1161: Fix replaying issues with dropped texture views r=urgent a=kvark
**Connections**
Fixes replaying of https://github.com/gfx-rs/wgpu/issues/1158#issuecomment-763272796
**Description**
The main fix is calling `maintain` in the player before getting the new view from the swapchain. That allows the device to properly remove the ID that is just about to be used.
**Testing**
Tested on that API trace.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1155: fix the size checks in transfer buffer to image commands r=cwfitzgerald a=Vengarioth
**Connections**
_none_
**Description**
It's fixing a trivial bug in which Extend3d was checked against a with of 0 two times, presumably instread checking depth as well.
**Testing**
Ran the tests and ran it with bevy - still worked.
Co-authored-by: Vengarioth | Andreas Fischer <opensource@deviru.de>
716: Update Wgpu-Core r=trivial a=cwfitzgerald
This rollup fixes a pretty bad segfault that always occurs.
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
1151: Update gfx and naga to gfx-6 tag r=cwfitzgerald a=kvark
**Connections**
Fixes#1149
Regression from https://github.com/gfx-rs/gfx/pull/3573
**Description**
The physical devices held references to `RawInstance` in Vulkan backend, and we weren't cleaning them up.
Note: an alternative fix could be in `gfx-backend-vulkan` - to move the `Entry` into `RawInstance`, thus delaying the shutdown. I didn't go this way because that would be less explicit.
**Testing**
Tested on wgpu-rs examples
Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
1128: Add pipeline statistics and timeline queries r=kvark a=cwfitzgerald
**Connections**
Closes#721
**Description**
This adds Pipeline Statistics queries and Timeline queries. Shoutout to @z2oh for doing a good chunk of the work to enable this feature.
Currently blocked on both https://github.com/gfx-rs/gfx/pull/3563 and rollup of gfx-hal mutability changes, but is ready to review.
**Testing**
Tested in various ways on the hello-compute and hello-triangle example, tested validation extensively. Queries will be permanently added to the mipmap example once the wgpu-rs PR is finished.
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
1147: Update render pass and framebuffers logic to the latest gfx-hal r=cwfitzgerald a=kvark
**Connections**
Updates gfx-hal with the latest major changes of https://github.com/gfx-rs/gfx/pull/3571
**Description**
There is no need to track framebuffers at all any more! Since they don't contain images, we can keep them permanently.
There is no need to create and delete framebuffers for the swapchain either. Lots of non-trivial code is gone 🎉
The PR also removes one of the mutexes that the device holds.
**Testing**
Will be tested on wgpu-rs
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
713: Expose texture format feature query (and update to latest wgpu) r=kvark a=Wumpf
Exposes texture format feature query from https://github.com/gfx-rs/wgpu/pull/1112
Updating wgpu-core led to breaking change of moving get_swap_chain_preferred_format from device to adapter.
Co-authored-by: Andreas Reich <r_andreas2@web.de>
1148: Remove comment introduced in #1144 r=cwfitzgerald a=DasEtwas
Tested to be unnecessary by reading
Co-authored-by: dasetwas <dasetwas@dasetwas.dasetwas>
698: Format PassErrorScope pipelines r=kvark a=scoopr
Pairing for gfx-rs/wgpu#1132, formats the pipeline ids
Co-authored-by: Mikko Lehtonen <scoopr@iki.fi>
1145: Support trailing render passes after the swapchain rendering r=kvark a=kvark
**Connections**
Discussed on the matrix.
**Description**
Fixes a case where we have any other render pass in a command encoder that draws to the swapchain (after the pass that draws to the swapchain).
**Testing**
Tested on a modified hello-triangle example.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
1144: Fix and simplify attachment dimension mismatch check r=kvark a=DasEtwas
This also improves error messages with detailed dimension and attachment type name info which developers can hopefully recognize.
**Connections**
None
**Description**
Before this commit, depth/stencil attachments are not considered when evaluating the renderpasses' dimension, which will result in a "NoAttachment" error with a confusing display message.
**Testing**
Tested by running an application making use of color-only and depth/stencil-only attachment renderpasses.
Tested by running boids, hello-triangle and shadow examples.
Co-authored-by: DasEtwas <18222134+DasEtwas@users.noreply.github.com>
709: Move get_swap_chain_preferred_format from device to adapter r=kvark a=niklaskorz
See https://github.com/gfx-rs/wgpu/pull/1142. This PR applies the necessary changes for the move of `get_swap_chain_preferred_format` from device to adapter.
Co-authored-by: Niklas Korz <niklas@niklaskorz.de>
1142: Move get_swap_chain_preferred_format from device to adapter and check formats r=kvark a=niklaskorz
**Connections**
https://github.com/gfx-rs/wgpu-rs/issues/123#issuecomment-757831458
**Description**
Not all devices actually support `Bgra8UnormSrgb`. If this is the case, a similar format like `Rgba8UnormSrgb` can be used.
**Testing**
e5ebfa70a4/learn-wgpu/src/lib.rs (L137)
- Output on MacBook Pro 13" Intel 2020 macOS Metal: `Format: Bgra8UnormSrgb`
- Output on OnePlus 7 Pro Android Vulkan: `Format: Rgba8UnormSrgb`
Co-authored-by: Niklas Korz <niklas@niklaskorz.de>
1140: Enable storage read/write format feature r=kvark a=Wumpf
**Connections**
#1112
**Description**
Use `wgt::TextureFormatFeatureFlags::STORAGE_READ_WRITE` in `get_texture_format_features`.
This has been waiting for a gfx update which was enabled by @ElArtista 's workaround in #1136 \o/
Co-authored-by: Andreas Reich <r_andreas2@web.de>
707: Fix compilation on aarch64-apple-darwin r=kvark a=andykilroy
wgpu-rs couldn't compile on the Apple M1 (arm64) macs due to a compilation
problem in winit. winit 0.24.0 has the fix so upgrade the dependency.
Link to the fix's pull request in winit:
https://github.com/rust-windowing/winit/pull/1752
Co-authored-by: Andrew Kilroy <7545475+andykilroy@users.noreply.github.com>
1135: Make all textures support usage COPY_SRC and COPY_DST r=kvark a=fintelia
**Connections**
N/A
**Description**
In WebGPU, all textures should be allowed to have usages COPY_SRC and COPY_DST. However, right now when Features::TEXTURE_ADAPTER_SPECIFIC_FORMAT_FEATURES is enabled any features that the backend says can't be blitted to/from will not gain these capabilities. In particular, compressed textures may not support blitting while still allowing access via copies to/from buffers.
You can also see the defaults used when TEXTURE_ADAPTER_SPECIFIC_FORMAT_FEATURES is not requested:
8c00cbcc8e/wgpu-types/src/lib.rs (L1224-L1342)
**Testing**
I ran a [patched version of the wgpu-rs cube example](https://github.com/fintelia/wgpu-rs/tree/cube-bc4/examples/cube).
Co-authored-by: Jonathan Behrens <fintelia@gmail.com>
