Texture format misreporting tracking issue

This is a tracking issue for texture format misreporting.

This tracking issue is a WIP.

## Testing
See #8646 

Some testing is added in #8653, which is where many of these issues were found.

## General
- [ ] #8662 
- [ ] #8647
- [ ] #8654 - instances of this are covered below. These have potential to be confusing and cause user errors but not to misreport capabilities
- [ ] P010 says "[Features::TEXTURE_FORMAT_P010](vscode-file://vscode-app/Users/supamaggie70/code/Visual%20Studio%20Code.app/Contents/Resources/app/out/vs/code/electron-browser/workbench/workbench.html) must be enabled to use this texture format." However, it also requires `wgpu::Features::TEXTURE_FORMAT_16BIT_NORM` which isn't documented anywhere to my knowledge
- [ ] R16Unorm and friends have nonsensical capabilities on all backends, supporting multisample but not being a render attachment

## Metal
Metal is especially bad. It is basically an attempt to copy down strange rules from a couple sources, including a huge table, which all change over time. Needless to say, there are about a billion bugs here.

See also #8594, which covers metal feature detection more broadly. #8439 fixes the crashes that could occur at adapter enumeration time, but leaves most of the texture format issues out.

### Etc2Rgb8A1UnormSrgb format issue
```
Trying to create view with format Etc2Rgb8A1UnormSrgb on texture with format Etc2Rgb8A1Unorm and caps TextureFormatFeatures { allowed_usages: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING), flags: TextureFormatFeatureFlags(FILTERABLE | BLENDABLE) }, and view format caps TextureFormatFeatures { allowed_usages: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING), flags: TextureFormatFeatureFlags(FILTERABLE | BLENDABLE) }
    _validateTextureView:745: failed assertion `Texture View Validation
    texture usage (0x01) doesn't specify MTLTextureUsagePixelFormatView (0x10)
```

## Vulkan
Misreported capabilities are SUPER common here. For example:

### General
* `Format Rgb9e5Ufloat supports multisample but doesn't support being a render attachment, with caps TextureFormatFeatures { allowed_usages: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING), flags: TextureFormatFeatureFlags(FILTERABLE | MULTISAMPLE_X4 | MULTISAMPLE_RESOLVE) }`

### R64Uint
```
Creating texture with format R64Uint and caps TextureFormatFeatures { allowed_usages: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING | STORAGE_BINDING), flags: TextureFormatFeatureFlags(STORAGE_READ_ONLY | STORAGE_WRITE_ONLY) }
    [2025-12-05T20:26:28Z ERROR wgpu_hal::vulkan::instance] VALIDATION [VUID-VkImageCreateInfo-imageCreateMaxMipLevels-02251 (0xbebcae79)]
        	vkCreateImage(): The following VkImageCreateInfo returned VK_ERROR_FORMAT_NOT_SUPPORTED when calling vkGetPhysicalDeviceImageFormatProperties2
        format (VK_FORMAT_R64_UINT)
        type (VK_IMAGE_TYPE_2D)
        tiling (VK_IMAGE_TILING_OPTIMAL)
        usage (VK_IMAGE_USAGE_TRANSFER_SRC_BIT|VK_IMAGE_USAGE_TRANSFER_DST_BIT|VK_IMAGE_USAGE_SAMPLED_BIT|VK_IMAGE_USAGE_STORAGE_BIT)
        flags (VkImageCreateFlags(0))
        VkImageCreateInfo::pNext is NULL.
        The Vulkan spec states: Each of the following values (as described in Image Creation Limits) must not be undefined : imageCreateMaxMipLevels, imageCreateMaxArrayLayers, imageCreateMaxExtent, and imageCreateSampleCounts (https://vulkan.lunarg.com/doc/view/1.4.328.1/linux/antora/spec/latest/chapters/resources.html#VUID-VkImageCreateInfo-imageCreateMaxMipLevels-02251)
```

### NV12
```
Trying to create view with format NV12 on texture with format NV12 and caps TextureFormatFeatures { allowed_usages: TextureUsages(TEXTURE_BINDING), flags: TextureFormatFeatureFlags(0x0) }, and view format caps TextureFormatFeatures { allowed_usages: TextureUsages(TEXTURE_BINDING), flags: TextureFormatFeatureFlags(0x0) }
    [2025-12-05T20:26:27Z ERROR wgpu_hal::vulkan::instance] VALIDATION [VUID-VkImageViewCreateInfo-subresourceRange-07818 (0x3ffdd96b)]
        	vkCreateImageView(): pCreateInfo->subresourceRange.aspectMask (VK_IMAGE_ASPECT_PLANE_0_BIT|VK_IMAGE_ASPECT_PLANE_1_BIT) is invalid for VK_FORMAT_G8_B8R8_2PLANE_420_UNORM.
        The Vulkan spec states: subresourceRange.aspectMask must only have at most 1 valid multi-planar aspect mask bit (https://vulkan.lunarg.com/doc/view/1.4.328.1/linux/antora/spec/latest/chapters/resources.html#VUID-VkImageViewCreateInfo-subresourceRange-07818)
```

## GLES

### R32Float "downlevel issue"

```

    Creating texture with format R32Float and caps TextureFormatFeatures { allowed_usages: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING | STORAGE_BINDING | RENDER_ATTACHMENT | TRANSIENT), flags: TextureFormatFeatureFlags(MULTISAMPLE_X4 | STORAGE_READ_ONLY | STORAGE_WRITE_ONLY | STORAGE_READ_WRITE) }
    [2025-12-05T20:26:05Z ERROR wgpu::backend::wgpu_core] Handling wgpu errors as fatal by default

    thread '<unnamed>' panicked at wgpu/src/backend/wgpu_core.rs:1607:26:
    wgpu error: Validation Error

    Caused by:
      In Device::create_texture
        Texture usages TextureUsages(RENDER_ATTACHMENT) are not allowed on a texture of type R32Float due to downlevel restrictions
```

### Copy command encoding
```
Encoding copy
    [2025-12-05T20:26:05Z INFO  wgpu_core::device::global] Device::create_texture(TextureDescriptor { label: None, size: (64, 64, 32), mip_level_count: 1, sample_count: 1, dimension: D2, format: Depth32FloatStencil8, usage: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING | RENDER_ATTACHMENT), view_formats: [Depth32FloatStencil8] }) -> Id(1,3)
    [2025-12-05T20:26:05Z INFO  wgpu_core::command::transfer] CommandEncoder::copy_texture_to_texture Id(0,4) -> Id(1,3) (64, 64, 1)
    [2025-12-05T20:26:05Z INFO  wgpu_core::device::global] Texture::drop Id(1,3)
    [2025-12-05T20:26:05Z INFO  wgpu_core::device::global] Texture::drop Id(0,4)
    Attempting to record & submit command buffer
    [2025-12-05T20:26:05Z INFO  wgpu_core::device::global] CommandEncoder::drop Id(0,1)
    [2025-12-05T20:26:05Z INFO  wgpu_core::command] Drop CommandEncoder with '' label
    [2025-12-05T20:26:05Z INFO  wgpu_core::device::queue] Queue::submit
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: depth attachment incomplete [-1]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glClear(incomplete framebuffer)
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: color attachment incomplete [0]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage2D(invalid readbuffer)
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: color attachment incomplete [0]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage2D(invalid readbuffer)
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: color attachment incomplete [0]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage2D(invalid readbuffer)
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: color attachment incomplete [0]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage2D(invalid readbuffer)
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: color attachment incomplete [0]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage3D(invalid readbuffer)
    [2025-12-05T20:26:05Z WARN  wgpu_hal::gles] GLES: [API/Other] ID 2 : FBO incomplete: color attachment incomplete [0]
        
    [2025-12-05T20:26:05Z ERROR wgpu_hal::gles] GLES: [API/Error] ID 1 : GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage3D(invalid readbuffer)
```

This also gets spammed below that
```
    [2025-12-05T20:26:04Z ERROR wgpu_test::expectations] Validation Error: GL_INVALID_FRAMEBUFFER_OPERATION in glCopyTexSubImage2D(invalid readbuffer)
```

These errors occur for many texture formats, not even exclusive to one family.

### Bc7RgbaUnormSrgb issue
```
Creating texture with format Bc7RgbaUnormSrgb and caps TextureFormatFeatures { allowed_usages: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING), flags: TextureFormatFeatureFlags(FILTERABLE | BLENDABLE) }
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::global] Device::create_texture(TextureDescriptor { label: None, size: (64, 64, 32), mip_level_count: 1, sample_count: 1, dimension: D2, format: Bc7RgbaUnormSrgb, usage: TextureUsages(COPY_SRC | COPY_DST | TEXTURE_BINDING), view_formats: [Bc7RgbaUnormSrgb] }) -> Id(0,42)
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::global] Texture::drop Id(0,42)
    [2025-12-05T20:26:04Z INFO  wgpu_core::resource] Destroy raw Texture with '' label
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::global] CommandEncoder::drop Id(0,1)
    [2025-12-05T20:26:04Z INFO  wgpu_core::command] Drop CommandEncoder with '' label
    [2025-12-05T20:26:04Z INFO  wgpu_core::instance] Adapter::drop Id(0,1)
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::global] Device::drop Id(0,1)
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::global] Queue::drop Id(0,1)
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::queue] Drop Queue with '' label
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::queue] Waiting on last submission. try: 1/6. timeout: 100ms
    [2025-12-05T20:26:04Z INFO  wgpu_core::resource] Destroy raw StagingBuffer
    [2025-12-05T20:26:04Z INFO  wgpu_core::device::resource] Drop Device with '' label
    [2025-12-05T20:26:04Z INFO  wgpu_core::global] Global::drop
    [2025-12-05T20:26:04Z ERROR wgpu_test::expectations] Validation Error: GL_INVALID_OPERATION in glTexStorage3D(internalformat = GL_COMPRESSED_RED_RGTC1)
    [2025-12-05T20:26:04Z ERROR wgpu_test::expectations] Validation Error: GL_INVALID_OPERATION in glTexStorage3D(internalformat = GL_COMPRESSED_SIGNED_RED_RGTC1)
    [2025-12-05T20:26:04Z ERROR wgpu_test::expectations] Validation Error: GL_INVALID_OPERATION in glTexStorage3D(internalformat = GL_COMPRESSED_RG_RGTC2)
    [2025-12-05T20:26:04Z ERROR wgpu_test::expectations] Validation Error: GL_INVALID_OPERATION in glTexStorage3D(internalformat = GL_COMPRESSED_SIGNED_RG_RGTC2)
```

### Invalid enum stuff
```
    [2025-12-05T20:26:04Z ERROR wgpu_test::expectations] Validation Error: GL_INVALID_ENUM in glTexStorage2DMultisample(target=GL_TEXTURE_2D)
```
This gets spammed in some tests at the end after they panic for other reasons

## DX12

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Texture format misreporting tracking issue #8670

Testing

General

Metal

Etc2Rgb8A1UnormSrgb format issue

Vulkan

General

R64Uint

NV12

GLES

R32Float "downlevel issue"

Copy command encoding

Bc7RgbaUnormSrgb issue

Invalid enum stuff

DX12

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Texture format misreporting tracking issue #8670

Description

Testing

General

Metal

Etc2Rgb8A1UnormSrgb format issue

Vulkan

General

R64Uint

NV12

GLES

R32Float "downlevel issue"

Copy command encoding

Bc7RgbaUnormSrgb issue

Invalid enum stuff

DX12

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions