB formats), then you need some other way in Vulkan for the app to say whether to sample encoded or linear (electrical or optical) values. And whether texture filtering is done in encoded or linear space, because that makes a difference too.

IOW, there are cases where the format mapping depends on the user of the buffer and not only on the contents of the buffer.

Therefore you simply cannot create a static mapping table between two format definition systems when the two systems are fundamentally different, like Vulkan and DRM fourcc.

Thanks, pq

Thanks all for your comments. We'll look into an alternative approach to achieve what we need.

I mostly agree with Pekka here. The fourcc formats as they currently are defined only specify a bit pattern and channel order, not an interpretation. Vulkan formats, on the other hand, have everything you need in order to know how to convert float vec4s to/from that format in a shader. That's not the same as knowing what the data represents (colorspace, wite values, etc.) but it's a lot more than fourcc.

That said, the Vulkan APIs for querying modifier support will give you much more fine-grained information about exactly the Vulkan formats you request. So if you ask for modifier support for VK_FORMAT_R16G16_UINT, that's what you'll get. I *think* it should be fine to use VK_FORMAT_R16G16_UINT with DRM_FOURCC_GR1616. Of course, the API on the other side will also need a more precise format specifier than fourcc if it's to know the difference between R16G16_UINT and R16G16_UNORM.

--Jason