Commit Graph
8323 Commits
Author SHA1 Message Date
bors[bot]andDzmitry Malyshau b5a73b6a80 Merge #1600
1600: Vulkan Timeline Semaphores r=kvark a=kvark

**Connections**
None

**Description**
One of the non-straightforward design decisions in wgpu-hal was to expose the fences with `D3D12Fence` semantics. Now that we got Vulkan, Metal, and GLES3, it became a bit concerning that all of these backends implemented this new fence API in the same convoluted way: via managing a pool of binary fences internally, which requires maintenance (done on submit() only). If that's the expected implementation, then the abstraction is wrong.

For this reason, it was important for me to try implementing this new semantics in a modern Vulkan way - via timeline semaphores. With this, we have 1:1 relation between `Api::Fence` and a Vulkan object. I believe the experiment is a success: Vulkan is now very efficient at fences!

**Testing**
Works on the examples!


Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
2021-07-04 04:57:40 +00:00
Dzmitry Malyshau 3de1681eb0 hal/vk: enable timeline semaphores 2021-07-04 00:50:05 -04:00
Dzmitry Malyshau 8ae7d51c82 hal/vk: request timeline semaphores if available 2021-07-04 00:50:05 -04:00
bors[bot]andDzmitry Malyshau 93eb33cf3b Merge #1598
1598: Fix memory leak in acquiring encoders r=kvark a=kvark

**Connections**
Fixes #1501 
Fixes #1531

**Description**
It was ... interesting. We were creating new command encoders all the time, but still trying to add them to the pool to re-use later. So we were constantly allocating and never freeing them, up until the end. Therefore, program as a whole was valid and non-leaking, if analyzed after termination.

**Testing**
Observed memory manually


Co-authored-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
2021-07-04 02:34:10 +00:00
Dzmitry Malyshau b704a0542f Fix memory leak in acquiring encoders 2021-07-03 22:25:59 -04:00
Connor Fitzgerald e2bcb72bf4 Please clippy 2021-07-03 22:15:34 -04:00
Connor Fitzgerald 861e84b178 Start adjusting CI 2021-07-03 22:15:16 -04:00
Connor Fitzgerald 26f3c03150 Add downlevel limit for vertex shader access to SSBO 2021-07-03 22:15:16 -04:00
Connor Fitzgerald 93cca6d182 Fix limits for rpi on vk 2021-07-03 22:15:16 -04:00
Connor Fitzgerald a13b54ced8 Fix limits for llvmpipe 2021-07-03 22:15:16 -04:00
bors[bot]andGordon-F 762cdb3953 Merge #1589
1589: [GL] Fix min_storage_buffer_offset_alignment param r=kvark a=Gordon-F

**Connections**
Android Emulator (11) error:
```
2021-07-02 14:49:12.500 8292-8317/rust.example.cube E/emuglGLESv2_enc: device/generic/goldfish-opengl/system/GLESv2_enc/GL2Encoder.cpp:s_glGetIntegerv:876 GL error 0x500
```

**Description**
As a workaround set this parameter to default value.


Co-authored-by: Gordon-F <ishaposhnik@icloud.com>
2021-07-04 00:33:32 +00:00
bors[bot]andDzmitry Malyshau fcba9325f8 Merge #1597
1597: hal/metal: bump storage limits r=kvark a=kvark

**Connections**
Fixes #1596

**Description**
The limits where just copied from the defaults. Then we updated the defaults...

**Testing**
Just ran locally


Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-07-04 00:27:35 +00:00
Dzmitry Malyshau a4a4d541b8 hal/metal: bump storage limits 2021-07-03 20:26:15 -04:00
bors[bot]andDzmitry Malyshau fd183a99ba Merge #1594
1594: Generation of memory reports for Hubs r=cwfitzgerald a=kvark

**Connections**
Part of #1501 

**Description**
Tried to see if we aren't accidentally growing the hubs.

**Testing**
Examples...


Co-authored-by: Dzmitry Malyshau <kvark@fastmail.com>
2021-07-03 21:45:57 +00:00
Gordon-F a077102457 [GL] Fix min_storage_buffer_offset_alignment param 2021-07-03 22:45:45 +03:00
Dzmitry Malyshau d2a4af35af Generation of memory reports for Hubs 2021-07-03 01:51:04 -04:00
bors[bot]andDzmitry Malyshau edbe3b33fa Merge #1593
1593: Fix VVL on halmark example r=kvark a=kvark

**Connections**
We have a bunch of VVL related to resource destruction on the halmark example.
It also seems to leak memory quite a ton and eventually hang the system.

**Description**
This PR fixes the logic of changing the command pools in halmark, its initialization sequence, and its destruction.

**Testing**
Tested on halmark and others


Co-authored-by: Dzmitry Malyshau <kvark@fastmail.com>
2021-07-03 05:13:18 +00:00
Dzmitry Malyshau e2f97ea659 hal/gles: detect llvmpipe instead of lavapipe 2021-07-03 01:09:39 -04:00
Dzmitry Malyshau 497c6647ee hal/vulkan: fix instance destruction logic 2021-07-03 01:07:21 -04:00
Dzmitry Malyshau b15f1ebfb7 halmark: fix the command pool recycling policy 2021-07-03 00:23:11 -04:00
bors[bot]andDzmitry Malyshau 7ab950e465 Merge #1592
1592: Detailed limits check and error r=cwfitzgerald a=kvark

**Connections**
Related to #1590

**Description**
The old comparison "A < B" was simply wrong, because it would do every field in order and could bail out mid-way without reaching for all limits.

**Testing**
Untested


Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-07-03 01:12:43 +00:00
Dzmitry Malyshau 81b382f5cd Readme and clippy fixes 2021-07-02 20:43:01 -04:00
bors[bot]andClayton Walker 62972c6c5f Merge #1591
1591: Expose calayer r=kvark a=Sineaggi

**Description**

This pr restores the ability to create a surface from a pre-existing CAMetalLayer. This is required for wgpu-native.

Co-authored-by: Clayton Walker <cwalker@sofi.org>
2021-07-03 00:23:25 +00:00
Dzmitry Malyshau 23b587246a Detailed limits check and error 2021-07-02 20:17:12 -04:00
Jasper St. Pierre 7a9fb864ad [glsl-in] Add support for gl_FragCoord / gl_FragDepth
And remove the ability to read gl_Position in fragment.
2021-07-03 00:29:44 +01:00
Clayton Walker 25c452cbdd Add metal layer back in 2021-07-02 15:51:11 -06:00
bors[bot]andDzmitry Malyshau df2b0b5843 Merge #1585
1585: hal/egl fix support for context extensions r=cwfitzgerald a=kvark

**Connections**
Follow-up to #1584
Closes  #1577

**Description**
TIL that EGL extensions work differently from the EGL-1.5 stuff, and we are now treating them properly.

**Testing**
Basic testing on AMD


Co-authored-by: Dzmitry Malyshau <kvark@fastmail.com>
2021-07-02 19:29:15 +00:00
Dzmitry Malyshau 67625b2a54 hal/egl fix support for context extensions 2021-07-02 15:23:06 -04:00
bors[bot]andDzmitry Malyshau 43c26e57e0 Merge #1588
1588: Update naga to 0b9af95 r=kvark a=kvark

**Connections**
Picks up https://github.com/gfx-rs/naga/pull/1060 for #1581

**Description**
Naga update

**Testing**
Untested


Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-07-02 06:00:13 +00:00
Dzmitry Malyshau ba8fcb937d Update naga to 0b9af95 2021-07-02 01:58:19 -04:00
bors[bot]andDzmitry Malyshau aee26e3e3e Merge #1587
1587: hal/metal: mark buffer immutability r=kvark a=kvark

**Connections**
This improves Metal performance.

**Description**
Extracting data from Naga to mark some buffers as immutable.

**Testing**
Just ran the examples.


Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>
2021-07-02 05:56:16 +00:00
Dzmitry Malyshau 34389396f9 hal/metal: mark buffer immutability 2021-07-02 01:52:28 -04:00
Jasper St. Pierre 0b9af95793 [glsl-in] Make sure to Load on a ValuePointer as well
Fixes issues with indexing vectors.
2021-07-02 01:44:57 -04:00
Jasper St. Pierre 78e1304d42 [glsl-in] Convert bool -> scalar cast to Select 2021-07-02 01:34:34 -04:00
bors[bot]andConnor Fitzgerald 5c17063592 Merge #1586
1586: Assorted Fixes of Testing-Infra r=kvark a=cwfitzgerald

**Connections**

Couple of low hanging fruit issues solved here. The biggest thing done is moving the masks to associated functions instead of being members of the bitflags, this is done so `all()` only includes the actual members.

Follows up on 579de425e5 to mark vk/mipmap as passing


Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2021-07-02 05:19:36 +00:00
Connor Fitzgerald 1fb77c68da Mark vk mipmap example as passing 2021-07-02 01:15:23 -04:00
Connor Fitzgerald 53e2c95dc8 Fix wgpu-info display by moving masks to associated functions 2021-07-02 01:15:19 -04:00
Connor Fitzgerald 96375dfa1b Fix display of test failure reason 2021-07-02 00:57:19 -04:00
bors[bot]andConnor Fitzgerald 871cbed18a Merge #1584
1584: Update testing limits for rpi4 r=kvark a=cwfitzgerald

**Connections**

Related to #1574. Exposed #1583.

**Description**

Changes the reftesting limits to be tolerant of how the RPI4 renders. Fixes the debug half of #1577. 

Fixes some naming issues as well.

**Testing**

It is


Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2021-07-02 04:34:55 +00:00
Dzmitry Malyshau a92f7689f2 [mtl] detect binding incomatibility 2021-07-02 00:25:48 -04:00
Connor Fitzgerald 2bdc54ad32 Formatting woes 2021-07-02 00:13:33 -04:00
Connor Fitzgerald a336c9e160 Spell tolerance right 2021-07-02 00:13:33 -04:00
Connor Fitzgerald fc03160d19 Fix backend_failure plurality 2021-07-02 00:13:33 -04:00
Connor Fitzgerald 6c79b71d87 Mention --no-fail-fast in readme 2021-07-02 00:13:33 -04:00
Connor Fitzgerald f76200e725 Update testing limits for rpi4 2021-07-02 00:13:33 -04:00
Dzmitry Malyshau 579de425e5 hal/vk: fix sampler default LOD range 2021-07-02 00:05:36 -04:00
bors[bot]andDzmitry Malyshau 9ee728bd53 Merge #1578
1578: Fix destruction of resources r=kvark a=kvark

**Connections**
Fixes #1552
Also fixes the VVL about swapchain frame destruction.

**Description**
Swapchain view was never registered in the device's root Hub. So the old logic of adding it to "suspected resources" didn't fire up correctly. The new logic goes straight into the submission tracker.
This path will cease to exist when either Hubs are removed, or API changes of https://github.com/webgpu-native/webgpu-headers/issues/89 are accepted.

The other resource destruction errors happened because we gathered all the resources that were abandoned by the user, and held alive only by the submitted command buffers, and we added them to the suspected list. Then we'd scan the list in `maintain`, see that they can be removed, and try to associate their destruction with one of the submissions. But the current submission was not added yet! So the logic thinks it can just remove the resources right away in this case, assuming the submission is not found because it's long gone in past.

**Testing**
Tested on our examples.


Co-authored-by: Dzmitry Malyshau <kvark@fastmail.com>
2021-07-02 02:36:01 +00:00
João Capucho 6e9410c056 [glsl-in] More implicit casts for builtins 2021-07-01 21:02:19 -04:00
João Capucho 1e2aee2263 [glsl-in] Clamp function fixes
Allow vector sources and scalar selector
Do implicit conversions
2021-07-01 21:02:19 -04:00
João Capucho e277cd1b12 [glsl-in] Fix shifts between vectors and scalars 2021-07-01 21:02:19 -04:00