Commit Graph
8897 Commits
Author SHA1 Message Date
bors[bot]andDzmitry Malyshau c691570c60 Merge #1127
1127: Fix API traced dynamic offsets and markers r=critical a=kvark

**Connections**
This was a regression from #1106, discovered while investigating #1123

**Description**
When running a pass, it was mutably used for "convenience". Once we moved the tracing of that pass from the beginning of the function to the end, the "convenience" eliminated our dynamic offsets and markers.
The solution is to stop mutating the passes inside our big functions.

**Testing**
Re-traced Terra

Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
2021-01-04 22:58:50 +00:00
Dzmitry Malyshau d70148725b Fix API traced dynamic offsets and object labels 2021-01-04 17:54:56 -05:00
bors[bot]andDzmitry Malyshau 66c987b226 Merge #1126
1126: Check swapchain frame to still be in use r=harmless a=kvark

**Connections**
Fixes #1123

**Description**
What happened there is - the same command buffer was used to record work into multiple swapchain frames.
The ID management of those is a bit special, so the error wasn't very descriptive, internal logic caught inconsistency.
Now there is a proper error, and the old error is better.
Note that ideally we'd want "present" to work like "destroy" on textures, but essentially the users would have to fix it anyway. It's not going to make their life easier (in fact - only harder), so we can defer this to a follow-up.

**Testing**
Tested on wgpu-rs examples as well as on Terra.

Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
2021-01-04 22:12:19 +00:00
Dzmitry Malyshau 922f2c0977 Check swapchain frame to still be in use 2021-01-04 17:03:16 -05:00
bors[bot]andDzmitry Malyshau 35a2c8c474 Merge #1125
1125: Fix UB in populating dynamic offsets r=kvark a=kvark

**Connections**
Related to the traces in #1123

**Description**
The dynamic offsets arrays appear to be empty in the traces, which makes no sense. We are populating them unconditionally, and I double-checked on the `shadow` example that the traces we record contain the offsets. So my only guess is that the reported traces are taken from a release build, and our code that does `slice::from_raw_parts()` on the offsets is UB because the pointer can be garbage (if length is 0). This is what the PR fixes.

**Testing**
Untested

Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
2021-01-04 19:49:21 +00:00
Dzmitry Malyshau f328995950 Fix UB in populating dynamic offsets 2021-01-04 14:46:04 -05:00
bors[bot]andDzmitry Malyshau 29cb9e687d Merge #1124
1124: Fix range tracker merging r=kvark a=kvark

**Connections**
Fixes #1123 

**Description**
The merge routine was not properly updating the end pointer in some branches.

**Testing**
A unit test is added with the data extracted from the user case.

Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
2021-01-04 18:42:02 +00:00
Dzmitry Malyshau a7e573d612 Fix range tracker merging 2021-01-04 13:41:25 -05:00
Pelle Johnsen 6710f5954e Fix clippy issues 2021-01-04 09:51:32 -05:00
Pelle Johnsen e062096a6a [spv-out] Handle import id for GLSL.std.450 2021-01-04 09:51:32 -05:00
bors[bot]andDzmitry Malyshau b82482f31c Merge #1121
1121: Implement render bundle freeing r=straightforward a=kvark

**Connections**
Fixes #1120

**Description**
Since render bundles were added the relevant bits of logic to free them were just missing.

**Testing**
Untested

Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-01-04 14:32:45 +00:00
Dzmitry Malyshau 5953098461 Implement render bundle freeing 2021-01-04 09:30:10 -05:00
bors[bot]andManuel Gil 7f0b8f3e34 [rs] Merge #693
693: Expose get_swap_chain_preferred_format r=kvark a=manugildev

Fixes #691

This functionality was committed in gfx-rs/wgpu@6f1d614 but never exposed to `wgpu-rs`.

Problem: Examples made use of `wgpu::TextureFormat::Bgra8UnormSrgb` but that's unreliable and according to the [API](https://gpuweb.github.io/gpuweb/#dom-gpucanvascontext-getswapchainpreferredformat) this method is the way of properly obtaining an optimal texture format for the `SwapChain` based on the platform.

- [x]  Expose `get_swap_chain_preferred_format()` on Device
- [x]  Use `TextureFormat::Bgra8Unorm` for [web.rs](https://github.com/gfx-rs/wgpu-rs/blob/e00ef5d5842a1fcfb46bb531789b1c340bdee3ab/src/backend/web.rs#L970)
- [x]  Update [framework.rs](https://github.com/gfx-rs/wgpu-rs/blob/e00ef5d5842a1fcfb46bb531789b1c340bdee3ab/examples/framework.rs#L227) and examples to call this new method

Co-authored-by: Manuel Gil <manugildev@gmail.com>
2021-01-04 14:12:35 +00:00
bors[bot]andDzmitry Malyshau d09b149f6d Merge #1119
1119: Lock device tracker after the resources r=kvark a=kvark

**Connections**
Fixes  #1118

**Description**
Added a comment above `struct Device` explaining the problem and constraints.
We may have a few other cases that need to be handled gracefully by the type system.
For this fix we at least ensure that the device tracker is always locked temporarily.

**Testing**
Untested.

Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-01-04 14:10:11 +00:00
Manuel Gil 81bcc72651 [rs] Expose get_swap_chain_preferred_format on Device
- Update examples to make use of the new method
2021-01-03 17:35:04 +00:00
Dzmitry Malyshau 2b02b7c8e4 Lock device tracker after the resources 2021-01-03 11:29:18 -05:00
bors[bot]andJonathan Behrens 36d274bc5a Merge #1115
1115: Request buffer FAST_DEVICE_ACCESS only with some non MAP or COPY usage r=kvark a=fintelia

**Connections**
Fixes #1113 

**Description**
On Linux AMD (and likely elsewhere) buffers with FAST_DEVICE_ACCESS are very slow to access when mapped. This makes sure that upload/download buffers don't request it unless those buffers are leveraging MAPPABLE_PRIMARY_BUFFERS for some non-map/copy usage.

**Testing**
I reran the failing reproduction case in the linked issue.


Co-authored-by: Jonathan Behrens <fintelia@gmail.com>
2021-01-03 03:32:08 +00:00
bors[bot]andDzmitry Malyshau b67c2b0299 Merge #1117
1117: Free associated staging buffers r=cwfitzgerald a=kvark

**Connections**
Reported on the Matrix. Having a loop that creates mapped buffers and drops them results in OOM.

**Description**
We are now checking for the map state when destroying buffers.

**Testing**
Ran through an ad-hoc example.

Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-01-02 23:55:05 +00:00
Dzmitry Malyshau d095dd39b7 Free associated staging buffers 2021-01-02 18:52:07 -05:00
Jonathan Behrens 0e84c29691 Limit when buffer FAST_DEVICE_ACCESS is requested 2021-01-02 14:18:14 -05:00
bors[bot]andAndreas Reich e9232127f0 Merge #1114
1114: Fixed error message for BindingTypeMaxCountError r=cwfitzgerald a=Wumpf

Previously, `BindingTypeMaxCountError` would incorrectly report the binding count as the violated limit. Now it reports both the limit and the count that violated it.

Example error message before:
```
wgpu error: Validation Error

Caused by:
    In Device::create_bind_group_layout
      note: label = `BindGroupLayout write scalar`
    too many bindings of type StorageTextures in stage COMPUTE, limit is 5
```


Example error message after:
```
wgpu error: Validation Error

Caused by:
    In Device::create_bind_group_layout
      note: label = `BindGroupLayout write scalar`
    too many bindings of type StorageTextures in stage COMPUTE, limit is 4, count was 5
```

Co-authored-by: Andreas Reich <r_andreas2@web.de>
2021-01-02 15:08:09 +00:00
Andreas Reich 9770959577 Fixed error message for BindingTypeMaxCountError
Previously, it would incorrectly report the binding count as the violated limit. Now it reports both the limit and the count that violated it.
2021-01-02 15:13:52 +01:00
Andreas Reich 757d12aca3 map depth/stencil feature to RENDER_ATTACHMENT 2020-12-30 12:23:13 +01:00
Andreas Reich c17b2dc6a0 Expose adapter_get_texture_format_features 2020-12-30 12:00:00 +01:00
Andreas Reich 4207e57c02 validation against adapter specific format features
TextureFormatFeatures is now a struct of two u32 flags. Query happens only at texture creation time
2020-12-30 10:38:55 +01:00
Andreas Reich 0ccad9a466 wip adapter specific format features & rw storage 2020-12-29 17:14:22 +01:00
bors[bot]and0x182d4454fb211940 ee3ca89105 [rs] Merge #690
690: Add `RenderEncoder` r=kvark a=0x182d4454fb211940

It seems like `RenderPass` and `RenderBundleEncoder` share many methods in common (`set_pipeline`, `draw_indexed`, etc). By creating a shared trait, it allows more reusable code. For instance:

```rust
impl RedCircle {
    fn new(...) -> Self { .. }
    fn draw<'a>(&'a self, encoder: &mut impl RenderEncoder<'a>) { .. }
}
```

This code would then work for both structs.

Moving the previous methods from their `impl`s to the trait's would break previous code, as to use these methods code would now have to import `RenderEncoder`. To avoid this, it is implemented by calling the direct methods.

It may be worth considering moving this to `wgpu::util`.

Co-authored-by: 0x182d4454fb211940 <nathanwoodformal@gmail.com>
2020-12-28 01:19:59 +00:00
bors[bot]andConnor Fitzgerald a90d2fcf8a Merge #1101
1101: Bump gfx-hal and gpu-alloc r=kvark a=cwfitzgerald

**Connections**

Absorbs https://github.com/zakarumych/gpu-alloc/pull/31 and https://github.com/gfx-rs/gfx/pull/3548#issuecomment-749158016

**Description**

This fixes the last of the DX11 issues listed in #1059.

**Testing**

Various examples ran.


Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2020-12-27 06:11:06 +00:00
Connor Fitzgerald 79973aaaff Bump gfx-hal and gpu-alloc 2020-12-27 01:10:24 -05:00
0x182d4454fb211940 cbbf25c6f8 [rs] Add set_push_constants to RenderEncoder 2020-12-25 14:01:04 +00:00
0x182d4454fb211940 a2ff43f989 [rs] Remove unneeded cast from encoder.rs 2020-12-25 13:50:36 +00:00
0x182d4454fb211940 8dc1c0896b [rs] Move RenderEncoder to util 2020-12-25 13:47:38 +00:00
0x182d4454fb211940 6ee00ab401 [rs] Add RenderEncoder 2020-12-24 17:37:06 +00:00
Pelle Johnsen 6c8cdea936 [glsl-in] Check number of function args 2020-12-23 14:54:53 -05:00
Pelle Johnsen 7d9054f2c1 [glsl-in] Implement more MathFunctions 2020-12-23 14:54:53 -05:00
Pelle Johnsen 9a51d82c39 [glsl-in] Implement ceil
Related to #330
2020-12-23 14:54:53 -05:00
Pelle Johnsen 511000d022 [glsl-in] Extract function logic out of parser 2020-12-23 14:54:53 -05:00
bors[bot]andDzmitry Malyshau aded405479 Merge #1106
1106: Encode render passes while command buffer is only read-locked r=cwfitzgerald a=kvark

**Connections**
Could fix #1104 to some extent

**Description**
The heaviest part of the render pass encoding doesn't need a write lock. Only in the end we need it.

**Testing**
Untested.

Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2020-12-23 16:21:03 +00:00
Dzmitry Malyshau ab10d36ac0 Rename get_encoder to get_encoder_mut 2020-12-23 11:20:08 -05:00
bors[bot]andJoshua Groves 4add83a8c5 Merge #1107
1107: Use `repr(transparent)` for shader flags r=cwfitzgerald a=grovesNL

**Connections**
None

**Description**
This allows cbindgen to generate bindings for shader flags in wgpu-native.

**Testing**
After `WGPUShaderFlags` is exposed in wgpu-native, bindings for `WGPUShaderFlags` will be generated in wgpu.h

Co-authored-by: Joshua Groves <josh@joshgroves.com>
2020-12-23 11:33:39 +00:00
Joshua Groves 45b5ddcb83 Use repr(transparent) for shader flags
This allows cbindgen to generate bindings for shader flags in
wgpu-native
2020-12-23 06:43:15 -03:30
Dzmitry Malyshau a7587b41e6 Move the render pass setup logic into a separate method/struct 2020-12-22 22:30:27 -05:00
Dzmitry Malyshau 35b163ebc8 Encode render passes while command buffer is only read-locked 2020-12-22 13:48:18 -05:00
Capucho 3f7ee8cd6e Generate locations with fixed naming scheme for linking 2020-12-22 13:02:22 -05:00
bors[bot]andVincent Jousse f1604cdc57 [rs] Merge #684
684: First step for WebGL support. r=grovesNL a=VincentFTS

This permits to test WebGL backend in Wgpu-rs.
Next step is to test examples to make them work with WebGL backend !

Co-authored-by: Vincent Jousse <contact@ftsoftware.fr>
2020-12-22 17:48:23 +00:00
Vincent Jousse 9f3a050fc4 [rs] First step for WebGL support. 2020-12-22 18:41:07 +01:00
bors[bot]andConnor Fitzgerald 2287ae3f8a Merge #1103
1103: Fix Assorted Low-Hanging Validation and Error Message Issues r=kvark a=cwfitzgerald

**Connections**

Closes #1085
Closes #393
Closes #1053

**Description**

These commits are independent and should be reviewed individually. Combined into a single PR to reduce noise.

Overview of what was done:
- Add validation for empty texture and buffer usage flags. (#393)
- Add allowed texture usage flags to `format.describe()` (#1085)
  Validate new textures follow the allowed usage flags.
- Properly validates vertex and buffers are bound. (#1053)
  Improves error messages when no vertex buffer is bound. (Before it said the limit was 0, now it says something is unbound)
- Improve the vertex buffer overrun messages by keeping track of which slot has the smallest index.

**Testing**

Tested on examples by artificially creating the situation I am trying to validate, as well as running clean examples to make sure they pass validation.

Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2020-12-22 04:29:38 +00:00
Connor Fitzgerald 1da153f3bd Improve vertex state error messages 2020-12-21 23:13:34 -05:00
Connor Fitzgerald 94579d8125 Properly validate vertex buffer state 2020-12-21 23:10:53 -05:00
Connor Fitzgerald 43814d258f Add format usage descriptions and validate on it. 2020-12-21 23:08:31 -05:00