Commit Graph
2124 Commits
Author SHA1 Message Date
Nicolas Silva b0e4734a7e Differentiate between device loss caused by an error and caused by dropping the device 2024-01-24 13:41:23 +01:00
Nicolas Silva 8d64915b3c Remove the lock around TextureView::parent (#5126) 2024-01-23 10:31:13 -05:00
wayneandAlphyr c4b5cc94ad don't panic if naga parsing of shader source fails (#5034)
* naga: glsl parser should return singular ParseError similar to wgsl

* wgpu: treat glsl the same as wgsl when creating ShaderModule

* naga: update glsl parser tests to use new ParseError type

* naga: glsl ParseError errors field should be public

* wgpu-core: add 'glsl' feature

* fix some minor bugs in glsl parse error refactor

* naga/wgpu/wgpu-core: improve spirv parse error handling

* wgpu-core: feature gate use of glsl and spv naga modules

* wgpu: enable wgpu-core glsl and spirv features when appropriate

* obey clippy

* naga: derive Clone in Type

* naga: don't feature gate Clone derivation for Type

* obey cargo fmt

* wgpu-core: use bytemuck instead of zerocopy

* wgpu-core: apply suggested edit

* wgpu-core: no need to borrow spirv code

* Update wgpu/src/backend/wgpu_core.rs

Co-authored-by: Alphyr <47725341+a1phyr@users.noreply.github.com>

---------

Co-authored-by: Alphyr <47725341+a1phyr@users.noreply.github.com>
2024-01-23 10:25:25 -05:00
Nicolas Silva 60a5739df2 d3d12: Propagate errors when closing command lists (#5125)
Before this commit, command lists that we failed to close were used anyway during submit, causing device loss.
2024-01-23 14:30:08 +01:00
Brad Werth ac8756c2d3 Release GPU resources from device.trackers, not from lifetime_tracker. (#5075) 2024-01-22 17:55:42 +01:00
John-John Tedro e5c62fb5bd vulkan: Replace fence with semaphore when acquiring surfaces (#4967) 2024-01-21 18:14:17 -05:00
Connor Fitzgerald 8b2098bd4e Release 0.19 (#5082) 2024-01-17 14:42:11 -05:00
Andreas ReichandEmil Ernerfeldt b8f27c7284 Expose maximum_frame_latency (#4899)
Co-authored-by: Emil Ernerfeldt <emil.ernerfeldt@gmail.com>
2024-01-17 12:46:34 -05:00
Connor Fitzgerald 2e38187954 Re-export public dependencies (#5063) 2024-01-16 19:21:51 +00:00
John-John Tedro 90b078f037 Improve ergonomics of gfx_select! (#5069) 2024-01-16 11:15:56 +01:00
dependabot[bot] c5a0b0b7b4 Bump the patch-updates group with 12 updates (#5061)
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-01-14 23:48:30 -05:00
Brad Werth f89bd3b978 Make DeviceLostClosure.from_c consume the closure before dropping it. (#5032)
This clarifies that the Rust and C-style callbacks/closures need to be
consumed (not called) before they are dropped. It also makes the from_c
function consume the param closure so that it can be dropped without
panicking.

It also relaxes the restriction that the callback/closure can only be
called once.
2024-01-14 02:53:40 -05:00
daxpedda 580340f2d3 Add cfg_aliases to wgpu-core and wgpu-hal (#5055) 2024-01-14 00:59:59 -05:00
Nicolas Silva 376d901d1e Triage suspected resources after triaging submissions (#5050) 2024-01-12 13:19:19 -05:00
Nicolas Silva 11c29c825f Prevent a few integer overflows (#5042)
* Prevent a few integer overflows

* Add a changelog entry

* use u64
2024-01-12 16:23:23 +01:00
Andreas Reich 4fd4a7142e Unify surface creation by introducing new SurfaceTarget enum (#4984) 2024-01-11 22:24:01 -05:00
Nicolas Silva bc65d84cdb Remove the Destroyed state from Storage (#4970)
* Remove the Destroyed state from Storage

It used to be how we handled destroying buffers and textures but we moved to different approach.

* Explicit check for destroyed textures/buffers in a few entry points

This used to be checked automatically when getting the resource from the registry, but has to be done manually now that we track we track the destroyed state in the resources.
2024-01-11 10:45:12 +00:00
Nicolas Silva b30c0c2e00 Remove the free_resources tracker (#5037) 2024-01-11 09:41:09 +00:00
Nicolas Silva c90f43bc2a Move the trace recording of destroy events out of triage_resources (#5036) 2024-01-11 00:48:25 -05:00
Nicolas Silva f27bb443c1 Don't crash if a texture is destroyed before queue submission (#5028) 2024-01-10 14:36:02 +01:00
Nicolas Silva 4400a58470 Fix a bounds check in remove_abandoned (#5024) 2024-01-09 11:24:18 -05:00
Nicolas Silva 37755b6985 Validate that a binding offset fits in the buffer 2024-01-09 16:07:47 +01:00
Nicolas Silva 58fe7eac48 Don't case the dynamic offsets count to u8 (#5026) 2024-01-09 15:41:39 +01:00
Nicolas Silva 1bc71eb8b2 Validate max_vertex_buffers in render bundles. (#5025) 2024-01-09 15:41:24 +01:00
Jim Blandy 8af6975d5e Prefer Arc::into_inner over Arc::try_unwrap. (#5018) 2024-01-08 19:24:33 -05:00
Jim Blandy a24bdbaa25 [wgpu-core] Minor doc fixes for initialization tracking. (#5017) 2024-01-08 16:46:57 -05:00
Nicolas Silva c6eea50b04 Track the actual size of the buffer for memory init (#5014) 2024-01-08 17:24:09 +01:00
Nicolas Silva 8358868526 Propagate errors when openning/closing a command encoder (#4999) 2024-01-06 10:14:38 +01:00
Nicolas Silva 4b82121501 Snatch the raw texture when destroying it 2024-01-05 11:42:28 +01:00
Nicolas Silva 95c451a3b8 Check thata bindgroup's textures are not destroyed before using it 2024-01-05 11:42:28 +01:00
Nicolas Silva defd43dff2 Rename Texture::as_raw into Texture::raw
For consistency with Buffer::raw.
2024-01-05 11:27:17 +01:00
Nicolas Silva bfe235a295 Put raw texture access behind snatch guards 2024-01-05 11:27:17 +01:00
Nick 00cf05c1d0 Fix Queue::write_texture, Fix DX12 write_texture_subset_2d and re-enable the test. (#4990) 2024-01-04 19:50:41 -05:00
Connor Fitzgerald d96d953025 Shorten Lock Lifetimes (#4976) 2024-01-04 09:24:54 +01:00
Connor Fitzgerald b989e56874 Use Custom Mesa for Building (#4977) 2024-01-03 22:49:40 -05:00
Connor Fitzgerald fd37dbfae9 Fix hang in multithreaded test (#4975) 2024-01-03 23:05:35 +01:00
Nathan AdamsandConnor Fitzgerald ec920e85d1 Fix incorrect ConfigureSurfaceError::TooLarge message (#4960)
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2024-01-02 15:32:51 -05:00
Connor Fitzgerald e26a853654 BGL Weak Pointer Deduplication Pool (#4927)
* Ho boy

* BGL pool finished

* Remove id32 feature

* Add BGL Test

* Iteration

* Working Dedupe

* Tests

* Iteration

* Re-iteration

* Hash Cleanup

* Add large slew of tests

* Whoops
2024-01-02 09:27:22 -05:00
Erich Gubler e1baa5a56e docs: inline document-features usage, remove dep.
Re-implements https://github.com/gfx-rs/wgpu/pull/4886 (CC @Wumpf)
without the `document-features` crate, which has issues integrating into
Firefox builds after being `cargo vendor`ed into its repository. This
issue is being tracked against
https://github.com/slint-ui/document-features/issues/20. Once resolved,
I expect that we will want to revert this PR in its entirety, since
`document-features` is still a good addition to `wgpu`'s documentation
story.

Internally, consensus has already been achieved for this change.
Firefox's ability to build unfortunately take priority over this
particular convenience. Hopefully, we won't have to compromise shortly!

I tested this by ensuring that the HTML output of our existing
`document_features::document_features!(…)` usage was exactly the same.
There should be exactly zero regressions in the current state of
documentation for users. For maintainers, I have added a disclaimer that
one needs to keep changes in sync. with the relevant `Cargo.toml`
manifests.
2023-12-22 14:31:12 -05:00
Andrew Farkas 2993650ede Fix typo "layout pipeline layout" -> "pipeline layout" (#4921) 2023-12-22 01:02:41 +00:00
Connor Fitzgerald 87ecc089dd Remove id32 feature 2023-12-20 21:39:24 -08:00
Brad WerthandConnor Fitzgerald 9eea31a4ae Eagerly release GPU resources when we lose the device. (#4851)
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2023-12-20 22:43:43 +00:00
Nicolas SilvaandConnor Fitzgerald 85b91c0408 Buffer snatching (#4896)
Co-authored-by: Connor Fitzgerald <connorwadefitzgerald@gmail.com>
2023-12-20 16:42:42 -05:00
Nicolas Silva 164e478fcb Simplify some code around buffer unmapping (#4892) 2023-12-20 15:21:32 -05:00
Nicolas Silva 0524c88c03 Check that raw buffers and raw bind groups are valid (#4895) 2023-12-20 15:13:20 -05:00
sotaroikedaandSotaro Ikeda 1d6c7a0263 Support Device fence sharing with dx12 on Windows (#4900)
Co-authored-by: Sotaro Ikeda <you@example.com>
2023-12-20 13:33:47 -05:00
Brad Werth 56d9d32a07 [wgpu-core] Ensure that DeviceLostCallback is always called exactly once
* Ensure device lost closure is called exactly once before being dropped.

This requires a change to the Rust callback signature, which is now Fn
instead of FnOnce. When the Rust callback or the C closure are dropped,
they will panic if they haven't been called. `device_drop` is changed
to call the closure with a message of "Device dropped." A test is added.
2023-12-19 23:16:10 +00:00
Nicolas Silva aade481bdf Remove some locks in BindGroup (#4894)
* Remove some locks in BindGroup

These are only written to clear the vectors when triaging bindgroups for destruction, which is not necessary. We can let the reference counts drop when the bind group is dropped.

* Make the mem_leak test pass again
2023-12-19 10:29:32 +01:00
Nicolas Silva a1b183f736 Avoid allocating memory every time we might log a label (#4893)
We allocate a String every time we want to get a label for logging. The string is also allocated when logging is disabled. Either way, the allocation is unnecessary. This commit replaces the String with a dyn Debug reference which does not need any allocation.
2023-12-18 13:08:18 +01:00
Nicolas Silva 192a2fe16c Simplify ResourceMaps (#4880)
* Remove the abstractions in resource maps

ResourceMaps had a rather convoluted system for erasing types that isn't needed anywhere except in one place in the triage_resource function. Everywhere else we are always dealing with specific types so using a member of the resource maps is simpler than going through an abstraction. More importantly there was a constraint that all contents of the resource maps implement the Resource trait which got in the way of some of the ongoing buffer snatching changes.

This commit simplifies this by removing the abstraction. Each resource type has its hash map directly in ResourceMaps and it is easier to have different requirements and behaviors depending on the type of the each resource.
2023-12-18 08:27:58 +00:00