710: Empty buffers are created internally with a size of 1 r=cwfitzgerald a=rukai
**Connections**
closes https://github.com/gfx-rs/wgpu/issues/709
**Description**
Empty buffers can be succesfully created.
**Testing**
Tested on my project and it resolves the above issue.
Co-authored-by: Rukai <rubickent@gmail.com>
708: Implement Mappable Primary Buffers Extension r=kvark a=cwfitzgerald
**Connections**
#675 made `MAP_WRITE | STORAGE` on buffers not possible. This extension re-enables it.
**Description**
UMA systems rejoice everywhere.
**Testing**
Hopefully it didn't break since it was usable a week or so ago, so this shouldn't need testing...
**Review Notes**
The name could be changed, particularly if primary has a certain meaning wrt buffers. It just seemed a reasonable description.
Knowing my luck, I got the bitflags call wrong...
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
707: Implement Bounds Checking on Buffers to Buffer and Texture Copies r=kvark a=cwfitzgerald
**Connections**
Closes#362, closes#554
This addresses issues found by @gretchenfrage when accidentally writing off the end of a texture causing a DEVICE_LOST error.
**Description**
Adds bounds checks to write_texture and write_buffer, and copy_texture_to_texture and friends. The bounds checking functions themselves follow guidance from the webgpu standard.
This doesn't make wgpu 100% to all the checks required by the standard, but takes care of the ones related to bounds and buffer overruns.
**Testing**
I tested this against all the examples, including intentionally making mistakes in texture copies, with this successfully catching those errors.
**Review Notes**
This code should be picked through fairly closely, as it's very likely there's some typos due to the close-but-not-quite repetition that these tests require. There's a bunch of LOC, but they are fairly boring and should be easy to understand.
I have tried to give assert messages that are as descriptive as possible, but if there is a message that could be changed to be more clear, let me know.
One thing that could change is the location of the functions. I left them where I originally wrote them, before I realized that the functions on queue also needed bounds checking, but they could have a better home elsewhere, maybe even their own file.
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
706: Return errors on create_render_pipeline r=cwfitzgerald a=kvark
This is a follow-up to #705 that switches `create_render_pipeline` to return `Result`.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
705: Shader input/output validation r=cwfitzgerald a=kvark
**Connections**
Fixes#269
It still has a few gaps (like the `storage_texture_format` matching, bugs, optional validation), but the main logic is there. Anything else should come in smaller issues as a follow-up.
**Description**
The main goal of this PR is to validate:
- vertex input stage against VS inputs
- VS outputs against FS inputs
- FS outputs against the pipeline attachment formats
I figured that `WGPU_SHADER_VALIDATION` environment is not a great path forward. It doesn't help engine authors, for example. So I'm switching it to just a boolean field in `DeviceDescriptor`. Hopefully, we'll remove it soon :)
**Testing**
Just running wgpu-rs examples - https://github.com/gfx-rs/wgpu-rs/pull/354
Review notes: the commit in the middle just moves stuff around. I think it's easier to just review the first and the last commit, ignoring the middle.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
704: Pipeline layout validation r=cwfitzgerald a=kvark
**Connections**
Implements a solid part of #269
Starts converting the function to return results, related to #638
cc @GabrielMajeri
**Description**
This change matches shader bindings against the pipeline layout. It's *mostly* complete, minus some bugs and not handling the `storage_texture_format` properly.
The risk here is that Naga reflection may have bugs, or our validation may have bugs, and we don't want to break the user content while this is in flux. So the PR introduces an internal `WGPU_SHADER_VALIDATION` environment variable. Switching it to "0" skips Naga shader parsing completely and allows the users to unsafely use the API.
Another aspect of the PR is that some of the functions now return `Result`. The way I see us proceeding is that any errors that we don't expect users to handle should result in panics when `wgpu` is used natively (i.e. not from a browser). These panics would happen in the "direct" backend of wgpu-rs (as well as in wgpu-native), but the `Result` would not be exposed to wgpu-rs, so that it matches the Web behavior.
At the same time, browser implementations (Gecko and Servo) will check the result on their GPU process and implement the WebGPU error model accordingly. This means `wgpu-core` can be super Rusty and safe.
**Testing**
Running on wgpu-rs examples. Most of them fail to get parsed by Naga, but `boids` succeeds and passes validation 🎉
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
703: Implement Extensions Interface r=kvark a=cwfitzgerald
## Description
This is the first step of implementing #691. This implements the described changes to all the core structs and to `SamplerDescriptor`. I intend this PR to be an example of what the plan looks like. If we think this works, we can roll out the rest of the changes either in bulk or as we need them.
The rolling out of this is also tied to the rolling out of #689. Hopefully the macro work done in https://github.com/gfx-rs/wgpu-native/pull/34 will make this easier.
## Questions
One outstanding question I had is what the default values for lod_min/max should be. As of right now I just derive default, which puts them at zero, which is probably a reasonable default, though most of the examples use -100, 100 which is normally what you use when doing mipmapping.
## TODO:
- [ ] Discuss if this meets our needs
- [x] Implement this in wgpu-rs (https://github.com/gfx-rs/wgpu-rs/pull/350)
- [ ] Implement this in wgpu-native
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
702: Fix SwapChainOutput Related Segfault r=kvark a=cwfitzgerald
If a user accidentally recreates a swapchain after grabbing a SwapChainOutput, then renders to it, this will trigger a segfault on render. This can be reproduced by replacing the get_next_frame code in any example with:
```rust
let frame = swap_chain
.get_next_frame()
.expect("Failed to acquire next swap chain texture")
.output;
device.create_swap_chain(&surface, &sc_desc);
```
I have tested the examples, and they still work after this change, and it correctly detects this condition.
Oddly enough the following code does not panic, nor actually raise any errors, even with this PR.
```rust
let tmp = device.create_swap_chain(&surface, &sc_desc);
let frame = swap_chain
.get_next_frame()
.expect("Failed to acquire next swap chain texture")
.output;
swap_chain = tmp;
```
I'm not entirely sure why.
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
701: Prevent internal thread from cleanup r=kvark a=kvark
**Connections**
Fixes#700
**Description**
The logic that removes command allocators for threads that no longer use it was accidentally removing the one for the "main" thread.
**Testing**
Untested
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
698: Add User Matrix Link r=kvark a=cwfitzgerald
Gone are the days of people not knowing there are two matrixes (matrices?)
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
696: Rustification of Extensions and SamplerDescriptor r=kvark a=cwfitzgerald
**Connections**
This begins the rustificaiton of `wgpu-types` as discussed in #689, starting with `Extensions` and `SamplerDescriptor`.
#691 and the resulting discussion was used to advise the shape of the extension struct.
**Description**
The PR should be fairly straight forward. As discussed, I have left extensions as a bitflag until the webgpu-native issue can be opened and discussed regarding allocation of extensions.
The most controversial part of this will be the `AnisotropyLevel` enum. I created it for two reasons, one that matters, one that doesn't:
- It means that, with the exception of enabling the aniso extension (and lod_bias), the sampler is correct by construction, which is helpful to both beginners and experts. It also better exposes the range of valid values and means less panics in user code.
- It saves a byte in the `Option<u8>` 😄
I definitely think that, if at all possible, we should have explicitly typed enums for "numbers" which have a limited amount of valid values (so _not_ lod bias, though that may be better expressed, idk) to make it very explicit which values can be accepted. This also feels more "rusty" and reduces the amount of "magicness" in the interface.
Ofc, I'm not married to the idea.
**Testing**
Follow up PR into wgpu-rs will include conversion of the examples.
**TODO**
- [x] wgpu-native pr rolling up this PR and #690 (https://github.com/gfx-rs/wgpu-native/pull/34)
- [x] wgpu-rs pr updating examples with the new structures (https://github.com/gfx-rs/wgpu-rs/pull/345)
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
695: Fix tracing of buffers that are mapped at creation r=kvark a=kvark
**Connections**
Fixes API tracing for the new API introduced in #675
**Description**
When buffers are mapped at creation, we need to create a staging buffer and copy into them. This implies COPY_DST, which isn't visible to the user. And this only applies to buffers without MAP_WRITE, in which case we are just mapping directly.
**Testing**
Tested locally on the `cube`. It would be great to automate API trace testing on CI.
Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
685: Read-only depth-stencil support (RODS) r=cwfitzgerald a=kvark
**Connections**
Fixes#680
- [ ] wgpu-native update
- [x] wgpu-rs update - https://github.com/gfx-rs/wgpu-rs/pull/340
**Description**
~~Interestingly, this requires us to create an extra pipeline state for RODS-compatible render pipelines (one main, and one with read-only depth-stencil). This also means we are creating another compatible render pass, internally (done once).~~
**Testing**
I don't know for sure that this works properly, but I'd be comfortable landing this at least if the existing functions aren't broken.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
694: player: enable x11 on Unix/Vulkan r=kvark a=kvark
Needed as it's no longer default in `wgpu-core` with #678
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
679: Enforce copy buffer-texture alignment r=cwfitzgerald a=kvark
**Connections**
This is a follow-up to #666
**Description**
We are now enforcing the `bytes_per_row` on copy-texture copies to `COPY_BYTES_PER_ROW_ALIGNMENT`.
We allow it being non-aligned for `write_texture`, which now has the code to properly align the staging space it uses, and to copy the rows one by one with proper alignment.
We are also moving `BufferSize` to wgpu-types, because it's required for wgpu-rs to build.
**Testing**
Testing this needs https://github.com/gfx-rs/wgpu-rs/pull/328, which is blocked by https://github.com/gfx-rs/wgpu-rs/pull/323
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
690: Implement Anisotropic Filtering Extension and Expose Extension Interface r=kvark a=cwfitzgerald
**Connections**
Resolves#568.
**Description**
This PR solves two tightly related problems:
- Adds the ability to create a sampler with anisotropic filtering
- Adds the webgpu interface for getting the limits/extensions on adapters and devies.
**Testing**
I have tested this against my codebase which uses it. Adding an example in wgpu-rs on how to use extensions/limits would be good, especially as we have native/webgpu extensions, but can come as a future enhancement, once we feel the extension situation has stabilized a bit.
**TODO**
- [x] Allow wgpu-rs to call limit/extension interface (https://github.com/gfx-rs/wgpu-rs/pull/338)
- [ ] Determine how wgpu-native is going to adopt this functionality.
- [ ] After discussing #691, what changes need to be made.
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
688: Fix recycling of command buffers r=kvark a=kvark
**Connections**
Fixes https://github.com/gfx-rs/wgpu-rs/issues/333
**Description**
We used `lowest_common_submission()` for figuring out when to reset command buffers, but it returns !0 when there are no active submissions. Instead, we are using the exising "last done" semantics here, which works better.
**Testing**
Tested on wgpu-rs examples.
Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
686: Move BufferSize to wgpu-types r=kvark a=kvark
This is required to be able to update wgpu-rs, since it uses `BufferSize` in the internal API for its backends.
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
678: Make libx11 optional r=kvark a=parasyte
**Connections**
Fixes#676
**Description**
Make `libx11` optional with a new non-default feature. Default features need to go "to the edge". In its current state, a crate depending on `wgpu` cannot disable the `x11` feature on `gfx-backend-vulkan`.
- `wgpu` will be updated with a similar new feature, but will enable the `wgc/x11` feature by default. This should hopefully be backward compatible for any downstream user of `wgpu`.
- `wgpu-native` will also be given the same treatment.
**Testing**
This was tested with repro steps included in the linked ticket.
Co-authored-by: Jay Oster <jay@kodewerx.org>
667: Add asserts to validate that a resource exists when used r=kvark a=DavidPeicho
Hi,
As discussed in #610, this is mostly for C/C++ users. Some questions:
* Would it be possible to display the ID of the resource that failed? I don't know what you think about implementing the `Display` trait on the `Id` struct?
* Do you see other places needing for checks?
Thanks!
Co-authored-by: David Peicho <david.peicho@gmail.com>
668: Return failures to the user in swap_chain_get_next_texture, rather than transparently reconfiguring. r=kvark a=AlphaModder
TODO:
- [x] Change `Global::swap_chain_get_next_texture` in `wgpu-core`.
- [x] Update `wgpu_swap_chain_get_next_texture` in `wgpu-native`. (https://github.com/gfx-rs/wgpu-native/pull/32)
- [x] Wrap `SwapChainOutput`/`SwapChainStatus` in a nice enum in `wgpu-rs`. (https://github.com/gfx-rs/wgpu-rs/pull/323)
- [ ] Update `wgpu_bindings` (?)
Co-authored-by: AlphaModder <quasiflux@gmail.com>
