1744: hal/dx12: fix depth formats and pipeline layout r=kvark a=kvark
**Connections**
Closes#1727
**Description**
There were 2 problems. One was that the reversing of the pipeline layout was done wrong (again). I don't know how it still worked on one of my machines :/
The other was that the backend relied on the depth format reinterpretation without using a typeless format. This is OK on higher-level platforms, but we should be doing it more careful. I tried to approach this maximally efficient: we only use the typeless format if all of the following are true:
1. it's a depth/stencil format
2. it needs to be seen as SRV/UAV
So in the case of a simple depth texture, we are still going to be using a fully-qualified format, and therefore get all of the benefit of HW compression on it.
**Testing**
Ran the examples on Intel
Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
1741: Improve Docs for DownlevelFlags r=cwfitzgerald a=zicklag
**Connections**
None
**Description**
Clarifies the documentation on the `DownlevelFlags`, and wraps the documentation of some of the flags to 100 chars.
**Testing**
Tested on Ubuntu 20.04
Co-authored-by: Zicklag <zicklag@katharostech.com>
1742: Mention WGPU_ADAPTER_NAME Env Var in WGPU README r=cwfitzgerald a=zicklag
**Connections**
None
**Description**
Mentions the new `WGPU_ADAPTER_NAME` environment variable in the WGPU readme.
**Testing**
Co-authored-by: Zicklag <zicklag@katharostech.com>
1737: Use Bitshift Style for All Bitflag Consts r=cwfitzgerald a=zicklag
**Connections**
None exactly, but somewhat related to https://github.com/gfx-rs/wgpu/pull/1686 as I was going to put some of this change into that PR, but am splitting it out for ease of review.
**Description**
This is purely code-style related, but it makes it easier to read bitflag constants and much eaiser to add bitflags in the future.
**Testing**
Tested on Ubuntu 20.04
Co-authored-by: Zicklag <zicklag@katharostech.com>
1738: Rename texture usage names according to binding r=kvark a=kvark
**Connections**
Matches https://github.com/gpuweb/gpuweb/pull/1989
**Description**
The `SAMPLED` usage name was confusing because integer and MSAA textures can't be sampled (with a sampler), but fetching data from them is under this usage.
**Testing**
No need
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1722: Document coordinate systems in README r=cwfitzgerald a=MichaelMMacLeod
I had to do a bit of digging to find information about coordinate systems in
wgpu (asking on Matrix). This commit adds a table of coordinate systems to help
clarify the issue.
The table comes from the DX12 backend documentation of gfx-rs:
https://github.com/gfx-rs/gfx/tree/master/src/backend/dx12
**Connections**
(none)
**Description**
It was difficult for me to find information about what coordinate systems wgpu uses. This commit addresses this issue by adding documentation about wgpu's coordinate systems to the README.
**Testing**
This is purely a documentation change (I have written no tests).
If you want to see how the table looks on Github, check out my fork at https://github.com/MichaelMMacLeod/wgpu/tree/add-coords-to-readme
Co-authored-by: Michael MacLeod <michaelmmacleod@gmail.com>
1729: Handle Multi-threaded EGL Context Access r=cwfitzgerald,kvark a=zicklag
**Connections**
https://github.com/gfx-rs/wgpu/discussions/1630, https://github.com/bevyengine/bevy/issues/841
**Description**
Implements the synchronization necessary to use the GL backend from multiple threads. Accomplishes this by using a mutex around the GL context with extra wrapping to bind and unbind the EGL context when locking and unlocking.
**Testing**
Tested on Ubunty 20.04 with a fork of the Bevy game engine and the WGPU examples ( not that the examples test the multi-threading ).
## Remaining Issues
There is only one Bevy example I cannot get to run yet and it's the `load_gltf` example. It fails with a shader translation error:
```
Jul 26 20:36:50.949 ERROR naga:🔙:glsl: Conflicting samplers for _group_3_binding_10
Jul 26 20:36:50.950 WARN wgpu::backend::direct: Shader translation error for stage FRAGMENT: A image was used with multiple samplers
Jul 26 20:36:50.950 WARN wgpu::backend::direct: Please report it to https://github.com/gfx-rs/naga
Jul 26 20:36:50.950 ERROR wgpu::backend::direct: wgpu error: Validation Error
Caused by:
In Device::create_render_pipeline
Internal error in FRAGMENT shader: A image was used with multiple samplers
```
Interestingly, I think the shader in question doesn't have a `group(3), binding(10)` anywhere that I know of so I'm going to have to drill down a bit more and find out exactly which shader translation is failing more.
This could potentially be fixed in a separate PR. I think the rest of this PR is rather straight-forward and the fix for the error above is probably mostly unrelated to the primary changes made in this PR.
Co-authored-by: Zicklag <zicklag@katharostech.com>
1732: Fix Tests on DX12/GL r=kvark a=cwfitzgerald
**Connections**
Marks #1730 as known failure
Closes#1695
**Description**
This fixes up tests on both DX12 and GL
**Testing**
Tests
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
1731: hal/dx12: support base vertex/instance r=cwfitzgerald a=kvark
**Connections**
Fixes#1725
Depends on https://github.com/gfx-rs/naga/pull/1141, https://github.com/gfx-rs/d3d12-rs/pull/34, and https://github.com/gfx-rs/d3d12-rs/pull/35 (thanks `@msiglreith` for quick reviews!).
**Description**
We allocate one root parameter covering the special constant buffer view with root constants. We set the root constants for each draw, but only on change.
This PR does *not* handle the indirect case. I'll follow-up with this.
**Testing**
Not thoroughly tested, but `@cwfitzgerald` already implemented a test for this, and this PR enables it on DX12.
Also, examples still seem to work :)
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1609: [Vulkan] Initialize wgpu objects from raw handles r=kvark a=zarik5
**Connections**
This PR is a successor of https://github.com/gfx-rs/gfx/pull/3762
**Description**
The `handle_is_external` flag mechanism was not enough to ensure safety, when for example the Instance is wrapped in `Arc<>` and a library cannot keep track of all its clones. This is the case with the bevy render rewrite. The solution was to let wgpu be in charge of destroying the handles, but it also has to keep a reference to a "drop guard" which is always dropped after the handle is destroyed. For the OpenXR integration, this drop guard will be `xr::Instance`.
For now this is just a proof of concept. Only instance creation is handled, and there is even type error in `wgc::Instance::from_hal()`.
I have a few concerns:
* Is it ok to expose `hal::Instance` from the wgpu crate? Or should the user pass all the parameters so `hal::Instance` can be constructed internally? This second options is more disruptive, since the wgpu-types crate would need to import all platform-specific crates to define the structure/enum to hold the raw handles.
* Without counting the call to create the `hal::Instance`, there are 4 indirection calls to save the raw instance. Can this be optimized in any way?
Do you think it is possible to merge wgpu-hal into wgpu-core? This could help with reducing the distance from the surface level API to the platform specific APIs even more.
**Testing**
Vulkan/Android (Oculus Quest) using [this sample](https://github.com/zarik5/openxrs/blob/wgpu-test/openxr/examples/vulkan.rs).
Co-authored-by: Riccardo Zaglia <riccardo.zaglia5@gmail.com>
1717: Work Around Fastclear Bug for Web and Native GL r=cwfitzgerald,kvark a=zicklag
**Connections**
Resolves#1627
**Description**
Works around Mesa fastclear bug by doing a manual shader clear on effected platforms
**Testing**
Tested on Mesa Intel(R) UHD Graphics (CML GT2) (Gl)
Co-authored-by: Zicklag <zicklag@katharostech.com>
I had to do a bit of digging to find information about coordinate systems in
wgpu (asking on Matrix). This commit adds a table of coordinate systems to help
clarify the issue.
The table comes from the DX12 backend documentation of gfx-rs:
https://github.com/gfx-rs/gfx/tree/master/src/backend/dx12
Remove trailing newlines from README
Host coordinate images locally
1721: Tweak Water Example to Show Off Water Shader a bit More r=cwfitzgerald a=zicklag
**Connections**
None
**Description**
Tweaks the water shader example terain and camera to show off the cool water shader a little bit more. 😄https://user-images.githubusercontent.com/25393315/126842645-e353c0eb-57fc-4b8e-9e0d-901f08c7e9f7.mp4
**Testing**
Tested on vulkan and GL on linux.
Co-authored-by: Zicklag <zicklag@katharostech.com>
1719: hal/dx12: pass null to mapping rects r=kvark a=kvark
**Connections**
See https://github.com/baldurk/renderdoc/issues/2323
**Description**
Passing NULL rects is better for tooling, as it turns out.
**Testing**
Ran the examples
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1710: Check render bundle encoding for RODS r=cwfitzgerald a=kvark
**Connections**
Matches https://github.com/gpuweb/gpuweb/pull/1979
**Description**
Allows render bundles to be encoded for RODS.
In addition, validates that bundle RODS matches the pass RODS when executing. This part of the spec isn't written yet, and it's safer to be on the strict side.
**Testing**
Untested
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1718: Update naga to bbfa9a0, tweak DX12 debug output r=kvark a=kvark
**Connections**
Picks up #1132
**Description**
Also reduces the spam output of DX12 backend
**Testing**
Got skybox and water examples running!
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1714: Update Naga to cc91c77 r=kvark a=kvark
**Connections**
Picks up https://github.com/gfx-rs/naga/pull/1128 and other HLSL fixes.
Complements #1713
**Description**
Just an update
**Testing**
Ran examples locally
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1713: hal/dx12: rewrite view creation, use arrays more aggressively r=kvark a=kvark
**Connections**
**Description**
D3D12 has this thing where you can't create a 2D view with non-zero base array index.
So we have special logic now to fall back to 2D array views in these cases.
**Testing**
Tested on the Shadow example.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1704: Fix buffer memory initialization to use the tracker consistently r=kvark a=kvark
**Connections**
Depends on https://github.com/gfx-rs/naga/pull/1125
**Description**
Contains a number of dx12 fixes in addition to the buffer initialization fix.
The problem there was - we used the "pending writes" command buffer to initialize buffers, but also used the device's resource tracker and its state to figure out the transitions *after* a command buffer is merged its states.
**Testing**
Ran the examples on Vulkan.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1705: Fix failure mode of adapter_request_device (direct backend) r=kvark a=dhardy
This function has a method for reporting errors, so it shouldn't panic IMO.
Possibly though we should adapt the `RequestDeviceError` type to have a `cause: Box<dyn Error>` payload? The `log::error` line is just because we have no other way of reporting this value.
I tried to test this, but...
- on v0.9 I couldn't make the request fail, even with requesting absurd limits like 1e9 something
- on master the adapter request succeeds, then the shader creation fails with `UnsupportedCapability(Float64)` (even though the request contains `wgpu::Features::SHADER_FLOAT64` and does not fail)
So the `direct` or Vulkan backend does not check features properly?
Co-authored-by: Diggory Hardy <git@dhardy.name>
1703: Add Rgb9e5Ufloat Format r=kvark a=cwfitzgerald
**Connections**
Addresses part of #967, a wgpu-hal followup to #955.
**Description**
This adds only `Rgb9e5Ufloat`. I removed the explicit numbering from the texture format so we don't have to constantly re-number things. The default numbering is identical to what we were doing.
**Testing**
All backends unconditionally support it, so little testing is needed.
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
