diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 6c287b927c..b4a6f86548 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -159,11 +159,44 @@ jobs:
           # check with no features
           cargo clippy --target ${{ matrix.target }} -p wgpu -p wgpu-core -p wgpu-info -p player --no-default-features
 
-          # check with all features
-          cargo clippy --target ${{ matrix.target }} -p wgpu -p wgpu-core -p wgpu-info -p player --tests --all-features
+          # Check with all features.
+          # (But watch out for backend-selection features in wgpu-core; some of
+          # those only build on the right platforms.)
+          cargo clippy --target ${{ matrix.target }} -p wgpu -p wgpu-info -p player --tests --all-features
+          cargo clippy --target ${{ matrix.target }} -p wgpu-core --tests \
+              --features="portable_features"
 
           # build docs
-          cargo doc --target ${{ matrix.target }} -p wgpu -p wgpu-core -p wgpu-info -p player --all-features --no-deps
+          # (Watch out for backend-selection features in wgpu-core; some of
+          # those only build on the right platforms.)
+          cargo doc --target ${{ matrix.target }} -p wgpu -p wgpu-info -p player --all-features --no-deps
+          cargo doc --target ${{ matrix.target }} -p wgpu-core --no-deps --features="portable_features"
+
+  wasm-test:
+    name: Test WebAssembly
+    runs-on: ubuntu-latest
+    steps:
+      - name: checkout repo
+        uses: actions/checkout@v3
+
+      - name: install rust stable
+        uses: actions-rs/toolchain@v1
+        with:
+          toolchain: stable
+          profile: minimal
+          override: true
+
+      - name: install wasm-pack  # install from fork until this is merged: https://github.com/rustwasm/wasm-pack/pull/1185
+        run: |
+          # replace with "install wasm-pack action", which doesn't work for this project because of https://github.com/rustwasm/wasm-pack/issues/1180
+          #      - name: install wasm-pack
+          #        uses: jetli/wasm-pack-action@v0.4.0
+          cargo install --git https://github.com/haraldreingruber/wasm-pack wasm-pack
+
+      - name: execute tests
+        run: |
+          cd wgpu
+          wasm-pack test --headless --chrome --features webgl
 
   gpu-test:
     strategy:
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 96f651e8e1..60150d709a 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -42,6 +42,42 @@ Bottom level categories:
 
 ### Major Changes
 
+#### Backend selection by features
+
+Whereas `wgpu-core` used to automatically select backends to enable
+based on the target OS and architecture, it now has separate features
+to enable each backend:
+
+- "metal", for the Metal API on macOS and iOS
+- "vulkan", for the Vulkan API (Linux, some Android, and occasionally Windows)
+- "dx12", for Microsoft's Direct3D 12 API
+- "gles", OpenGL ES, available on many systems
+- "dx11", for Microsoft's Direct3D 11 API
+
+None are enabled by default, but the `wgpu` crate automatically
+selects these features based on the target operating system and
+architecture, using the same rules that `wgpu-core` used to, so users
+of `wgpu` should be unaffected by this change. However, other crates
+using `wgpu-core` directly will need to copy `wgpu`'s logic or write
+their own. See the `[target]` section of `wgpu/Cargo.toml` for
+details.
+
+Similarly, `wgpu-core` now has `emscripten` and `renderdoc` features
+that `wgpu` enables on appropriate platforms.
+
+In previous releases, the `wgpu-core` crate decided which backends to
+support. However, this left `wgpu-core`'s users with no way to
+override those choices. (Firefox doesn't want the GLES back end, for
+example.) There doesn't seem to be any way to have a crate select
+backends based on target OS and architecture that users of that crate
+can still override. Default features can't be selected based on the
+target, for example. That implies that we should do the selection as
+late in the dependency DAG as feasible. Having `wgpu` (and
+`wgpu-core`'s other dependents) choose backends seems like the best
+option.
+
+By @jimblandy in [#3254](https://github.com/gfx-rs/wgpu/pull/3254).
+
 #### Surface Capabilities API
 
 The various surface capability functions were combined into a single call that gives you all the capabilities.
@@ -63,6 +99,10 @@ Additionally `Surface::get_default_config` now returns an Option and returns Non
 + let config = surface.get_default_config(&adapter).expect("Surface unsupported by adapter");
 ```
 
+#### Fallible surface creation
+
+`Instance::create_surface()` now returns `Result<Surface, CreateSurfaceError>` instead of `Surface`. This allows an error to be returned instead of panicking if the given window is a HTML canvas and obtaining a WebGPU or WebGL 2 context fails. (No other platforms currently report any errors through this path.) By @kpreid in [#3052](https://github.com/gfx-rs/wgpu/pull/3052/)
+
 ### Changes
 
 #### General
@@ -73,6 +113,7 @@ Additionally `Surface::get_default_config` now returns an Option and returns Non
 - New downlevel feature `UNRESTRICTED_INDEX_BUFFER` to indicate support for using `INDEX` together with other non-copy/map usages (unsupported on WebGL). By @Wumpf in [#3157](https://github.com/gfx-rs/wgpu/pull/3157)
 - Combine `Surface::get_supported_formats`, `Surface::get_supported_present_modes`, and `Surface::get_supported_alpha_modes` into `Surface::get_capabilities` and `SurfaceCapabilities`. By @cwfitzgerald in [#3157](https://github.com/gfx-rs/wgpu/pull/3157)
 - Make `Surface::get_default_config` return an Option to prevent panics. By @cwfitzgerald in [#3157](https://github.com/gfx-rs/wgpu/pull/3157)
+- Lower the `max_buffer_size` limit value for compatibility with Apple2 and WebGPU compliance. By @jinleili in [#3255](https://github.com/gfx-rs/wgpu/pull/3255)
 
 #### WebGPU
 
@@ -92,11 +133,13 @@ Additionally `Surface::get_default_config` now returns an Option and returns Non
 - Add `get_default_config` for `Surface` to simplify user creation of `SurfaceConfiguration`. By @jinleili in [#3034](https://github.com/gfx-rs/wgpu/pull/3034)
 - Native adapters can now use MSAA x2 and x8 if it's supported , previously only x1 and x4 were supported . By @39ali in [3140](https://github.com/gfx-rs/wgpu/pull/3140)
 - Implemented correleation between user timestamps and platform specific presentation timestamps via [`Adapter::correlate_presentation_timestamp`]. By @cwfitzgerald in [#3240](https://github.com/gfx-rs/wgpu/pull/3240)
+- Added support for `Features::SHADER_PRIMITIVE_INDEX` on all backends. By @cwfitzgerald in [#3272](https://github.com/gfx-rs/wgpu/pull/3272)
 
 #### GLES
 
 - Surfaces support now `TextureFormat::Rgba8Unorm` and (non-web only) `TextureFormat::Bgra8Unorm`. By @Wumpf in [#3070](https://github.com/gfx-rs/wgpu/pull/3070)
 - Support alpha to coverage. By @Wumpf in [#3156](https://github.com/gfx-rs/wgpu/pull/3156)
+- Support filtering f32 textures. By @expenses in [#3261](https://github.com/gfx-rs/wgpu/pull/3261)
 
 #### WebGPU
 
@@ -115,6 +158,8 @@ Additionally `Surface::get_default_config` now returns an Option and returns Non
 - Fix an integer overflow in `queue_write_texture` by @nical in (#3146)[https://github.com/gfx-rs/wgpu/pull/3146]
 - Make `RenderPassCompatibilityError` and `CreateShaderModuleError` not so huge. By @jimblandy in (#3226)[https://github.com/gfx-rs/wgpu/pull/3226]
 - Check for invalid bitflag bits in wgpu-core and allow them to be captured/replayed by @nical in (#3229)[https://github.com/gfx-rs/wgpu/pull/3229]
+- Evaluate `gfx_select!`'s `#[cfg]` conditions at the right time. By @jimblandy in [#3253](https://github.com/gfx-rs/wgpu/pull/3253)
+- Improve error messages when binding bind group with dynamic offsets. By @cwfitzgerald in [#3294](https://github.com/gfx-rs/wgpu/pull/3294)
 
 #### WebGPU
 
@@ -136,6 +181,10 @@ Additionally `Surface::get_default_config` now returns an Option and returns Non
 
 - Let the wgpu examples `framework.rs` compile again under Emscripten. By @jimblandy in [#3246](https://github.com/gfx-rs/wgpu/pull/3246)
 
+#### Vulkan
+
+- Update ash to 0.37.1+1.3.235 to fix CI breaking by changing a call to the deprecated `debug_utils_set_object_name()` function to `set_debug_utils_object_name()` by @elabajaba in [#3273](https://github.com/gfx-rs/wgpu/pull/3273)
+
 ### Examples
 
 - Log adapter info in hello example on wasm target by @JolifantoBambla in [#2858](https://github.com/gfx-rs/wgpu/pull/2858)
@@ -143,8 +192,9 @@ Additionally `Surface::get_default_config` now returns an Option and returns Non
 ### Testing/Internal
 
 - Update the `minimum supported rust version` to 1.64
-- Use cargo 1.64 workspace inheritance feature. By @jinleili in [#3107](https://github.com/gfx-rs/wgpu/pull/3107)
 - Move `ResourceMetadata` into its own module. By @jimblandy in [#3213](https://github.com/gfx-rs/wgpu/pull/3213)
+- Add WebAssembly testing infrastructure. By @haraldreingruber in [#3238](https://github.com/gfx-rs/wgpu/pull/3238)
+- Error message when you forget to use cargo-nextest. By @cwfitzgerald in [#3293](https://github.com/gfx-rs/wgpu/pull/3293)
 
 #### Vulkan
 
diff --git a/Cargo.lock b/Cargo.lock
index e68dd9d134..14720620d4 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -66,9 +66,9 @@ dependencies = [
 
 [[package]]
 name = "ash"
-version = "0.37.0+1.3.209"
+version = "0.37.1+1.3.235"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "006ca68e0f2b03f22d6fa9f2860f85aed430d257fec20f8879b2145e7c7ae1a6"
+checksum = "911015c962d56e2e4052f40182ca5462ba60a3d2ff04e827c365a0ab3d65726d"
 dependencies = [
  "libloading",
 ]
@@ -131,12 +131,6 @@ dependencies = [
  "safemem",
 ]
 
-[[package]]
-name = "base64"
-version = "0.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b41b7ea54a0c9d92199de89e20e58d49f02f8e699814ef3fdf266f6f748d15c7"
-
 [[package]]
 name = "base64"
 version = "0.13.0"
@@ -145,9 +139,9 @@ checksum = "904dfeac50f3cdaba28fc6f57fdcddb75f49ed61346676a78c4ffe55877802fd"
 
 [[package]]
 name = "base64-simd"
-version = "0.6.2"
+version = "0.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "278c7ba87265587c4823cf1b2fdf57834151540b2e509574adb03627f8c7f22d"
+checksum = "781dd20c3aff0bd194fe7d2a977dd92f21c173891f3a03b677359e5fa457e5d5"
 dependencies = [
  "simd-abstraction",
 ]
@@ -259,12 +253,6 @@ version = "1.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
-[[package]]
-name = "cfg_aliases"
-version = "0.1.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fd16c4719339c4530435d38e511904438d07cce7950afa3718a84ac36c10e89e"
-
 [[package]]
 name = "cgl"
 version = "0.3.2"
@@ -523,20 +511,21 @@ dependencies = [
 
 [[package]]
 name = "deno_console"
-version = "0.69.0"
+version = "0.80.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "02cd684bd0097101aa8b411dac1e955ded1bc9c0945a453b607af929ca33b380"
+checksum = "68a13f7497eaec2c5b7f766d7cecbdd8547611dabd6ce8f615c507d275887049"
 dependencies = [
  "deno_core",
 ]
 
 [[package]]
 name = "deno_core"
-version = "0.151.0"
+version = "0.162.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "390bdf983f9f20d403b09894ce4d9baeb980fa5faa33bf859c47ffa729abd157"
+checksum = "3f2aa18d22faadc44c9c287b4d6f5a0f2aeea75a120d2a7d400a98e4b9255584"
 dependencies = [
  "anyhow",
+ "bytes",
  "deno_ops",
  "futures",
  "indexmap",
@@ -548,6 +537,7 @@ dependencies = [
  "serde",
  "serde_json",
  "serde_v8",
+ "smallvec",
  "sourcemap",
  "url",
  "v8",
@@ -555,11 +545,12 @@ dependencies = [
 
 [[package]]
 name = "deno_ops"
-version = "0.29.0"
+version = "0.40.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "73e5775de06dc4589c43fa5d81d962f8db9640ccf214291625300e6bf6f3e806"
+checksum = "83ddd1b980dcf7c7b4ad586338704c78c6423608f4b8fd622d72bfa76006333f"
 dependencies = [
  "once_cell",
+ "pmutil",
  "proc-macro-crate",
  "proc-macro2",
  "quote",
@@ -569,9 +560,9 @@ dependencies = [
 
 [[package]]
 name = "deno_url"
-version = "0.69.0"
+version = "0.80.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fcc823c3e01d24a2a55e8e9b23fcdcfdf376c039a24b3e3571b9b17630f05186"
+checksum = "4d0cf9b99857820d594cc816b84507ba062f276427aab815c380cdeec6554021"
 dependencies = [
  "deno_core",
  "serde",
@@ -581,9 +572,9 @@ dependencies = [
 
 [[package]]
 name = "deno_web"
-version = "0.100.0"
+version = "0.111.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "204620cb1ce3ec06b74d51fb3370c4fcb25d34d42b0557b6e97bbe84ea64e770"
+checksum = "54f0a00df81eee23798d2a516bdc54d7541e3008310b26e7f79533df7d22bd7b"
 dependencies = [
  "async-trait",
  "base64-simd",
@@ -597,7 +588,7 @@ dependencies = [
 
 [[package]]
 name = "deno_webgpu"
-version = "0.63.0"
+version = "0.81.0"
 dependencies = [
  "deno_core",
  "serde",
@@ -608,9 +599,9 @@ dependencies = [
 
 [[package]]
 name = "deno_webidl"
-version = "0.69.0"
+version = "0.80.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a3bbfcb5416924c4b7ed50514d6577d8a87a61772a043daabe00d81734f5cb07"
+checksum = "92a564d515fb807b1f2712717b4de5082cc59f3f86a7e233c04231a760c001e5"
 dependencies = [
  "deno_core",
 ]
@@ -1257,9 +1248,9 @@ checksum = "884e2677b40cc8c339eaefcb701c32ef1fd2493d71118dc0ca4b6a736c93bd67"
 
 [[package]]
 name = "libc"
-version = "0.2.134"
+version = "0.2.126"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "329c933548736bc49fd575ee68c89e8be4d260064184389a5b77517cddd99ffb"
+checksum = "349d5a591cd28b49e1d1037471617a32ddcda5731b99419008085f72d5a53836"
 
 [[package]]
 name = "libloading"
@@ -1558,9 +1549,9 @@ dependencies = [
 
 [[package]]
 name = "once_cell"
-version = "1.15.0"
+version = "1.14.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e82dad04139b71a90c080c8463fe0dc7902db5192d939bd0950f074d014339e1"
+checksum = "2f7254b99e31cad77da24b08ebf628882739a608578bb1bcdfc1f9c21260d7c0"
 
 [[package]]
 name = "osmesa-sys"
@@ -1571,6 +1562,12 @@ dependencies = [
  "shared_library",
 ]
 
+[[package]]
+name = "outref"
+version = "0.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f222829ae9293e33a9f5e9f440c6760a3d450a64affe1846486b140db81c1f4"
+
 [[package]]
 name = "parking"
 version = "2.0.0"
@@ -1699,6 +1696,17 @@ dependencies = [
  "winit",
 ]
 
+[[package]]
+name = "pmutil"
+version = "0.5.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3894e5d549cccbe44afecf72922f277f603cd4bb0219c8342631ef18fffbe004"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
 [[package]]
 name = "png"
 version = "0.17.6"
@@ -1998,9 +2006,9 @@ checksum = "388a1df253eca08550bef6c72392cfe7c30914bf41df5269b68cbd6ff8f570a3"
 
 [[package]]
 name = "serde"
-version = "1.0.145"
+version = "1.0.144"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "728eb6351430bccb993660dfffc5a72f91ccc1295abaa8ce19b27ebe4f75568b"
+checksum = "0f747710de3dcd43b88c9168773254e809d8ddbdf9653b84e2554ab219f17860"
 dependencies = [
  "serde_derive",
 ]
@@ -2016,9 +2024,9 @@ dependencies = [
 
 [[package]]
 name = "serde_derive"
-version = "1.0.145"
+version = "1.0.144"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "81fa1584d3d1bcacd84c277a0dfe21f5b0f6accf4a23d04d4c6d61f1af522b4c"
+checksum = "94ed3a816fb1d101812f83e789f888322c34e291f894f19590dc310963e87a00"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -2050,9 +2058,9 @@ dependencies = [
 
 [[package]]
 name = "serde_v8"
-version = "0.62.0"
+version = "0.73.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b1f30310f753d4b1347acdd669a30b8e6208029cfbb28d3d91012b19333eeff1"
+checksum = "17891f7f7e138e2d25fafd19b5f7a95c2cf1e72be0e6804343c63aa6e90b0287"
 dependencies = [
  "bytes",
  "derive_more",
@@ -2104,9 +2112,12 @@ dependencies = [
 
 [[package]]
 name = "simd-abstraction"
-version = "0.6.2"
+version = "0.7.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c2880f3f7b392823ee65bbcc681961cd8e698c6a30e91ab9b4eef1f9c6c226d8"
+checksum = "9cadb29c57caadc51ff8346233b5cec1d240b68ce55cf1afc764818791876987"
+dependencies = [
+ "outref",
+]
 
 [[package]]
 name = "slab"
@@ -2163,11 +2174,11 @@ dependencies = [
 
 [[package]]
 name = "sourcemap"
-version = "6.0.1"
+version = "6.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6e031f2463ecbdd5f34c950f89f5c1e1032f22c0f8e3dc4bdb2e8b6658cf61eb"
+checksum = "c46fdc1838ff49cf692226f5c2b0f5b7538f556863d0eca602984714667ac6e7"
 dependencies = [
- "base64 0.11.0",
+ "base64 0.13.0",
  "if_chain",
  "lazy_static",
  "regex",
@@ -2295,9 +2306,9 @@ checksum = "cda74da7e1a664f795bb1f8a87ec406fb89a02522cf6e50620d016add6dbbf5c"
 
 [[package]]
 name = "tokio"
-version = "1.21.2"
+version = "1.21.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a9e03c497dc955702ba729190dc4aac6f2a0ce97f913e5b1b5912fc5039d9099"
+checksum = "0020c875007ad96677dcc890298f4b942882c5d4eb7cc8f439fc3bf813dc9c95"
 dependencies = [
  "autocfg",
  "bytes",
@@ -2305,6 +2316,7 @@ dependencies = [
  "memchr",
  "mio",
  "num_cpus",
+ "once_cell",
  "parking_lot 0.12.1",
  "pin-project-lite",
  "signal-hook-registry",
@@ -2440,9 +2452,9 @@ dependencies = [
 
 [[package]]
 name = "uuid"
-version = "1.2.1"
+version = "1.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "feb41e78f93363bb2df8b0e86a2ca30eed7806ea16ea0c790d757cf93f79be83"
+checksum = "dd6469f4314d5f1ffec476e05f17cc9a78bc7a27a6a857842170bdf8d6f98d2f"
 dependencies = [
  "getrandom 0.2.7",
  "serde",
@@ -2450,14 +2462,13 @@ dependencies = [
 
 [[package]]
 name = "v8"
-version = "0.49.0"
+version = "0.58.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5a1cbad73336d67babcbe5e3b03c907c8d2ff77fc6f997570af219bbd9fdb6ce"
+checksum = "8e9b88668afedf6ec9f8f6d30b446f622498da2ef0b3991a52e10f0ea8c6cc09"
 dependencies = [
  "bitflags",
  "fslock",
  "lazy_static",
- "libc",
  "which",
 ]
 
@@ -2627,6 +2638,30 @@ version = "0.2.83"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1c38c045535d93ec4f0b4defec448e4291638ee608530863b1e2ba115d4fff7f"
 
+[[package]]
+name = "wasm-bindgen-test"
+version = "0.3.33"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "09d2fff962180c3fadf677438054b1db62bee4aa32af26a45388af07d1287e1d"
+dependencies = [
+ "console_error_panic_hook",
+ "js-sys",
+ "scoped-tls",
+ "wasm-bindgen",
+ "wasm-bindgen-futures",
+ "wasm-bindgen-test-macro",
+]
+
+[[package]]
+name = "wasm-bindgen-test-macro"
+version = "0.3.33"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4683da3dfc016f704c9f82cf401520c4f1cb3ee440f7f52b3d6ac29506a49ca7"
+dependencies = [
+ "proc-macro2",
+ "quote",
+]
+
 [[package]]
 name = "wasm-bindgen-threads-xform"
 version = "0.2.83"
@@ -2804,6 +2839,7 @@ dependencies = [
  "static_assertions",
  "wasm-bindgen",
  "wasm-bindgen-futures",
+ "wasm-bindgen-test",
  "web-sys",
  "wgpu-core",
  "wgpu-hal",
@@ -2818,7 +2854,6 @@ dependencies = [
  "arrayvec 0.7.2",
  "bit-vec",
  "bitflags",
- "cfg_aliases",
  "codespan-reporting",
  "fxhash",
  "log",
diff --git a/Cargo.toml b/Cargo.toml
index 01ea9934e1..a2d537a0ff 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -89,7 +89,7 @@ objc = "0.2.5"
 core-graphics-types = "0.1"
 
 # Vulkan dependencies
-ash = "0.37"
+ash = "0.37.1"
 gpu-alloc = "0.5"
 gpu-descriptor = "0.2"
 android_system_properties = "0.1.1"
@@ -113,14 +113,15 @@ console_log = "0.2"
 js-sys = "0.3.60"
 wasm-bindgen = "0.2.83"
 wasm-bindgen-futures = "0.4.33"
+wasm-bindgen-test = "0.3"
 web-sys = "0.3.60"
 
 # deno dependencies
-deno_console = "0.69.0"
-deno_core = "0.151.0"
-deno_url = "0.69.0"
-deno_web = "0.100.0"
-deno_webidl = "0.69.0"
+deno_console = "0.80.0"
+deno_core = "0.162.0"
+deno_url = "0.80.0"
+deno_web = "0.111.0"
+deno_webidl = "0.80.0"
 deno_webgpu = { path = "./deno_webgpu" }
 tokio = "1.19.0"
 termcolor = "1.1.2"
diff --git a/deno_webgpu/Cargo.toml b/deno_webgpu/Cargo.toml
index 3091c9c088..4a47a9418e 100644
--- a/deno_webgpu/Cargo.toml
+++ b/deno_webgpu/Cargo.toml
@@ -2,7 +2,7 @@
 
 [package]
 name = "deno_webgpu"
-version = "0.63.0"
+version = "0.81.0"
 authors = ["the Deno authors"]
 edition.workspace = true
 license = "MIT"
@@ -14,5 +14,23 @@ description = "WebGPU implementation for Deno"
 deno_core.workspace = true
 serde = { workspace = true, features = ["derive"] }
 tokio = { workspace = true, features = ["full"] }
-wgpu-core = { workspace = true, features = ["trace", "replay", "serde", "strict_asserts", "wgsl"] }
 wgpu-types = { workspace = true, features = ["trace", "replay", "serde"] }
+
+[dependencies.wgpu-core]
+workspace = true
+features = ["trace", "replay", "serde", "strict_asserts", "wgsl", "gles"]
+
+# We want the wgpu-core Metal backend on macOS and iOS.
+[target.'cfg(any(target_os = "macos", target_os = "ios"))'.dependencies.wgpu-core]
+workspace = true
+features = ["metal"]
+
+# We want the wgpu-core Direct3D backends on Windows.
+[target.'cfg(windows)'.dependencies.wgpu-core]
+workspace = true
+features = ["dx11", "dx12"]
+
+# We want the wgpu-core Vulkan backend on Unix (but not Emscripten) and Windows.
+[target.'cfg(any(windows, all(unix, not(target_arch = "emscripten"))))'.dependencies.wgpu-core]
+workspace = true
+features = ["vulkan"]
diff --git a/deno_webgpu/src/01_webgpu.js b/deno_webgpu/src/01_webgpu.js
index c01911af52..373fe58a8b 100644
--- a/deno_webgpu/src/01_webgpu.js
+++ b/deno_webgpu/src/01_webgpu.js
@@ -10,6 +10,7 @@
 
 ((window) => {
   const core = window.Deno.core;
+  const ops = core.ops;
   const webidl = window.__bootstrap.webidl;
   const eventTarget = window.__bootstrap.eventTarget;
   const { DOMException } = window.__bootstrap.domException;
@@ -26,12 +27,12 @@
     ObjectDefineProperty,
     ObjectPrototypeIsPrototypeOf,
     Promise,
-    PromiseAll,
     PromisePrototypeCatch,
     PromisePrototypeThen,
     PromiseReject,
     PromiseResolve,
     SafeArrayIterator,
+    SafePromiseAll,
     Set,
     SetPrototypeEntries,
     SetPrototypeForEach,
@@ -946,8 +947,7 @@
         context: "Argument 1",
       });
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_buffer",
+      const { rid, err } = ops.op_webgpu_create_buffer(
         device.rid,
         descriptor.label,
         descriptor.size,
@@ -997,7 +997,7 @@
         context: "Argument 1",
       });
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync("op_webgpu_create_texture", {
+      const { rid, err } = ops.op_webgpu_create_texture({
         deviceRid: device.rid,
         ...descriptor,
         size: normalizeGPUExtent3D(descriptor.size),
@@ -1025,7 +1025,7 @@
         context: "Argument 1",
       });
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync("op_webgpu_create_sampler", {
+      const { rid, err } = ops.op_webgpu_create_texture({
         deviceRid: device.rid,
         ...descriptor,
       });
@@ -1065,8 +1065,7 @@
         }
       }
 
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_bind_group_layout",
+      const { rid, err } = ops.op_webgpu_create_bind_group_layout(
         device.rid,
         descriptor.label,
         descriptor.entries,
@@ -1108,8 +1107,7 @@
           return rid;
         },
       );
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_pipeline_layout",
+      const { rid, err } = ops.op_webgpu_create_pipeline_layout(
         device.rid,
         descriptor.label,
         bindGroupLayouts,
@@ -1203,8 +1201,7 @@
         }
       });
 
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_bind_group",
+      const { rid, err } = ops.op_webgpu_create_bind_group(
         device.rid,
         descriptor.label,
         layout,
@@ -1233,8 +1230,7 @@
         context: "Argument 1",
       });
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_shader_module",
+      const { rid, err } = ops.op_webgpu_create_shader_module(
         device.rid,
         descriptor.label,
         descriptor.code,
@@ -1283,8 +1279,7 @@
         selfContext: "this",
       });
 
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_compute_pipeline",
+      const { rid, err } = ops.op_webgpu_create_compute_pipeline(
         device.rid,
         descriptor.label,
         layout,
@@ -1355,7 +1350,7 @@
         };
       }
 
-      const { rid, err } = core.opSync("op_webgpu_create_render_pipeline", {
+      const { rid, err } = ops.op_webgpu_create_render_pipeline({
         deviceRid: device.rid,
         label: descriptor.label,
         layout,
@@ -1402,8 +1397,7 @@
         context: "Argument 1",
       });
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_command_encoder",
+      const { rid, err } = ops.op_webgpu_create_command_encoder(
         device.rid,
         descriptor.label,
       );
@@ -1435,13 +1429,10 @@
         },
       );
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync(
-        "op_webgpu_create_render_bundle_encoder",
-        {
-          deviceRid: device.rid,
-          ...descriptor,
-        },
-      );
+      const { rid, err } = ops.op_webgpu_create_render_bundle_encoder({
+        deviceRid: device.rid,
+        ...descriptor,
+      });
       device.pushError(err);
 
       const renderBundleEncoder = createGPURenderBundleEncoder(
@@ -1469,7 +1460,7 @@
         },
       );
       const device = assertDevice(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync("op_webgpu_create_query_set", {
+      const { rid, err } = ops.op_webgpu_create_query_set({
         deviceRid: device.rid,
         ...descriptor,
       });
@@ -1530,7 +1521,7 @@
           "OperationError",
         );
       }
-      const operations = PromiseAll(scope.operations);
+      const operations = SafePromiseAll(scope.operations);
       return PromisePrototypeThen(
         operations,
         () => PromiseResolve(null),
@@ -1600,11 +1591,7 @@
           return rid;
         },
       );
-      const { err } = core.opSync(
-        "op_webgpu_queue_submit",
-        device.rid,
-        commandBufferRids,
-      );
+      const { err } = ops.op_webgpu_queue_submit(device.rid, commandBufferRids);
       for (const commandBuffer of commandBuffers) {
         commandBuffer[_rid] = undefined;
       }
@@ -1659,8 +1646,7 @@
         selfContext: "this",
         resourceContext: "Argument 1",
       });
-      const { err } = core.opSync(
-        "op_webgpu_write_buffer",
+      const { err } = ops.op_webgpu_write_buffer(
         device.rid,
         bufferRid,
         bufferOffset,
@@ -1707,8 +1693,7 @@
         selfContext: "this",
         resourceContext: "texture",
       });
-      const { err } = core.opSync(
-        "op_webgpu_write_texture",
+      const { err } = ops.op_webgpu_write_texture(
         device.rid,
         {
           texture: textureRid,
@@ -1979,8 +1964,7 @@
       }
 
       const buffer = new ArrayBuffer(rangeSize);
-      const { rid } = core.opSync(
-        "op_webgpu_buffer_get_mapped_range",
+      const { rid } = ops.op_webgpu_buffer_get_mapped_range(
         bufferRid,
         offset,
         size,
@@ -2034,8 +2018,7 @@
           throw new DOMException(`${prefix}: invalid state.`, "OperationError");
         }
         for (const [buffer, mappedRid] of mappedRanges) {
-          const { err } = core.opSync(
-            "op_webgpu_buffer_unmap",
+          const { err } = ops.op_webgpu_buffer_unmap(
             bufferRid,
             mappedRid,
             ...new SafeArrayIterator(write ? [new Uint8Array(buffer)] : []),
@@ -2198,7 +2181,7 @@
       });
       const device = assertDevice(this, { prefix, context: "this" });
       const textureRid = assertResource(this, { prefix, context: "this" });
-      const { rid, err } = core.opSync("op_webgpu_create_texture_view", {
+      const { rid, err } = ops.op_webgpu_create_texture_view({
         textureRid,
         ...descriptor,
       });
@@ -2622,11 +2605,11 @@
         prefix,
         context: "this",
       });
-      const { rid, label, err } = core.opSync(
-        "op_webgpu_compute_pipeline_get_bind_group_layout",
-        computePipelineRid,
-        index,
-      );
+      const { rid, label, err } =
+        ops.op_webgpu_compute_pipeline_get_bind_group_layout(
+          computePipelineRid,
+          index,
+        );
       device.pushError(err);
 
       const bindGroupLayout = createGPUBindGroupLayout(
@@ -2699,11 +2682,11 @@
         prefix,
         context: "this",
       });
-      const { rid, label, err } = core.opSync(
-        "op_webgpu_render_pipeline_get_bind_group_layout",
-        renderPipelineRid,
-        index,
-      );
+      const { rid, label, err } =
+        ops.op_webgpu_render_pipeline_get_bind_group_layout(
+          renderPipelineRid,
+          index,
+        );
       device.pushError(err);
 
       const bindGroupLayout = createGPUBindGroupLayout(
@@ -2893,8 +2876,7 @@
         },
       );
 
-      const { rid } = core.opSync(
-        "op_webgpu_command_encoder_begin_render_pass",
+      const { rid } = ops.op_webgpu_command_encoder_begin_render_pass(
         commandEncoderRid,
         descriptor.label,
         colorAttachments,
@@ -2928,8 +2910,7 @@
         context: "this",
       });
 
-      const { rid } = core.opSync(
-        "op_webgpu_command_encoder_begin_compute_pass",
+      const { rid } = ops.op_webgpu_command_encoder_begin_compute_pass(
         commandEncoderRid,
         descriptor.label,
       );
@@ -3005,8 +2986,7 @@
         selfContext: "this",
       });
 
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_copy_buffer_to_buffer",
+      const { err } = ops.op_webgpu_command_encoder_copy_buffer_to_buffer(
         commandEncoderRid,
         sourceRid,
         sourceOffset,
@@ -3063,8 +3043,7 @@
         selfContext: "this",
       });
 
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_copy_buffer_to_texture",
+      const { err } = ops.op_webgpu_command_encoder_copy_buffer_to_texture(
         commandEncoderRid,
         {
           ...source,
@@ -3128,8 +3107,7 @@
         resourceContext: "buffer in Argument 2",
         selfContext: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_copy_texture_to_buffer",
+      const { err } = ops.op_webgpu_command_encoder_copy_texture_to_buffer(
         commandEncoderRid,
         {
           texture: sourceTextureRid,
@@ -3193,8 +3171,7 @@
         resourceContext: "texture in Argument 2",
         selfContext: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_copy_texture_to_texture",
+      const { err } = ops.op_webgpu_command_encoder_copy_texture_to_texture(
         commandEncoderRid,
         {
           texture: sourceTextureRid,
@@ -3247,8 +3224,7 @@
         prefix,
         context: "Argument 1",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_clear_buffer",
+      const { err } = ops.op_webgpu_command_encoder_clear_buffer(
         commandEncoderRid,
         bufferRid,
         offset,
@@ -3274,8 +3250,7 @@
         prefix,
         context: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_push_debug_group",
+      const { err } = ops.op_webgpu_command_encoder_push_debug_group(
         commandEncoderRid,
         groupLabel,
       );
@@ -3290,8 +3265,7 @@
         prefix,
         context: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_pop_debug_group",
+      const { err } = ops.op_webgpu_command_encoder_pop_debug_group(
         commandEncoderRid,
       );
       device.pushError(err);
@@ -3314,8 +3288,7 @@
         prefix,
         context: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_insert_debug_marker",
+      const { err } = ops.op_webgpu_command_encoder_insert_debug_marker(
         commandEncoderRid,
         markerLabel,
       );
@@ -3353,8 +3326,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_write_timestamp",
+      const { err } = ops.op_webgpu_command_encoder_write_timestamp(
         commandEncoderRid,
         querySetRid,
         queryIndex,
@@ -3423,8 +3395,7 @@
         resourceContext: "Argument 3",
         selfContext: "this",
       });
-      const { err } = core.opSync(
-        "op_webgpu_command_encoder_resolve_query_set",
+      const { err } = ops.op_webgpu_command_encoder_resolve_query_set(
         commandEncoderRid,
         querySetRid,
         firstQuery,
@@ -3451,8 +3422,7 @@
         prefix,
         context: "this",
       });
-      const { rid, err } = core.opSync(
-        "op_webgpu_command_encoder_finish",
+      const { rid, err } = ops.op_webgpu_command_encoder_finish(
         commandEncoderRid,
         descriptor.label,
       );
@@ -3551,7 +3521,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync("op_webgpu_render_pass_set_viewport", {
+      ops.op_webgpu_render_pass_set_viewport({
         renderPassRid,
         x,
         y,
@@ -3598,8 +3568,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_set_scissor_rect",
+      ops.op_webgpu_render_pass_set_scissor_rect(
         renderPassRid,
         x,
         y,
@@ -3629,8 +3598,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_set_blend_constant",
+      ops.op_webgpu_render_pass_set_blend_constant(
         renderPassRid,
         normalizeGPUColor(color),
       );
@@ -3657,8 +3625,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_set_stencil_reference",
+      ops.op_webgpu_render_pass_set_stencil_reference(
         renderPassRid,
         reference,
       );
@@ -3707,8 +3674,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_begin_pipeline_statistics_query",
+      ops.op_webgpu_render_pass_begin_pipeline_statistics_query(
         renderPassRid,
         querySetRid,
         queryIndex,
@@ -3728,10 +3694,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_end_pipeline_statistics_query",
-        renderPassRid,
-      );
+      ops.op_webgpu_render_pass_end_pipeline_statistics_query(renderPassRid);
     }
 
     /**
@@ -3769,8 +3732,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_write_timestamp",
+      ops.op_webgpu_render_pass_write_timestamp(
         renderPassRid,
         querySetRid,
         queryIndex,
@@ -3808,11 +3770,7 @@
         });
         return rid;
       });
-      core.opSync(
-        "op_webgpu_render_pass_execute_bundles",
-        renderPassRid,
-        bundleRids,
-      );
+      ops.op_webgpu_render_pass_execute_bundles(renderPassRid, bundleRids);
     }
 
     end() {
@@ -3827,8 +3785,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      const { err } = core.opSync(
-        "op_webgpu_render_pass_end",
+      const { err } = ops.op_webgpu_render_pass_end(
         commandEncoderRid,
         renderPassRid,
       );
@@ -3875,8 +3832,7 @@
         dynamicOffsetsDataStart = 0;
         dynamicOffsetsDataLength = dynamicOffsetsData.length;
       }
-      core.opSync(
-        "op_webgpu_render_pass_set_bind_group",
+      ops.op_webgpu_render_pass_set_bind_group(
         renderPassRid,
         index,
         bindGroupRid,
@@ -3907,11 +3863,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_push_debug_group",
-        renderPassRid,
-        groupLabel,
-      );
+      ops.op_webgpu_render_pass_push_debug_group(renderPassRid, groupLabel);
     }
 
     popDebugGroup() {
@@ -3927,7 +3879,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync("op_webgpu_render_pass_pop_debug_group", renderPassRid);
+      ops.op_webgpu_render_pass_pop_debug_group(renderPassRid);
     }
 
     /**
@@ -3951,11 +3903,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_insert_debug_marker",
-        renderPassRid,
-        markerLabel,
-      );
+      ops.op_webgpu_render_pass_insert_debug_marker(renderPassRid, markerLabel);
     }
 
     /**
@@ -3988,11 +3936,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_set_pipeline",
-        renderPassRid,
-        pipelineRid,
-      );
+      ops.op_webgpu_render_pass_set_pipeline(renderPassRid, pipelineRid);
     }
 
     /**
@@ -4042,8 +3986,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_set_index_buffer",
+      ops.op_webgpu_render_pass_set_index_buffer(
         renderPassRid,
         bufferRid,
         indexFormat,
@@ -4099,8 +4042,7 @@
         resourceContext: "Argument 2",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_set_vertex_buffer",
+      ops.op_webgpu_render_pass_set_vertex_buffer(
         renderPassRid,
         slot,
         bufferRid,
@@ -4144,8 +4086,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_draw",
+      ops.op_webgpu_render_pass_draw(
         renderPassRid,
         vertexCount,
         instanceCount,
@@ -4201,8 +4142,7 @@
         context: "encoder referenced by this",
       });
       const renderPassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_render_pass_draw_indexed",
+      ops.op_webgpu_render_pass_draw_indexed(
         renderPassRid,
         indexCount,
         instanceCount,
@@ -4247,8 +4187,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_draw_indirect",
+      ops.op_webgpu_render_pass_draw_indirect(
         renderPassRid,
         indirectBufferRid,
         indirectOffset,
@@ -4290,8 +4229,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_pass_draw_indexed_indirect",
+      ops.op_webgpu_render_pass_draw_indexed_indirect(
         renderPassRid,
         indirectBufferRid,
         indirectOffset,
@@ -4374,11 +4312,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_compute_pass_set_pipeline",
-        computePassRid,
-        pipelineRid,
-      );
+      ops.op_webgpu_compute_pass_set_pipeline(computePassRid, pipelineRid);
     }
 
     /**
@@ -4416,8 +4350,7 @@
         context: "encoder referenced by this",
       });
       const computePassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_compute_pass_dispatch_workgroups",
+      ops.op_webgpu_compute_pass_dispatch_workgroups(
         computePassRid,
         workgroupCountX,
         workgroupCountY,
@@ -4460,8 +4393,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_compute_pass_dispatch_workgroups_indirect",
+      ops.op_webgpu_compute_pass_dispatch_workgroups_indirect(
         computePassRid,
         indirectBufferRid,
         indirectOffset,
@@ -4503,8 +4435,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_compute_pass_begin_pipeline_statistics_query",
+      ops.op_webgpu_compute_pass_begin_pipeline_statistics_query(
         computePassRid,
         querySetRid,
         queryIndex,
@@ -4524,10 +4455,7 @@
         context: "encoder referenced by this",
       });
       const computePassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_compute_pass_end_pipeline_statistics_query",
-        computePassRid,
-      );
+      ops.op_webgpu_compute_pass_end_pipeline_statistics_query(computePassRid);
     }
 
     /**
@@ -4565,8 +4493,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_compute_pass_write_timestamp",
+      ops.op_webgpu_compute_pass_write_timestamp(
         computePassRid,
         querySetRid,
         queryIndex,
@@ -4585,8 +4512,7 @@
         context: "encoder referenced by this",
       });
       const computePassRid = assertResource(this, { prefix, context: "this" });
-      const { err } = core.opSync(
-        "op_webgpu_compute_pass_end",
+      const { err } = ops.op_webgpu_compute_pass_end(
         commandEncoderRid,
         computePassRid,
       );
@@ -4633,8 +4559,7 @@
         dynamicOffsetsDataStart = 0;
         dynamicOffsetsDataLength = dynamicOffsetsData.length;
       }
-      core.opSync(
-        "op_webgpu_compute_pass_set_bind_group",
+      ops.op_webgpu_compute_pass_set_bind_group(
         computePassRid,
         index,
         bindGroupRid,
@@ -4665,11 +4590,7 @@
         context: "encoder referenced by this",
       });
       const computePassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_compute_pass_push_debug_group",
-        computePassRid,
-        groupLabel,
-      );
+      ops.op_webgpu_compute_pass_push_debug_group(computePassRid, groupLabel);
     }
 
     popDebugGroup() {
@@ -4685,7 +4606,7 @@
         context: "encoder referenced by this",
       });
       const computePassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync("op_webgpu_compute_pass_pop_debug_group", computePassRid);
+      ops.op_webgpu_compute_pass_pop_debug_group(computePassRid);
     }
 
     /**
@@ -4709,8 +4630,7 @@
         context: "encoder referenced by this",
       });
       const computePassRid = assertResource(this, { prefix, context: "this" });
-      core.opSync(
-        "op_webgpu_compute_pass_insert_debug_marker",
+      ops.op_webgpu_compute_pass_insert_debug_marker(
         computePassRid,
         markerLabel,
       );
@@ -4820,8 +4740,7 @@
         prefix,
         context: "this",
       });
-      const { rid, err } = core.opSync(
-        "op_webgpu_render_bundle_encoder_finish",
+      const { rid, err } = ops.op_webgpu_render_bundle_encoder_finish(
         renderBundleEncoderRid,
         descriptor.label,
       );
@@ -4872,8 +4791,7 @@
         dynamicOffsetsDataStart = 0;
         dynamicOffsetsDataLength = dynamicOffsetsData.length;
       }
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_set_bind_group",
+      ops.op_webgpu_render_bundle_encoder_set_bind_group(
         renderBundleEncoderRid,
         index,
         bindGroupRid,
@@ -4900,8 +4818,7 @@
         prefix,
         context: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_push_debug_group",
+      ops.op_webgpu_render_bundle_encoder_push_debug_group(
         renderBundleEncoderRid,
         groupLabel,
       );
@@ -4916,8 +4833,7 @@
         prefix,
         context: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_pop_debug_group",
+      ops.op_webgpu_render_bundle_encoder_pop_debug_group(
         renderBundleEncoderRid,
       );
     }
@@ -4939,8 +4855,7 @@
         prefix,
         context: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_insert_debug_marker",
+      ops.op_webgpu_render_bundle_encoder_insert_debug_marker(
         renderBundleEncoderRid,
         markerLabel,
       );
@@ -4972,8 +4887,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_set_pipeline",
+      ops.op_webgpu_render_bundle_encoder_set_pipeline(
         renderBundleEncoderRid,
         pipelineRid,
       );
@@ -5020,8 +4934,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_set_index_buffer",
+      ops.op_webgpu_render_bundle_encoder_set_index_buffer(
         renderBundleEncoderRid,
         bufferRid,
         indexFormat,
@@ -5071,8 +4984,7 @@
         resourceContext: "Argument 2",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_set_vertex_buffer",
+      ops.op_webgpu_render_bundle_encoder_set_vertex_buffer(
         renderBundleEncoderRid,
         slot,
         bufferRid,
@@ -5112,8 +5024,7 @@
         prefix,
         context: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_draw",
+      ops.op_webgpu_render_bundle_encoder_draw(
         renderBundleEncoderRid,
         vertexCount,
         instanceCount,
@@ -5165,8 +5076,7 @@
         prefix,
         context: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_draw_indexed",
+      ops.op_webgpu_render_bundle_encoder_draw_indexed(
         renderBundleEncoderRid,
         indexCount,
         instanceCount,
@@ -5207,8 +5117,7 @@
         resourceContext: "Argument 1",
         selfContext: "this",
       });
-      core.opSync(
-        "op_webgpu_render_bundle_encoder_draw_indirect",
+      ops.op_webgpu_render_bundle_encoder_draw_indirect(
         renderBundleEncoderRid,
         indirectBufferRid,
         indirectOffset,
diff --git a/deno_webgpu/webgpu.idl b/deno_webgpu/webgpu.idl
index f0fec0b1b9..30b99a67c8 100644
--- a/deno_webgpu/webgpu.idl
+++ b/deno_webgpu/webgpu.idl
@@ -790,6 +790,23 @@ dictionary GPUVertexAttribute {
     required GPUIndex32 shaderLocation;
 };
 
+dictionary GPUImageDataLayout {
+    GPUSize64 offset = 0;
+    GPUSize32 bytesPerRow;
+    GPUSize32 rowsPerImage;
+};
+
+dictionary GPUImageCopyBuffer : GPUImageDataLayout {
+    required GPUBuffer buffer;
+};
+
+dictionary GPUImageCopyTexture {
+    required GPUTexture texture;
+    GPUIntegerCoordinate mipLevel = 0;
+    GPUOrigin3D origin = {};
+    GPUTextureAspect aspect = "all";
+};
+
 [Exposed=(Window, DedicatedWorker), SecureContext]
 interface GPUCommandBuffer {
 };
@@ -851,23 +868,6 @@ GPUCommandEncoder includes GPUDebugCommandsMixin;
 dictionary GPUCommandEncoderDescriptor : GPUObjectDescriptorBase {
 };
 
-dictionary GPUImageDataLayout {
-    GPUSize64 offset = 0;
-    GPUSize32 bytesPerRow;
-    GPUSize32 rowsPerImage;
-};
-
-dictionary GPUImageCopyBuffer : GPUImageDataLayout {
-    required GPUBuffer buffer;
-};
-
-dictionary GPUImageCopyTexture {
-    required GPUTexture texture;
-    GPUIntegerCoordinate mipLevel = 0;
-    GPUOrigin3D origin = {};
-    GPUTextureAspect aspect = "all";
-};
-
 interface mixin GPUBindingCommandsMixin {
     undefined setBindGroup(GPUIndex32 index, GPUBindGroup bindGroup,
         optional sequence<GPUBufferDynamicOffset> dynamicOffsets = []);
@@ -972,7 +972,7 @@ enum GPUStoreOp {
     "discard"
 };
 
-dictionary GPURenderPassLayout: GPUObjectDescriptorBase {
+dictionary GPURenderPassLayout : GPUObjectDescriptorBase {
     required sequence<GPUTextureFormat?> colorFormats;
     GPUTextureFormat depthStencilFormat;
     GPUSize32 sampleCount = 1;
diff --git a/wgpu-core/Cargo.toml b/wgpu-core/Cargo.toml
index c92865b73d..fb7c90d5ec 100644
--- a/wgpu-core/Cargo.toml
+++ b/wgpu-core/Cargo.toml
@@ -1,13 +1,13 @@
 [package]
 name = "wgpu-core"
-version.workspace = true
-authors.workspace = true
-edition.workspace = true
+version = "0.14.0"
+authors = ["wgpu developers"]
+edition = "2021"
 description = "WebGPU core logic on wgpu-hal"
-homepage.workspace = true
-repository.workspace = true
-keywords.workspace = true
-license.workspace = true
+homepage = "https://wgpu.rs/"
+repository = "https://github.com/gfx-rs/wgpu"
+keywords = ["graphics"]
+license = "MIT OR Apache-2.0"
 
 [package.metadata.docs.rs]
 all-features = true
@@ -17,6 +17,21 @@ rustdoc-args = ["--cfg", "docsrs"]
 
 [features]
 default = []
+
+# Backends, passed through to wgpu-hal
+metal = ["hal/metal"]
+vulkan = ["hal/vulkan"]
+gles = ["hal/gles"]
+dx11 = ["hal/dx11"]
+dx12 = ["hal/dx12"]
+
+# Support the Renderdoc graphics debugger:
+# https://renderdoc.org/
+renderdoc = ["hal/renderdoc"]
+
+# Compile for the Emscripten POSIX-in-a-web-page emulation environment.
+emscripten = ["hal/emscripten"]
+
 # Apply run-time checks, even in release builds. These are in addition
 # to the validation carried out at public APIs in all builds.
 strict_asserts = []
@@ -32,49 +47,38 @@ id32 = []
 wgsl = ["naga/wgsl-in"]
 vulkan-portability = ["hal/vulkan"]
 
+# Features that are intended to work on all platforms.
+portable_features = ["gles", "strict_asserts", "trace", "replay", "serial-pass", "id32", "wgsl"]
+
 [dependencies]
-arrayvec.workspace = true
-bitflags.workspace = true
-bit-vec.workspace = true
-codespan-reporting.workspace = true
-fxhash.workspace = true
-log.workspace = true
-parking_lot.workspace = true
-profiling.workspace = true
-raw-window-handle = { workspace = true, optional = true }
-ron = { workspace = true, optional = true }
-serde = { workspace = true, features = ["serde_derive"], optional = true }
-smallvec.workspace = true
-thiserror.workspace = true
+arrayvec = "0.7"
+bitflags = "1"
+bit-vec = "0.6"
+codespan-reporting = "0.11"
+fxhash = "0.2.1"
+log = "0.4"
+# parking_lot 0.12 switches from `winapi` to `windows`; permit either
+parking_lot = ">=0.11,<0.13"
+profiling = { version = "1", default-features = false }
+raw-window-handle = { version = "0.5", optional = true }
+ron = { version = "0.8", optional = true }
+serde = { version = "1", features = ["serde_derive"], optional = true }
+smallvec = "1"
+thiserror = "1"
 
 [dependencies.naga]
-workspace = true
+git = "https://github.com/gfx-rs/naga"
+rev = "e7fc8e6"
+version = "0.10"
 features = ["clone", "span", "validate"]
 
 [dependencies.wgt]
-workspace = true
+package = "wgpu-types"
+path = "../wgpu-types"
 
 [dependencies.hal]
-workspace = true
+package = "wgpu-hal"
+path = "../wgpu-hal"
 
 [target.'cfg(all(target_arch = "wasm32", not(target_os = "emscripten")))'.dependencies]
-web-sys = { workspace = true, features = ["HtmlCanvasElement", "OffscreenCanvas"] }
-
-[target.'cfg(target_arch = "wasm32")'.dependencies]
-hal = { workspace = true, features = ["gles"] }
-
-[target.'cfg(all(not(target_arch = "wasm32"), any(target_os = "ios", target_os = "macos")))'.dependencies]
-hal = { workspace = true, features = ["metal"] }
-#Note: could also enable "vulkan" for Vulkan Portability
-
-[target.'cfg(all(not(target_arch = "wasm32"), unix, not(target_os = "ios"), not(target_os = "macos")))'.dependencies]
-hal = { workspace = true, features = ["vulkan", "gles", "renderdoc"] }
-
-[target.'cfg(all(not(target_arch = "wasm32"), windows))'.dependencies]
-hal = { workspace = true, features = ["vulkan", "dx12", "dx11", "renderdoc"] }
-
-[target.'cfg(target_os = "emscripten")'.dependencies]
-hal = { workspace = true, features = ["emscripten"] }
-
-[build-dependencies]
-cfg_aliases.workspace = true
+web-sys = { version = "0.3.60", features = ["HtmlCanvasElement", "OffscreenCanvas"] }
diff --git a/wgpu-core/build.rs b/wgpu-core/build.rs
deleted file mode 100644
index 445ea245d2..0000000000
--- a/wgpu-core/build.rs
+++ /dev/null
@@ -1,22 +0,0 @@
-fn main() {
-    // Setup cfg aliases
-    cfg_aliases::cfg_aliases! {
-        // Vendors/systems
-        wasm: { target_arch = "wasm32" },
-        apple: { any(target_os = "ios", target_os = "macos") },
-        unix_wo_apple: {all(unix, not(apple))},
-
-        // Backends
-        vulkan: { all(not(wasm), any(windows, unix_wo_apple, feature = "vulkan-portability")) },
-        metal: { all(not(wasm), apple) },
-        dx12: { all(not(wasm), windows) },
-        dx11: { all(not(wasm), windows) },
-        gl: {
-            any(
-                unix_wo_apple,
-                feature = "angle",
-                wasm
-            )
-        },
-    }
-}
diff --git a/wgpu-core/src/binding_model.rs b/wgpu-core/src/binding_model.rs
index 5feef3bab4..d64d563b25 100644
--- a/wgpu-core/src/binding_model.rs
+++ b/wgpu-core/src/binding_model.rs
@@ -683,23 +683,56 @@ pub enum BindingResource<'a> {
 
 #[derive(Clone, Debug, Error)]
 pub enum BindError {
-    #[error("number of dynamic offsets ({actual}) doesn't match the number of dynamic bindings in the bind group layout ({expected})")]
-    MismatchedDynamicOffsetCount { actual: usize, expected: usize },
     #[error(
-        "dynamic binding at index {idx}: offset {offset} does not respect device's requested `{limit_name}` limit {alignment}"
+        "Bind group {group} expects {expected} dynamic offset{s0}. However {actual} dynamic offset{s1} were provided.",
+        s0 = if *.expected >= 2 { "s" } else { "" },
+        s1 = if *.actual >= 2 { "s" } else { "" },
+    )]
+    MismatchedDynamicOffsetCount {
+        group: u8,
+        actual: usize,
+        expected: usize,
+    },
+    #[error(
+        "Dynamic binding index {idx} (targeting bind group {group}, binding {binding}) with value {offset}, does not respect device's requested `{limit_name}` limit: {alignment}"
     )]
     UnalignedDynamicBinding {
         idx: usize,
+        group: u8,
+        binding: u32,
         offset: u32,
         alignment: u32,
         limit_name: &'static str,
     },
-    #[error("dynamic binding at index {idx} with offset {offset} would overrun the buffer (limit: {max})")]
-    DynamicBindingOutOfBounds { idx: usize, offset: u32, max: u64 },
+    #[error(
+        "Dynamic binding offset index {idx} with offset {offset} would overrun the buffer bound to bind group {group} -> binding {binding}. \
+         Buffer size is {buffer_size} bytes, the binding binds bytes {binding_range:?}, meaning the maximum the binding can be offset is {maximum_dynamic_offset} bytes",
+    )]
+    DynamicBindingOutOfBounds {
+        idx: usize,
+        group: u8,
+        binding: u32,
+        offset: u32,
+        buffer_size: wgt::BufferAddress,
+        binding_range: Range<wgt::BufferAddress>,
+        maximum_dynamic_offset: wgt::BufferAddress,
+    },
 }
 
 #[derive(Debug)]
 pub struct BindGroupDynamicBindingData {
+    /// The index of the binding.
+    ///
+    /// Used for more descriptive errors.
+    pub(crate) binding_idx: u32,
+    /// The size of the buffer.
+    ///
+    /// Used for more descriptive errors.
+    pub(crate) buffer_size: wgt::BufferAddress,
+    /// The range that the binding covers.
+    ///
+    /// Used for more descriptive errors.
+    pub(crate) binding_range: Range<wgt::BufferAddress>,
     /// The maximum value the dynamic offset can have before running off the end of the buffer.
     pub(crate) maximum_dynamic_offset: wgt::BufferAddress,
     /// The binding type.
@@ -739,11 +772,13 @@ pub struct BindGroup<A: HalApi> {
 impl<A: HalApi> BindGroup<A> {
     pub(crate) fn validate_dynamic_bindings(
         &self,
+        bind_group_index: u8,
         offsets: &[wgt::DynamicOffset],
         limits: &wgt::Limits,
     ) -> Result<(), BindError> {
         if self.dynamic_binding_info.len() != offsets.len() {
             return Err(BindError::MismatchedDynamicOffsetCount {
+                group: bind_group_index,
                 expected: self.dynamic_binding_info.len(),
                 actual: offsets.len(),
             });
@@ -758,6 +793,8 @@ impl<A: HalApi> BindGroup<A> {
             let (alignment, limit_name) = buffer_binding_type_alignment(limits, info.binding_type);
             if offset as wgt::BufferAddress % alignment as u64 != 0 {
                 return Err(BindError::UnalignedDynamicBinding {
+                    group: bind_group_index,
+                    binding: info.binding_idx,
                     idx,
                     offset,
                     alignment,
@@ -767,9 +804,13 @@ impl<A: HalApi> BindGroup<A> {
 
             if offset as wgt::BufferAddress > info.maximum_dynamic_offset {
                 return Err(BindError::DynamicBindingOutOfBounds {
+                    group: bind_group_index,
+                    binding: info.binding_idx,
                     idx,
                     offset,
-                    max: info.maximum_dynamic_offset,
+                    buffer_size: info.buffer_size,
+                    binding_range: info.binding_range.clone(),
+                    maximum_dynamic_offset: info.maximum_dynamic_offset,
                 });
             }
         }
diff --git a/wgpu-core/src/command/compute.rs b/wgpu-core/src/command/compute.rs
index c228519595..dec6211a73 100644
--- a/wgpu-core/src/command/compute.rs
+++ b/wgpu-core/src/command/compute.rs
@@ -428,7 +428,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
                         .ok_or(ComputePassErrorInner::InvalidBindGroup(bind_group_id))
                         .map_pass_err(scope)?;
                     bind_group
-                        .validate_dynamic_bindings(&temp_offsets, &cmd_buf.limits)
+                        .validate_dynamic_bindings(index, &temp_offsets, &cmd_buf.limits)
                         .map_pass_err(scope)?;
 
                     cmd_buf.buffer_memory_init_actions.extend(
diff --git a/wgpu-core/src/command/render.rs b/wgpu-core/src/command/render.rs
index 09af0bbe6a..04e27ec062 100644
--- a/wgpu-core/src/command/render.rs
+++ b/wgpu-core/src/command/render.rs
@@ -1238,7 +1238,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
                             .ok_or(RenderCommandError::InvalidBindGroup(bind_group_id))
                             .map_pass_err(scope)?;
                         bind_group
-                            .validate_dynamic_bindings(&temp_offsets, &cmd_buf.limits)
+                            .validate_dynamic_bindings(index, &temp_offsets, &cmd_buf.limits)
                             .map_pass_err(scope)?;
 
                         // merge the resource tracker in
diff --git a/wgpu-core/src/device/mod.rs b/wgpu-core/src/device/mod.rs
index 86c79a9a6d..cdfc0bea4f 100644
--- a/wgpu-core/src/device/mod.rs
+++ b/wgpu-core/src/device/mod.rs
@@ -1732,6 +1732,9 @@ impl<A: HalApi> Device<A> {
         // Record binding info for validating dynamic offsets
         if dynamic {
             dynamic_binding_info.push(binding_model::BindGroupDynamicBindingData {
+                binding_idx: binding,
+                buffer_size: buffer.size,
+                binding_range: bb.offset..bind_end,
                 maximum_dynamic_offset: buffer.size - bind_end,
                 binding_type: binding_ty,
             });
@@ -5407,27 +5410,27 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
         let mut closures = UserClosures::default();
         let mut all_queue_empty = true;
 
-        #[cfg(vulkan)]
+        #[cfg(feature = "vulkan")]
         {
             all_queue_empty = self.poll_devices::<hal::api::Vulkan>(force_wait, &mut closures)?
                 && all_queue_empty;
         }
-        #[cfg(metal)]
+        #[cfg(feature = "metal")]
         {
             all_queue_empty =
                 self.poll_devices::<hal::api::Metal>(force_wait, &mut closures)? && all_queue_empty;
         }
-        #[cfg(dx12)]
+        #[cfg(feature = "dx12")]
         {
             all_queue_empty =
                 self.poll_devices::<hal::api::Dx12>(force_wait, &mut closures)? && all_queue_empty;
         }
-        #[cfg(dx11)]
+        #[cfg(feature = "dx11")]
         {
             all_queue_empty =
                 self.poll_devices::<hal::api::Dx11>(force_wait, &mut closures)? && all_queue_empty;
         }
-        #[cfg(gl)]
+        #[cfg(feature = "gles")]
         {
             all_queue_empty =
                 self.poll_devices::<hal::api::Gles>(force_wait, &mut closures)? && all_queue_empty;
diff --git a/wgpu-core/src/hub.rs b/wgpu-core/src/hub.rs
index f07010477c..290df3c716 100644
--- a/wgpu-core/src/hub.rs
+++ b/wgpu-core/src/hub.rs
@@ -1057,30 +1057,30 @@ impl<A: HalApi, F: GlobalIdentityHandlerFactory> Hub<A, F> {
 }
 
 pub struct Hubs<F: GlobalIdentityHandlerFactory> {
-    #[cfg(vulkan)]
+    #[cfg(feature = "vulkan")]
     vulkan: Hub<hal::api::Vulkan, F>,
-    #[cfg(metal)]
+    #[cfg(feature = "metal")]
     metal: Hub<hal::api::Metal, F>,
-    #[cfg(dx12)]
+    #[cfg(feature = "dx12")]
     dx12: Hub<hal::api::Dx12, F>,
-    #[cfg(dx11)]
+    #[cfg(feature = "dx11")]
     dx11: Hub<hal::api::Dx11, F>,
-    #[cfg(gl)]
+    #[cfg(feature = "gles")]
     gl: Hub<hal::api::Gles, F>,
 }
 
 impl<F: GlobalIdentityHandlerFactory> Hubs<F> {
     fn new(factory: &F) -> Self {
         Self {
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             vulkan: Hub::new(factory),
-            #[cfg(metal)]
+            #[cfg(feature = "metal")]
             metal: Hub::new(factory),
-            #[cfg(dx12)]
+            #[cfg(feature = "dx12")]
             dx12: Hub::new(factory),
-            #[cfg(dx11)]
+            #[cfg(feature = "dx11")]
             dx11: Hub::new(factory),
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             gl: Hub::new(factory),
         }
     }
@@ -1089,15 +1089,15 @@ impl<F: GlobalIdentityHandlerFactory> Hubs<F> {
 #[derive(Debug)]
 pub struct GlobalReport {
     pub surfaces: StorageReport,
-    #[cfg(vulkan)]
+    #[cfg(feature = "vulkan")]
     pub vulkan: Option<HubReport>,
-    #[cfg(metal)]
+    #[cfg(feature = "metal")]
     pub metal: Option<HubReport>,
-    #[cfg(dx12)]
+    #[cfg(feature = "dx12")]
     pub dx12: Option<HubReport>,
-    #[cfg(dx11)]
+    #[cfg(feature = "dx11")]
     pub dx11: Option<HubReport>,
-    #[cfg(gl)]
+    #[cfg(feature = "gles")]
     pub gl: Option<HubReport>,
 }
 
@@ -1162,31 +1162,31 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
     pub fn generate_report(&self) -> GlobalReport {
         GlobalReport {
             surfaces: self.surfaces.data.read().generate_report(),
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             vulkan: if self.instance.vulkan.is_some() {
                 Some(self.hubs.vulkan.generate_report())
             } else {
                 None
             },
-            #[cfg(metal)]
+            #[cfg(feature = "metal")]
             metal: if self.instance.metal.is_some() {
                 Some(self.hubs.metal.generate_report())
             } else {
                 None
             },
-            #[cfg(dx12)]
+            #[cfg(feature = "dx12")]
             dx12: if self.instance.dx12.is_some() {
                 Some(self.hubs.dx12.generate_report())
             } else {
                 None
             },
-            #[cfg(dx11)]
+            #[cfg(feature = "dx11")]
             dx11: if self.instance.dx11.is_some() {
                 Some(self.hubs.dx11.generate_report())
             } else {
                 None
             },
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             gl: if self.instance.gl.is_some() {
                 Some(self.hubs.gl.generate_report())
             } else {
@@ -1203,23 +1203,23 @@ impl<G: GlobalIdentityHandlerFactory> Drop for Global<G> {
         let mut surface_guard = self.surfaces.data.write();
 
         // destroy hubs before the instance gets dropped
-        #[cfg(vulkan)]
+        #[cfg(feature = "vulkan")]
         {
             self.hubs.vulkan.clear(&mut surface_guard, true);
         }
-        #[cfg(metal)]
+        #[cfg(feature = "metal")]
         {
             self.hubs.metal.clear(&mut surface_guard, true);
         }
-        #[cfg(dx12)]
+        #[cfg(feature = "dx12")]
         {
             self.hubs.dx12.clear(&mut surface_guard, true);
         }
-        #[cfg(dx11)]
+        #[cfg(feature = "dx11")]
         {
             self.hubs.dx11.clear(&mut surface_guard, true);
         }
-        #[cfg(gl)]
+        #[cfg(feature = "gles")]
         {
             self.hubs.gl.clear(&mut surface_guard, true);
         }
@@ -1261,7 +1261,7 @@ impl HalApi for hal::api::Empty {
     }
 }
 
-#[cfg(vulkan)]
+#[cfg(feature = "vulkan")]
 impl HalApi for hal::api::Vulkan {
     const VARIANT: Backend = Backend::Vulkan;
     fn create_instance_from_hal(name: &str, hal_instance: Self::Instance) -> Instance {
@@ -1285,7 +1285,7 @@ impl HalApi for hal::api::Vulkan {
     }
 }
 
-#[cfg(metal)]
+#[cfg(feature = "metal")]
 impl HalApi for hal::api::Metal {
     const VARIANT: Backend = Backend::Metal;
     fn create_instance_from_hal(name: &str, hal_instance: Self::Instance) -> Instance {
@@ -1309,7 +1309,7 @@ impl HalApi for hal::api::Metal {
     }
 }
 
-#[cfg(dx12)]
+#[cfg(feature = "dx12")]
 impl HalApi for hal::api::Dx12 {
     const VARIANT: Backend = Backend::Dx12;
     fn create_instance_from_hal(name: &str, hal_instance: Self::Instance) -> Instance {
@@ -1333,7 +1333,7 @@ impl HalApi for hal::api::Dx12 {
     }
 }
 
-#[cfg(dx11)]
+#[cfg(feature = "dx11")]
 impl HalApi for hal::api::Dx11 {
     const VARIANT: Backend = Backend::Dx11;
     fn create_instance_from_hal(name: &str, hal_instance: Self::Instance) -> Instance {
@@ -1357,7 +1357,7 @@ impl HalApi for hal::api::Dx11 {
     }
 }
 
-#[cfg(gl)]
+#[cfg(feature = "gles")]
 impl HalApi for hal::api::Gles {
     const VARIANT: Backend = Backend::Gl;
     fn create_instance_from_hal(name: &str, hal_instance: Self::Instance) -> Instance {
diff --git a/wgpu-core/src/instance.rs b/wgpu-core/src/instance.rs
index f139409906..7cb8a4f1fd 100644
--- a/wgpu-core/src/instance.rs
+++ b/wgpu-core/src/instance.rs
@@ -54,15 +54,15 @@ fn downlevel_default_limits_less_than_default_limits() {
 pub struct Instance {
     #[allow(dead_code)]
     pub name: String,
-    #[cfg(vulkan)]
+    #[cfg(feature = "vulkan")]
     pub vulkan: Option<HalInstance<hal::api::Vulkan>>,
-    #[cfg(metal)]
+    #[cfg(feature = "metal")]
     pub metal: Option<HalInstance<hal::api::Metal>>,
-    #[cfg(dx12)]
+    #[cfg(feature = "dx12")]
     pub dx12: Option<HalInstance<hal::api::Dx12>>,
-    #[cfg(dx11)]
+    #[cfg(feature = "dx11")]
     pub dx11: Option<HalInstance<hal::api::Dx11>>,
-    #[cfg(gl)]
+    #[cfg(feature = "gles")]
     pub gl: Option<HalInstance<hal::api::Gles>>,
 }
 
@@ -87,15 +87,15 @@ impl Instance {
 
         Self {
             name: name.to_string(),
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             vulkan: init(hal::api::Vulkan, backends),
-            #[cfg(metal)]
+            #[cfg(feature = "metal")]
             metal: init(hal::api::Metal, backends),
-            #[cfg(dx12)]
+            #[cfg(feature = "dx12")]
             dx12: init(hal::api::Dx12, backends),
-            #[cfg(dx11)]
+            #[cfg(feature = "dx11")]
             dx11: init(hal::api::Dx11, backends),
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             gl: init(hal::api::Gles, backends),
         }
     }
@@ -112,30 +112,30 @@ impl Instance {
                 }
             }
         }
-        #[cfg(vulkan)]
+        #[cfg(feature = "vulkan")]
         destroy(hal::api::Vulkan, &self.vulkan, surface.vulkan);
-        #[cfg(metal)]
+        #[cfg(feature = "metal")]
         destroy(hal::api::Metal, &self.metal, surface.metal);
-        #[cfg(dx12)]
+        #[cfg(feature = "dx12")]
         destroy(hal::api::Dx12, &self.dx12, surface.dx12);
-        #[cfg(dx11)]
+        #[cfg(feature = "dx11")]
         destroy(hal::api::Dx11, &self.dx11, surface.dx11);
-        #[cfg(gl)]
+        #[cfg(feature = "gles")]
         destroy(hal::api::Gles, &self.gl, surface.gl);
     }
 }
 
 pub struct Surface {
     pub(crate) presentation: Option<Presentation>,
-    #[cfg(vulkan)]
+    #[cfg(feature = "vulkan")]
     pub vulkan: Option<HalSurface<hal::api::Vulkan>>,
-    #[cfg(metal)]
+    #[cfg(feature = "metal")]
     pub metal: Option<HalSurface<hal::api::Metal>>,
-    #[cfg(dx12)]
+    #[cfg(feature = "dx12")]
     pub dx12: Option<HalSurface<hal::api::Dx12>>,
-    #[cfg(dx11)]
+    #[cfg(feature = "dx11")]
     pub dx11: Option<HalSurface<hal::api::Dx11>>,
-    #[cfg(gl)]
+    #[cfg(feature = "gles")]
     pub gl: Option<HalSurface<hal::api::Gles>>,
 }
 
@@ -451,15 +451,15 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
 
         let surface = Surface {
             presentation: None,
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             vulkan: init::<hal::api::Vulkan>(&self.instance.vulkan, display_handle, window_handle),
-            #[cfg(metal)]
+            #[cfg(feature = "metal")]
             metal: init::<hal::api::Metal>(&self.instance.metal, display_handle, window_handle),
-            #[cfg(dx12)]
+            #[cfg(feature = "dx12")]
             dx12: init::<hal::api::Dx12>(&self.instance.dx12, display_handle, window_handle),
-            #[cfg(dx11)]
+            #[cfg(feature = "dx11")]
             dx11: init::<hal::api::Dx11>(&self.instance.dx11, display_handle, window_handle),
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             gl: init::<hal::api::Gles>(&self.instance.gl, display_handle, window_handle),
         };
 
@@ -468,7 +468,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
         id.0
     }
 
-    #[cfg(metal)]
+    #[cfg(feature = "metal")]
     pub fn instance_create_surface_metal(
         &self,
         layer: *mut std::ffi::c_void,
@@ -486,9 +486,9 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
                 },
                 //acquired_texture: None,
             }),
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             vulkan: None,
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             gl: None,
         };
 
@@ -502,22 +502,26 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
         &self,
         canvas: &web_sys::HtmlCanvasElement,
         id_in: Input<G, SurfaceId>,
-    ) -> SurfaceId {
+    ) -> Result<SurfaceId, hal::InstanceError> {
         profiling::scope!("Instance::create_surface_webgl_canvas");
 
         let surface = Surface {
             presentation: None,
-            gl: self.instance.gl.as_ref().map(|inst| HalSurface {
-                raw: {
-                    inst.create_surface_from_canvas(canvas)
-                        .expect("Create surface from canvas")
-                },
-            }),
+            gl: self
+                .instance
+                .gl
+                .as_ref()
+                .map(|inst| {
+                    Ok(HalSurface {
+                        raw: inst.create_surface_from_canvas(canvas)?,
+                    })
+                })
+                .transpose()?,
         };
 
         let mut token = Token::root();
         let id = self.surfaces.prepare(id_in).assign(surface, &mut token);
-        id.0
+        Ok(id.0)
     }
 
     #[cfg(all(target_arch = "wasm32", not(target_os = "emscripten")))]
@@ -525,25 +529,29 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
         &self,
         canvas: &web_sys::OffscreenCanvas,
         id_in: Input<G, SurfaceId>,
-    ) -> SurfaceId {
+    ) -> Result<SurfaceId, hal::InstanceError> {
         profiling::scope!("Instance::create_surface_webgl_offscreen_canvas");
 
         let surface = Surface {
             presentation: None,
-            gl: self.instance.gl.as_ref().map(|inst| HalSurface {
-                raw: {
-                    inst.create_surface_from_offscreen_canvas(canvas)
-                        .expect("Create surface from offscreen canvas")
-                },
-            }),
+            gl: self
+                .instance
+                .gl
+                .as_ref()
+                .map(|inst| {
+                    Ok(HalSurface {
+                        raw: inst.create_surface_from_offscreen_canvas(canvas)?,
+                    })
+                })
+                .transpose()?,
         };
 
         let mut token = Token::root();
         let id = self.surfaces.prepare(id_in).assign(surface, &mut token);
-        id.0
+        Ok(id.0)
     }
 
-    #[cfg(dx12)]
+    #[cfg(feature = "dx12")]
     /// # Safety
     ///
     /// The visual must be valid and able to be used to make a swapchain with.
@@ -556,13 +564,13 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
 
         let surface = Surface {
             presentation: None,
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             vulkan: None,
             dx12: self.instance.dx12.as_ref().map(|inst| HalSurface {
                 raw: unsafe { inst.create_surface_from_visual(visual as _) },
             }),
             dx11: None,
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             gl: None,
         };
 
@@ -615,25 +623,25 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
 
         let mut adapters = Vec::new();
 
-        #[cfg(vulkan)]
+        #[cfg(feature = "vulkan")]
         self.enumerate(
             hal::api::Vulkan,
             &self.instance.vulkan,
             &inputs,
             &mut adapters,
         );
-        #[cfg(metal)]
+        #[cfg(feature = "metal")]
         self.enumerate(
             hal::api::Metal,
             &self.instance.metal,
             &inputs,
             &mut adapters,
         );
-        #[cfg(dx12)]
+        #[cfg(feature = "dx12")]
         self.enumerate(hal::api::Dx12, &self.instance.dx12, &inputs, &mut adapters);
-        #[cfg(dx11)]
+        #[cfg(feature = "dx11")]
         self.enumerate(hal::api::Dx11, &self.instance.dx11, &inputs, &mut adapters);
-        #[cfg(gl)]
+        #[cfg(feature = "gles")]
         self.enumerate(hal::api::Gles, &self.instance.gl, &inputs, &mut adapters);
 
         adapters
@@ -716,7 +724,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
             .transpose()?;
         let mut device_types = Vec::new();
 
-        #[cfg(vulkan)]
+        #[cfg(feature = "vulkan")]
         let (id_vulkan, adapters_vk) = gather(
             hal::api::Vulkan,
             self.instance.vulkan.as_ref(),
@@ -725,7 +733,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
             desc.force_fallback_adapter,
             &mut device_types,
         );
-        #[cfg(metal)]
+        #[cfg(feature = "metal")]
         let (id_metal, adapters_metal) = gather(
             hal::api::Metal,
             self.instance.metal.as_ref(),
@@ -734,7 +742,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
             desc.force_fallback_adapter,
             &mut device_types,
         );
-        #[cfg(dx12)]
+        #[cfg(feature = "dx12")]
         let (id_dx12, adapters_dx12) = gather(
             hal::api::Dx12,
             self.instance.dx12.as_ref(),
@@ -743,7 +751,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
             desc.force_fallback_adapter,
             &mut device_types,
         );
-        #[cfg(dx11)]
+        #[cfg(feature = "dx11")]
         let (id_dx11, adapters_dx11) = gather(
             hal::api::Dx11,
             self.instance.dx11.as_ref(),
@@ -752,7 +760,7 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
             desc.force_fallback_adapter,
             &mut device_types,
         );
-        #[cfg(gl)]
+        #[cfg(feature = "gles")]
         let (id_gl, adapters_gl) = gather(
             hal::api::Gles,
             self.instance.gl.as_ref(),
@@ -805,23 +813,23 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
         };
 
         let mut selected = preferred_gpu.unwrap_or(0);
-        #[cfg(vulkan)]
+        #[cfg(feature = "vulkan")]
         if let Some(id) = self.select(&mut selected, id_vulkan, adapters_vk) {
             return Ok(id);
         }
-        #[cfg(metal)]
+        #[cfg(feature = "metal")]
         if let Some(id) = self.select(&mut selected, id_metal, adapters_metal) {
             return Ok(id);
         }
-        #[cfg(dx12)]
+        #[cfg(feature = "dx12")]
         if let Some(id) = self.select(&mut selected, id_dx12, adapters_dx12) {
             return Ok(id);
         }
-        #[cfg(dx11)]
+        #[cfg(feature = "dx11")]
         if let Some(id) = self.select(&mut selected, id_dx11, adapters_dx11) {
             return Ok(id);
         }
-        #[cfg(gl)]
+        #[cfg(feature = "gles")]
         if let Some(id) = self.select(&mut selected, id_gl, adapters_gl) {
             return Ok(id);
         }
@@ -845,15 +853,15 @@ impl<G: GlobalIdentityHandlerFactory> Global<G> {
         let fid = A::hub(self).adapters.prepare(input);
 
         match A::VARIANT {
-            #[cfg(vulkan)]
+            #[cfg(feature = "vulkan")]
             Backend::Vulkan => fid.assign(Adapter::new(hal_adapter), &mut token).0,
-            #[cfg(metal)]
+            #[cfg(feature = "metal")]
             Backend::Metal => fid.assign(Adapter::new(hal_adapter), &mut token).0,
-            #[cfg(dx12)]
+            #[cfg(feature = "dx12")]
             Backend::Dx12 => fid.assign(Adapter::new(hal_adapter), &mut token).0,
-            #[cfg(dx11)]
+            #[cfg(feature = "dx11")]
             Backend::Dx11 => fid.assign(Adapter::new(hal_adapter), &mut token).0,
-            #[cfg(gl)]
+            #[cfg(feature = "gles")]
             Backend::Gl => fid.assign(Adapter::new(hal_adapter), &mut token).0,
             _ => unreachable!(),
         }
diff --git a/wgpu-core/src/lib.rs b/wgpu-core/src/lib.rs
index ce0f7f5087..c8b3fb7296 100644
--- a/wgpu-core/src/lib.rs
+++ b/wgpu-core/src/lib.rs
@@ -243,6 +243,91 @@ If you are running this program on native and not in a browser and wish to work
 Adapter::downlevel_properties or Device::downlevel_properties to get a listing of the features the current \
 platform supports.";
 
+// #[cfg] attributes in exported macros are interesting!
+//
+// The #[cfg] conditions in a macro's expansion are evaluated using the
+// configuration options (features, target architecture and os, etc.) in force
+// where the macro is *used*, not where it is *defined*. That is, if crate A
+// defines a macro like this:
+//
+//     #[macro_export]
+//     macro_rules! if_bleep {
+//         { } => {
+//             #[cfg(feature = "bleep")]
+//             bleep();
+//         }
+//     }
+//
+// and then crate B uses it like this:
+//
+//     fn f() {
+//         if_bleep! { }
+//     }
+//
+// then it is crate B's `"bleep"` feature, not crate A's, that determines
+// whether the macro expands to a function call or an empty statement. The
+// entire configuration predicate is evaluated in the use's context, not the
+// definition's.
+//
+// Since `wgpu-core` selects back ends using features, we need to make sure the
+// arms of the `gfx_select!` macro are pruned according to `wgpu-core`'s
+// features, not those of whatever crate happens to be using `gfx_select!`. This
+// means we can't use `#[cfg]` attributes in `gfx_select!`s definition itself.
+// Instead, for each backend, `gfx_select!` must use a macro whose definition is
+// selected by `#[cfg]` in `wgpu-core`. The configuration predicate is still
+// evaluated when the macro is used; we've just moved the `#[cfg]` into a macro
+// used by `wgpu-core` itself.
+
+/// Define an exported macro named `$public` that expands to an expression if
+/// the feature `$feature` is enabled, or to a panic otherwise.
+///
+/// For a call like this:
+///
+///    define_backend_caller! { name, hidden_name, feature }
+///
+/// define a macro `name`, used like this:
+///
+///    name!(expr)
+///
+/// that expands to `expr` if `feature` is enabled, or a panic otherwise.
+///
+/// Because of odd technical limitations on exporting macros expanded by other
+/// macros, you must supply both a public-facing name for the macro and a
+/// private name, which is never used outside this macro. For details:
+/// <https://github.com/rust-lang/rust/pull/52234#issuecomment-976702997>
+macro_rules! define_backend_caller {
+    { $public:ident, $private:ident if $feature:literal } => {
+        #[cfg(feature = $feature )]
+        #[macro_export]
+        macro_rules! $private {
+            ( $call:expr ) => ( $call )
+        }
+
+        #[cfg(not(feature = $feature ))]
+        #[macro_export]
+        macro_rules! $private {
+            ( $call:expr ) => (
+                panic!("Identifier refers to disabled backend feature {:?}", $feature)
+            )
+        }
+
+        // See note about rust-lang#52234 above.
+        #[doc(hidden)] pub use $private as $public;
+    }
+}
+
+// Define a macro for each `gfx_select!` match arm. For example,
+//
+//     gfx_if_vulkan!(expr)
+//
+// expands to `expr` if the `"vulkan"` feature is enabled, or to a panic
+// otherwise.
+define_backend_caller! { gfx_if_vulkan, gfx_if_vulkan_hidden if "vulkan" }
+define_backend_caller! { gfx_if_metal, gfx_if_metal_hidden if "metal" }
+define_backend_caller! { gfx_if_dx12, gfx_if_dx12_hidden if "dx12" }
+define_backend_caller! { gfx_if_dx11, gfx_if_dx11_hidden if "dx11" }
+define_backend_caller! { gfx_if_gles, gfx_if_gles_hidden if "gles" }
+
 /// Dispatch on an [`Id`]'s backend to a backend-generic method.
 ///
 /// Uses of this macro have the form:
@@ -291,29 +376,13 @@ platform supports.";
 #[macro_export]
 macro_rules! gfx_select {
     ($id:expr => $global:ident.$method:ident( $($param:expr),* )) => {
-        // Note: For some reason the cfg aliases defined in build.rs
-        // don't succesfully apply in this macro so we must specify
-        // their equivalents manually.
         match $id.backend() {
-            #[cfg(any(
-                all(not(target_arch = "wasm32"), not(target_os = "ios"), not(target_os = "macos")),
-                feature = "vulkan-portability"
-            ))]
-            wgt::Backend::Vulkan => $global.$method::<$crate::api::Vulkan>( $($param),* ),
-            #[cfg(all(not(target_arch = "wasm32"), any(target_os = "ios", target_os = "macos")))]
-            wgt::Backend::Metal => $global.$method::<$crate::api::Metal>( $($param),* ),
-            #[cfg(all(not(target_arch = "wasm32"), windows))]
-            wgt::Backend::Dx12 => $global.$method::<$crate::api::Dx12>( $($param),* ),
-            #[cfg(all(not(target_arch = "wasm32"), windows))]
-            wgt::Backend::Dx11 => $global.$method::<$crate::api::Dx11>( $($param),* ),
-            #[cfg(any(
-                all(unix, not(target_os = "macos"), not(target_os = "ios")),
-                feature = "angle",
-                target_arch = "wasm32"
-            ))]
-            wgt::Backend::Gl => $global.$method::<$crate::api::Gles>( $($param),+ ),
+            wgt::Backend::Vulkan => $crate::gfx_if_vulkan!($global.$method::<$crate::api::Vulkan>( $($param),* )),
+            wgt::Backend::Metal => $crate::gfx_if_metal!($global.$method::<$crate::api::Metal>( $($param),* )),
+            wgt::Backend::Dx12 => $crate::gfx_if_dx12!($global.$method::<$crate::api::Dx12>( $($param),* )),
+            wgt::Backend::Dx11 => $crate::gfx_if_dx11!($global.$method::<$crate::api::Dx11>( $($param),* )),
+            wgt::Backend::Gl => $crate::gfx_if_gles!($global.$method::<$crate::api::Gles>( $($param),+ )),
             other => panic!("Unexpected backend {:?}", other),
-
         }
     };
 }
diff --git a/wgpu-core/src/resource.rs b/wgpu-core/src/resource.rs
index 4302df7e61..e59ec0aaa1 100644
--- a/wgpu-core/src/resource.rs
+++ b/wgpu-core/src/resource.rs
@@ -551,9 +551,9 @@ pub struct TextureViewDescriptor<'a> {
     pub format: Option<wgt::TextureFormat>,
     /// The dimension of the texture view.
     ///
-    /// - For 1D textures, this must be `1D`.
+    /// - For 1D textures, this must be `D1`.
     /// - For 2D textures it must be one of `D2`, `D2Array`, `Cube`, or `CubeArray`.
-    /// - For 3D textures it must be `3D`.
+    /// - For 3D textures it must be `D3`.
     pub dimension: Option<wgt::TextureViewDimension>,
     /// Range within the texture that is accessible via this view.
     pub range: wgt::ImageSubresourceRange,
diff --git a/wgpu-hal/Cargo.toml b/wgpu-hal/Cargo.toml
index 2f79c2e51e..2d333930c6 100644
--- a/wgpu-hal/Cargo.toml
+++ b/wgpu-hal/Cargo.toml
@@ -1,13 +1,13 @@
 [package]
 name = "wgpu-hal"
-version.workspace = true
-authors.workspace = true
-edition.workspace = true
+version = "0.14.0"
+authors = ["wgpu developers"]
+edition = "2021"
 description = "WebGPU hardware abstraction layer"
-homepage.workspace = true
-repository.workspace = true
-keywords.workspace = true
-license.workspace = true
+homepage = "https://wgpu.rs/"
+repository = "https://github.com/gfx-rs/wgpu"
+keywords = ["graphics"]
+license = "MIT OR Apache-2.0"
 
 # Override the workspace's `rust-version` key. Firefox uses `cargo vendor` to
 # copy the crates it actually uses out of the workspace, so it's meaningful for
@@ -27,7 +27,7 @@ rustdoc-args = ["--cfg", "docsrs"]
 [lib]
 
 [features]
-default = ["gles"]
+default = []
 metal = ["naga/msl-out", "block", "foreign-types"]
 vulkan = ["naga/spv-out", "ash", "gpu-alloc", "gpu-descriptor", "libloading", "smallvec"]
 gles = ["naga/glsl-out", "glow", "egl", "libloading"]
@@ -44,79 +44,84 @@ name = "raw-gles"
 required-features = ["gles"]
 
 [dependencies]
-bitflags.workspace = true
-parking_lot.workspace = true
-profiling.workspace = true
-raw-window-handle.workspace = true
-thiserror.workspace = true
+bitflags = "1"
+parking_lot = ">=0.11,<0.13"
+profiling = { version = "1", default-features = false }
+raw-window-handle = "0.5"
+thiserror = "1"
 
 # backends common
-arrayvec.workspace = true
-fxhash.workspace = true
-log.workspace = true
-renderdoc-sys = { workspace = true, optional = true }
+arrayvec = "0.7"
+fxhash = "0.2.1"
+log = "0.4"
+renderdoc-sys = { version = "0.7.1", optional = true }
 
 # backend: Metal
-block = { workspace = true, optional = true }
-foreign-types = { workspace = true, optional = true }
+block = { version = "0.1", optional = true }
+foreign-types = { version = "0.3", optional = true }
 
 # backend: Vulkan
-ash = { workspace = true, optional = true }
-gpu-alloc = { workspace = true, optional = true }
-gpu-descriptor = { workspace = true, optional = true }
-smallvec = { workspace = true, optional = true, features = ["union"] }
+ash = { version = "0.37.1", optional = true }
+gpu-alloc = { version = "0.5", optional = true }
+gpu-descriptor = { version = "0.2", optional = true }
+smallvec = { version = "1", optional = true, features = ["union"] }
 
 # backend: Gles
-glow = { workspace = true, optional = true }
+glow = { git = "https://github.com/grovesNL/glow", rev = "c8a011fcd57a5c68cc917ed394baa484bdefc909", optional = true }
 
 # backend: Dx12
-bit-set = { workspace = true, optional = true }
-range-alloc = { workspace = true, optional = true }
+bit-set = { version = "0.5", optional = true }
+range-alloc = { version = "0.1", optional = true }
 
 [dependencies.wgt]
-workspace = true
+package = "wgpu-types"
+path = "../wgpu-types"
 
 [target.'cfg(not(target_arch = "wasm32"))'.dependencies]
-egl = { workspace = true, features = ["dynamic"], optional = true }
-libloading = { workspace = true, optional = true }
+egl = { package = "khronos-egl", version = "4.1", features = ["dynamic"], optional = true }
+libloading = { version = "0.7", optional = true }
 
 [target.'cfg(target_os = "emscripten")'.dependencies]
-egl = { workspace = true, features = ["static", "no-pkg-config"] }
+egl = { package = "khronos-egl", version = "4.1", features = ["static", "no-pkg-config"] }
 #Note: it's unused by emscripten, but we keep it to have single code base in egl.rs
-libloading = { workspace = true, optional = true }
+libloading = { version = "0.7", optional = true }
 
 [target.'cfg(windows)'.dependencies]
-winapi = { workspace = true, features = ["profileapi", "libloaderapi", "windef", "winuser", "dcomp"] }
-native = { workspace = true, features = ["libloading"], optional = true }
+winapi = { version = "0.3", features = ["profileapi", "libloaderapi", "windef", "winuser", "dcomp"] }
+native = { package = "d3d12", version = "0.5.0", features = ["libloading"], optional = true }
 
 [target.'cfg(any(target_os="macos", target_os="ios"))'.dependencies]
-mtl.workspace = true
-objc.workspace = true
-core-graphics-types.workspace = true
+mtl = { package = "metal", version = "0.24.0" }
+objc = "0.2.5"
+core-graphics-types = "0.1"
 
 [target.'cfg(all(target_arch = "wasm32", not(target_os = "emscripten")))'.dependencies]
-wasm-bindgen.workspace = true
-web-sys = { workspace = true, features = ["Window", "HtmlCanvasElement", "WebGl2RenderingContext", "OffscreenCanvas"] }
-js-sys.workspace = true
+wasm-bindgen = "0.2.83"
+web-sys = { version = "0.3.60", features = ["Window", "HtmlCanvasElement", "WebGl2RenderingContext", "OffscreenCanvas"] }
+js-sys = "0.3.60"
 
 [target.'cfg(unix)'.dependencies]
-libc.workspace = true
+libc = "0.2"
 
 [target.'cfg(target_os = "android")'.dependencies]
-android_system_properties.workspace = true
+android_system_properties = "0.1.1"
 
 [dependencies.naga]
-workspace = true
+git = "https://github.com/gfx-rs/naga"
+rev = "e7fc8e6"
+version = "0.10"
 features = ["clone"]
 
 # DEV dependencies
 [dev-dependencies.naga]
-workspace = true
+git = "https://github.com/gfx-rs/naga"
+rev = "e7fc8e6"
+version = "0.10"
 features = ["wgsl-in"]
 
 [dev-dependencies]
-env_logger.workspace = true
-winit.workspace = true      # for "halmark" example
+env_logger = "0.9"
+winit = "0.27.1"      # for "halmark" example
 
 [target.'cfg(not(target_arch = "wasm32"))'.dev-dependencies]
-glutin.workspace = true # for "gles" example
+glutin = "0.29.1" # for "gles" example
diff --git a/wgpu-hal/examples/halmark/shader.wgsl b/wgpu-hal/examples/halmark/shader.wgsl
index 60aac686b8..ffa7264591 100644
--- a/wgpu-hal/examples/halmark/shader.wgsl
+++ b/wgpu-hal/examples/halmark/shader.wgsl
@@ -1,12 +1,17 @@
 struct Globals {
     mvp: mat4x4<f32>,
     size: vec2<f32>,
+    _pad0: u32,
+    _pad1: u32,
 };
 
 struct Locals {
     position: vec2<f32>,
     velocity: vec2<f32>,
     color: u32,
+    _pad0: u32,
+    _pad1: u32,
+    _pad2: u32,
 };
 
 @group(0)
diff --git a/wgpu-hal/src/dx11/adapter.rs b/wgpu-hal/src/dx11/adapter.rs
index a040e087cb..19de44ac8e 100644
--- a/wgpu-hal/src/dx11/adapter.rs
+++ b/wgpu-hal/src/dx11/adapter.rs
@@ -127,6 +127,7 @@ impl super::Adapter {
             features |= wgt::Features::DEPTH_CLIP_CONTROL;
             features |= wgt::Features::TIMESTAMP_QUERY;
             features |= wgt::Features::PIPELINE_STATISTICS_QUERY;
+            features |= wgt::Features::SHADER_PRIMITIVE_INDEX;
         }
 
         if feature_level >= FL10_1 {
diff --git a/wgpu-hal/src/dx12/adapter.rs b/wgpu-hal/src/dx12/adapter.rs
index 57b238c9f8..2119ebd522 100644
--- a/wgpu-hal/src/dx12/adapter.rs
+++ b/wgpu-hal/src/dx12/adapter.rs
@@ -209,7 +209,8 @@ impl super::Adapter {
             | wgt::Features::TEXTURE_COMPRESSION_BC
             | wgt::Features::CLEAR_TEXTURE
             | wgt::Features::TEXTURE_FORMAT_16BIT_NORM
-            | wgt::Features::PUSH_CONSTANTS;
+            | wgt::Features::PUSH_CONSTANTS
+            | wgt::Features::SHADER_PRIMITIVE_INDEX;
         //TODO: in order to expose this, we need to run a compute shader
         // that extract the necessary statistics out of the D3D12 result.
         // Alternatively, we could allocate a buffer for the query set,
diff --git a/wgpu-hal/src/gles/adapter.rs b/wgpu-hal/src/gles/adapter.rs
index aa7d1f002c..3794909ea5 100644
--- a/wgpu-hal/src/gles/adapter.rs
+++ b/wgpu-hal/src/gles/adapter.rs
@@ -338,6 +338,10 @@ impl super::Adapter {
             wgt::Features::MULTIVIEW,
             extensions.contains("OVR_multiview2"),
         );
+        features.set(
+            wgt::Features::SHADER_PRIMITIVE_INDEX,
+            ver >= (3, 2) || extensions.contains("OES_geometry_shader"),
+        );
         let gles_bcn_exts = [
             "GL_EXT_texture_compression_s3tc_srgb",
             "GL_EXT_texture_compression_rgtc",
@@ -421,6 +425,10 @@ impl super::Adapter {
             super::PrivateCapabilities::COLOR_BUFFER_FLOAT,
             color_buffer_float,
         );
+        private_caps.set(
+            super::PrivateCapabilities::TEXTURE_FLOAT_LINEAR,
+            extensions.contains("OES_texture_float_linear"),
+        );
 
         let max_texture_size = unsafe { gl.get_parameter_i32(glow::MAX_TEXTURE_SIZE) } as u32;
         let max_texture_3d_size = unsafe { gl.get_parameter_i32(glow::MAX_3D_TEXTURE_SIZE) } as u32;
@@ -730,6 +738,9 @@ impl crate::Adapter<super::Api> for super::Adapter {
                 | Tfc::MULTISAMPLE_RESOLVE,
         );
 
+        let texture_float_linear =
+            private_caps_fn(super::PrivateCapabilities::TEXTURE_FLOAT_LINEAR, filterable);
+
         match format {
             Tf::R8Unorm => filterable_renderable,
             Tf::R8Snorm => filterable,
@@ -746,7 +757,7 @@ impl crate::Adapter<super::Api> for super::Adapter {
             Tf::Rg8Sint => renderable,
             Tf::R32Uint => renderable | storage,
             Tf::R32Sint => renderable | storage,
-            Tf::R32Float => unfilterable | storage | float_renderable,
+            Tf::R32Float => unfilterable | storage | float_renderable | texture_float_linear,
             Tf::Rg16Uint => renderable,
             Tf::Rg16Sint => renderable,
             Tf::Rg16Unorm => empty,
@@ -761,7 +772,7 @@ impl crate::Adapter<super::Api> for super::Adapter {
             Tf::Rg11b10Float => filterable | float_renderable,
             Tf::Rg32Uint => renderable,
             Tf::Rg32Sint => renderable,
-            Tf::Rg32Float => unfilterable | float_renderable,
+            Tf::Rg32Float => unfilterable | float_renderable | texture_float_linear,
             Tf::Rgba16Uint => renderable | storage,
             Tf::Rgba16Sint => renderable | storage,
             Tf::Rgba16Unorm => empty,
@@ -769,7 +780,7 @@ impl crate::Adapter<super::Api> for super::Adapter {
             Tf::Rgba16Float => filterable | storage | half_float_renderable,
             Tf::Rgba32Uint => renderable | storage,
             Tf::Rgba32Sint => renderable | storage,
-            Tf::Rgba32Float => unfilterable | storage | float_renderable,
+            Tf::Rgba32Float => unfilterable | storage | float_renderable | texture_float_linear,
             //Tf::Stencil8 |
             Tf::Depth16Unorm
             | Tf::Depth32Float
diff --git a/wgpu-hal/src/gles/mod.rs b/wgpu-hal/src/gles/mod.rs
index bad5c08b91..e57b05a979 100644
--- a/wgpu-hal/src/gles/mod.rs
+++ b/wgpu-hal/src/gles/mod.rs
@@ -148,6 +148,8 @@ bitflags::bitflags! {
         const COLOR_BUFFER_HALF_FLOAT = 1 << 8;
         /// Supports `f11/f10` and `f32` color buffers
         const COLOR_BUFFER_FLOAT = 1 << 9;
+        /// Supports linear flitering `f32` textures.
+        const TEXTURE_FLOAT_LINEAR = 1 << 10;
     }
 }
 
diff --git a/wgpu-hal/src/gles/web.rs b/wgpu-hal/src/gles/web.rs
index b9e7302182..091c494ddc 100644
--- a/wgpu-hal/src/gles/web.rs
+++ b/wgpu-hal/src/gles/web.rs
@@ -33,32 +33,53 @@ impl Instance {
         &self,
         canvas: &web_sys::HtmlCanvasElement,
     ) -> Result<Surface, crate::InstanceError> {
-        let webgl2_context = canvas
-            .get_context_with_context_options("webgl2", &Self::create_context_options())
-            .expect("Cannot create WebGL2 context")
-            .and_then(|context| context.dyn_into::<web_sys::WebGl2RenderingContext>().ok())
-            .expect("Cannot convert into WebGL2 context");
-
-        *self.webgl2_context.lock() = Some(webgl2_context.clone());
-
-        Ok(Surface {
-            webgl2_context,
-            srgb_present_program: None,
-            swapchain: None,
-            texture: None,
-            presentable: true,
-        })
+        self.create_surface_from_context(
+            canvas.get_context_with_context_options("webgl2", &Self::create_context_options()),
+        )
     }
 
     pub fn create_surface_from_offscreen_canvas(
         &self,
         canvas: &web_sys::OffscreenCanvas,
     ) -> Result<Surface, crate::InstanceError> {
-        let webgl2_context = canvas
-            .get_context_with_context_options("webgl2", &Self::create_context_options())
-            .expect("Cannot create WebGL2 context")
-            .and_then(|context| context.dyn_into::<web_sys::WebGl2RenderingContext>().ok())
-            .expect("Cannot convert into WebGL2 context");
+        self.create_surface_from_context(
+            canvas.get_context_with_context_options("webgl2", &Self::create_context_options()),
+        )
+    }
+
+    /// Common portion of public `create_surface_from_*` functions.
+    ///
+    /// Note: Analogous code also exists in the WebGPU backend at
+    /// `wgpu::backend::web::Context`.
+    fn create_surface_from_context(
+        &self,
+        context_result: Result<Option<js_sys::Object>, wasm_bindgen::JsValue>,
+    ) -> Result<Surface, crate::InstanceError> {
+        let context_object: js_sys::Object = match context_result {
+            Ok(Some(context)) => context,
+            Ok(None) => {
+                // <https://html.spec.whatwg.org/multipage/canvas.html#dom-canvas-getcontext-dev>
+                // A getContext() call “returns null if contextId is not supported, or if the
+                // canvas has already been initialized with another context type”. Additionally,
+                // “not supported” could include “insufficient GPU resources” or “the GPU process
+                // previously crashed”. So, we must return it as an `Err` since it could occur
+                // for circumstances outside the application author's control.
+                return Err(crate::InstanceError);
+            }
+            Err(js_error) => {
+                // <https://html.spec.whatwg.org/multipage/canvas.html#dom-canvas-getcontext>
+                // A thrown exception indicates misuse of the canvas state. Ideally we wouldn't
+                // panic in this case, but for now, `InstanceError` conveys no detail, so it
+                // is more informative to panic with a specific message.
+                panic!("canvas.getContext() threw {js_error:?}")
+            }
+        };
+
+        // Not returning this error because it is a type error that shouldn't happen unless
+        // the browser, JS builtin objects, or wasm bindings are misbehaving somehow.
+        let webgl2_context: web_sys::WebGl2RenderingContext = context_object
+            .dyn_into()
+            .expect("canvas context is not a WebGl2RenderingContext");
 
         *self.webgl2_context.lock() = Some(webgl2_context.clone());
 
diff --git a/wgpu-hal/src/lib.rs b/wgpu-hal/src/lib.rs
index f9ce21e122..9a82b54a25 100644
--- a/wgpu-hal/src/lib.rs
+++ b/wgpu-hal/src/lib.rs
@@ -49,6 +49,15 @@
     clippy::pattern_type_mismatch,
 )]
 
+#[cfg(not(any(
+    feature = "dx11",
+    feature = "dx12",
+    feature = "gles",
+    feature = "metal",
+    feature = "vulkan"
+)))]
+compile_error!("No back ends enabled in `wgpu-hal`. Enable at least one backend feature.");
+
 #[cfg(all(feature = "metal", not(any(target_os = "macos", target_os = "ios"))))]
 compile_error!("Metal API enabled on non-Apple OS. If your project is not using resolver=\"2\" in Cargo.toml, it should.");
 #[cfg(all(feature = "dx12", not(windows)))]
diff --git a/wgpu-hal/src/metal/adapter.rs b/wgpu-hal/src/metal/adapter.rs
index 63264a70d8..5ade092c24 100644
--- a/wgpu-hal/src/metal/adapter.rs
+++ b/wgpu-hal/src/metal/adapter.rs
@@ -732,6 +732,8 @@ impl super::PrivateCapabilities {
             supports_depth_clip_control: os_is_mac
                 || device.supports_feature_set(MTLFeatureSet::iOS_GPUFamily4_v1),
             supports_preserve_invariance: version.at_least((11, 0), (13, 0)),
+            // Metal 2.2 on mac, 2.3 on iOS.
+            supports_shader_primitive_index: version.at_least((10, 15), (14, 0)),
             has_unified_memory: if version.at_least((10, 15), (13, 0)) {
                 Some(device.has_unified_memory())
             } else {
@@ -770,6 +772,10 @@ impl super::PrivateCapabilities {
         features.set(F::TEXTURE_COMPRESSION_ETC2, self.format_eac_etc);
 
         features.set(F::DEPTH_CLIP_CONTROL, self.supports_depth_clip_control);
+        features.set(
+            F::SHADER_PRIMITIVE_INDEX,
+            self.supports_shader_primitive_index,
+        );
 
         features.set(
             F::TEXTURE_BINDING_ARRAY
diff --git a/wgpu-hal/src/metal/mod.rs b/wgpu-hal/src/metal/mod.rs
index c0b0b5d320..37f101cff7 100644
--- a/wgpu-hal/src/metal/mod.rs
+++ b/wgpu-hal/src/metal/mod.rs
@@ -231,6 +231,7 @@ struct PrivateCapabilities {
     supports_mutability: bool,
     supports_depth_clip_control: bool,
     supports_preserve_invariance: bool,
+    supports_shader_primitive_index: bool,
     has_unified_memory: Option<bool>,
 }
 
diff --git a/wgpu-hal/src/vulkan/device.rs b/wgpu-hal/src/vulkan/device.rs
index 084be72de9..dafd15f500 100644
--- a/wgpu-hal/src/vulkan/device.rs
+++ b/wgpu-hal/src/vulkan/device.rs
@@ -52,7 +52,7 @@ impl super::DeviceShared {
         let name = unsafe { CStr::from_bytes_with_nul_unchecked(name_bytes) };
 
         let _result = unsafe {
-            extension.debug_utils_set_object_name(
+            extension.set_debug_utils_object_name(
                 self.raw.handle(),
                 &vk::DebugUtilsObjectNameInfoEXT::builder()
                     .object_type(object_type)
diff --git a/wgpu-types/Cargo.toml b/wgpu-types/Cargo.toml
index 0748296a81..4b608b472f 100644
--- a/wgpu-types/Cargo.toml
+++ b/wgpu-types/Cargo.toml
@@ -1,13 +1,13 @@
 [package]
 name = "wgpu-types"
-version.workspace = true
-authors.workspace = true
-edition.workspace = true
+version = "0.14.0"
+authors = ["wgpu developers"]
+edition = "2021"
 description = "WebGPU types"
-homepage.workspace = true
-repository.workspace = true
-keywords.workspace = true
-license.workspace = true
+homepage = "https://wgpu.rs/"
+repository = "https://github.com/gfx-rs/wgpu"
+keywords = ["graphics"]
+license = "MIT OR Apache-2.0"
 
 [package.metadata.docs.rs]
 all-features = true
@@ -20,8 +20,8 @@ trace = ["serde"]
 replay = ["serde"]
 
 [dependencies]
-bitflags.workspace = true
-serde = { workspace = true, features = ["serde_derive"], optional = true }
+bitflags = "1"
+serde = { version = "1", features = ["serde_derive"], optional = true }
 
 [dev-dependencies]
-serde_json.workspace = true
+serde_json = "1.0.85"
diff --git a/wgpu-types/src/lib.rs b/wgpu-types/src/lib.rs
index f333d2d603..573a5771fd 100644
--- a/wgpu-types/src/lib.rs
+++ b/wgpu-types/src/lib.rs
@@ -661,7 +661,7 @@ bitflags::bitflags! {
         ///
         /// This is a native-only feature.
         const TEXTURE_COMPRESSION_ASTC_HDR = 1 << 40;
-        /// Allows for timestamp queries inside renderpasses. Metal does not allow this
+        /// Allows for timestamp queries inside render passes. Metal does not allow this
         /// on Apple GPUs.
         ///
         /// Implies [`Features::TIMESTAMP_QUERIES`] is supported.
@@ -698,9 +698,9 @@ impl Features {
 /// - [`Limits::downlevel_webgl2_defaults()`] This is a set of limits that is lower even than the
 ///   [`downlevel_defaults()`], configured to be low enough to support running in the browser using
 ///   WebGL2.
-/// - [`Limits::default()`]. This is the set of limits that is guarenteed to work on all modern
-///   backends and is guarenteed to be supported by WebGPU. Applications needing more modern
-///   features can use this as a reasonable set of limits if they are targetting only desktop and
+/// - [`Limits::default()`]. This is the set of limits that is guaranteed to work on all modern
+///   backends and is guaranteed to be supported by WebGPU. Applications needing more modern
+///   features can use this as a reasonable set of limits if they are targeting only desktop and
 ///   modern mobile devices.
 ///
 /// We recommend starting with the most restrictive limits you can and manually increasing the
@@ -813,7 +813,7 @@ pub struct Limits {
     pub max_compute_workgroups_per_dimension: u32,
     /// A limit above which buffer allocations are guaranteed to fail.
     ///
-    /// Buffer allocations below the maximum buffer size may not succed depending on available memory,
+    /// Buffer allocations below the maximum buffer size may not succeed depending on available memory,
     /// fragmentation and other factors.
     pub max_buffer_size: u64,
 }
@@ -849,13 +849,13 @@ impl Default for Limits {
             max_compute_workgroup_size_y: 256,
             max_compute_workgroup_size_z: 64,
             max_compute_workgroups_per_dimension: 65535,
-            max_buffer_size: 1 << 30,
+            max_buffer_size: 1 << 28,
         }
     }
 }
 
 impl Limits {
-    /// These default limits are guarenteed to be compatible with GLES-3.1, and D3D11
+    /// These default limits are guaranteed to be compatible with GLES-3.1, and D3D11
     pub fn downlevel_defaults() -> Self {
         Self {
             max_texture_dimension_1d: 2048,
@@ -890,7 +890,7 @@ impl Limits {
         }
     }
 
-    /// These default limits are guarenteed to be compatible with GLES-3.0, and D3D11, and WebGL2
+    /// These default limits are guaranteed to be compatible with GLES-3.0, and D3D11, and WebGL2
     pub fn downlevel_webgl2_defaults() -> Self {
         Self {
             max_uniform_buffers_per_shader_stage: 11,
@@ -1218,7 +1218,7 @@ pub struct DeviceDescriptor<L> {
 }
 
 impl<L> DeviceDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the device descriptor into another.
     pub fn map_label<K>(&self, fun: impl FnOnce(&L) -> K) -> DeviceDescriptor<K> {
         DeviceDescriptor {
             label: fun(&self.label),
@@ -2086,7 +2086,7 @@ pub enum TextureFormat {
     Astc {
         /// compressed block dimensions
         block: AstcBlock,
-        ///
+        /// ASTC RGBA channel
         channel: AstcChannel,
     },
 }
@@ -3707,7 +3707,7 @@ pub struct BufferDescriptor<L> {
 }
 
 impl<L> BufferDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the buffer descriptor into another.
     pub fn map_label<K>(&self, fun: impl FnOnce(&L) -> K) -> BufferDescriptor<K> {
         BufferDescriptor {
             label: fun(&self.label),
@@ -3732,7 +3732,7 @@ pub struct CommandEncoderDescriptor<L> {
 }
 
 impl<L> CommandEncoderDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the command encoder descriptor into another.
     pub fn map_label<K>(&self, fun: impl FnOnce(&L) -> K) -> CommandEncoderDescriptor<K> {
         CommandEncoderDescriptor {
             label: fun(&self.label),
@@ -3876,7 +3876,7 @@ bitflags::bitflags! {
         const TEXTURE_BINDING = 1 << 2;
         /// Allows a texture to be a [`BindingType::StorageTexture`] in a bind group.
         const STORAGE_BINDING = 1 << 3;
-        /// Allows a texture to be an output attachment of a renderpass.
+        /// Allows a texture to be an output attachment of a render pass.
         const RENDER_ATTACHMENT = 1 << 4;
     }
 }
@@ -3998,13 +3998,13 @@ impl PresentationTimestamp {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "serde", serde(rename_all = "camelCase"))]
 pub struct Color {
-    ///
+    /// Red component of the color
     pub r: f64,
-    ///
+    /// Green component of the color
     pub g: f64,
-    ///
+    /// Blue component of the color
     pub b: f64,
-    ///
+    /// Alpha component of the color
     pub a: f64,
 }
 
@@ -4078,11 +4078,11 @@ pub enum TextureDimension {
 #[cfg_attr(feature = "replay", derive(Deserialize))]
 #[cfg_attr(feature = "serde", serde(rename_all = "camelCase"))]
 pub struct Origin3d {
-    ///
+    /// X position of the origin
     pub x: u32,
-    ///
+    /// Y position of the origin
     pub y: u32,
-    ///
+    /// Z position of the origin
     pub z: u32,
 }
 
@@ -4107,11 +4107,11 @@ impl Default for Origin3d {
 #[cfg_attr(feature = "replay", derive(Deserialize))]
 #[cfg_attr(feature = "serde", serde(rename_all = "camelCase"))]
 pub struct Extent3d {
-    ///
+    /// Width of the extent
     pub width: u32,
-    ///
+    /// Height of the extent
     pub height: u32,
-    ///
+    /// The depth of the extent or the number of array layers
     #[cfg_attr(feature = "serde", serde(default = "default_depth"))]
     pub depth_or_array_layers: u32,
 }
@@ -4316,7 +4316,7 @@ pub struct TextureDescriptor<L> {
 }
 
 impl<L> TextureDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the texture descriptor into another.
     pub fn map_label<K>(&self, fun: impl FnOnce(&L) -> K) -> TextureDescriptor<K> {
         TextureDescriptor {
             label: fun(&self.label),
@@ -4478,7 +4478,7 @@ pub struct CommandBufferDescriptor<L> {
 }
 
 impl<L> CommandBufferDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the command buffer descriptor into another.
     pub fn map_label<K>(&self, fun: impl FnOnce(&L) -> K) -> CommandBufferDescriptor<K> {
         CommandBufferDescriptor {
             label: fun(&self.label),
@@ -4517,7 +4517,7 @@ pub struct RenderBundleDescriptor<L> {
 }
 
 impl<L> RenderBundleDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the render bundle descriptor into another.
     pub fn map_label<K>(&self, fun: impl FnOnce(&L) -> K) -> RenderBundleDescriptor<K> {
         RenderBundleDescriptor {
             label: fun(&self.label),
@@ -5069,7 +5069,7 @@ pub struct QuerySetDescriptor<L> {
 }
 
 impl<L> QuerySetDescriptor<L> {
-    ///
+    /// Takes a closure and maps the label of the query set descriptor into another.
     pub fn map_label<'a, K>(&'a self, fun: impl FnOnce(&'a L) -> K) -> QuerySetDescriptor<K> {
         QuerySetDescriptor {
             label: fun(&self.label),
diff --git a/wgpu/Cargo.toml b/wgpu/Cargo.toml
index 36272c0fda..bb446b7d8e 100644
--- a/wgpu/Cargo.toml
+++ b/wgpu/Cargo.toml
@@ -83,25 +83,62 @@ wgsl = ["wgc?/wgsl"]
 trace = ["serde", "wgc/trace"]
 replay = ["serde", "wgc/replay"]
 angle = ["wgc/angle"]
-webgl = ["wgc"]
+webgl = ["hal", "wgc"]
 emscripten = ["webgl"]
 vulkan-portability = ["wgc/vulkan-portability"]
 expose-ids = []
 
-[target.'cfg(not(target_arch = "wasm32"))'.dependencies.wgc]
+# wgpu-core is always available as an optional dependency, "wgc".
+# Whenever wgpu-core is selected, we want the GLES backend and raw
+# window handle support.
+[dependencies.wgc]
+optional = true
 workspace = true
-features = ["raw-window-handle"]
+features = ["raw-window-handle", "gles"]
 
-[target.'cfg(target_arch = "wasm32")'.dependencies.wgc]
+# wgpu-core is required whenever not targeting web APIs directly.
+# Whenever wgpu-core is selected, we want the GLES backend and raw
+# window handle support.
+[target.'cfg(any(not(target_arch = "wasm32"), target_os = "emscripten"))'.dependencies.wgc]
 workspace = true
-features = ["raw-window-handle"]
-optional = true
+features = ["raw-window-handle", "gles"]
+
+# We want the wgpu-core Metal backend on macOS and iOS.
+# (We should consider also enabling "vulkan" for Vulkan Portability.)
+[target.'cfg(any(target_os = "macos", target_os = "ios"))'.dependencies.wgc]
+workspace = true
+features = ["metal"]
+
+# We want the wgpu-core Direct3D backends on Windows.
+[target.'cfg(windows)'.dependencies.wgc]
+workspace = true
+features = ["dx11", "dx12"]
+
+# We want the wgpu-core Vulkan backend on Unix (but not Emscripten) and Windows.
+[target.'cfg(any(windows, all(unix, not(target_arch = "emscripten"))))'.dependencies.wgc]
+workspace = true
+features = ["vulkan"]
+
+[target.'cfg(target_os = "emscripten")'.dependencies.wgc]
+workspace = true
+features = ["emscripten"]
 
 [dependencies.wgt]
 workspace = true
 
-[target.'cfg(any(not(target_arch = "wasm32"), target_os = "emscripten"))'.dependencies.hal]
+# We need wgpu-hal unless we're targeting the web APIs.
+[target.'cfg(any(not(target_arch = "wasm32"), target_os = "emscripten"))'.dependencies]
+hal = { workspace = true }
+
+[target.'cfg(all(not(target_arch = "wasm32"), unix, not(target_os = "ios"), not(target_os = "macos")))'.dependencies]
+hal = { workspace = true, features = ["renderdoc"] }
+
+[target.'cfg(windows)'.dependencies]
+hal = { workspace = true, features = ["renderdoc"] }
+
+[target.'cfg(target_arch = "wasm32")'.dependencies.hal]
 workspace = true
+optional = true
 
 [dependencies]
 arrayvec.workspace = true
@@ -126,7 +163,8 @@ obj.workspace = true
 pollster.workspace = true
 png.workspace = true
 nanorand = { workspace = true, features = ["wyrand"] }
-winit.workspace = true                                 # for "halmark" example                                                        # for "halmark" example
+wasm-bindgen-test.workspace = true
+winit.workspace = true # for "halmark" example
 
 [target.'cfg(not(target_arch = "wasm32"))'.dev-dependencies]
 async-executor.workspace = true
@@ -283,12 +321,13 @@ parking_lot.workspace = true
 [target.'cfg(target_arch = "wasm32")'.dev-dependencies]
 console_error_panic_hook.workspace = true
 console_log.workspace = true
-# We need these features in the framework examples
+# We need these features in the framework examples and tests
 web-sys = { workspace = true, features = [
     "Location",
     "Blob",
     "RequestInit",
     "RequestMode",
     "Request",
-    "Response"
+    "Response",
+    "WebGl2RenderingContext"
 ] }
diff --git a/wgpu/examples/framework.rs b/wgpu/examples/framework.rs
index 122c88170d..9af044bf66 100644
--- a/wgpu/examples/framework.rs
+++ b/wgpu/examples/framework.rs
@@ -164,7 +164,7 @@ async fn setup<E: Example>(title: &str) -> Setup {
         let size = window.inner_size();
 
         #[cfg(any(not(target_arch = "wasm32"), target_os = "emscripten"))]
-        let surface = instance.create_surface(&window);
+        let surface = instance.create_surface(&window).unwrap();
         #[cfg(all(target_arch = "wasm32", not(target_os = "emscripten")))]
         let surface = {
             if let Some(offscreen_canvas_setup) = &offscreen_canvas_setup {
@@ -174,7 +174,8 @@ async fn setup<E: Example>(title: &str) -> Setup {
             } else {
                 instance.create_surface(&window)
             }
-        };
+        }
+        .unwrap();
 
         (size, surface)
     };
diff --git a/wgpu/examples/hello-triangle/main.rs b/wgpu/examples/hello-triangle/main.rs
index 4c19babf3c..58e7ba1a90 100644
--- a/wgpu/examples/hello-triangle/main.rs
+++ b/wgpu/examples/hello-triangle/main.rs
@@ -8,7 +8,7 @@ use winit::{
 async fn run(event_loop: EventLoop<()>, window: Window) {
     let size = window.inner_size();
     let instance = wgpu::Instance::new(wgpu::Backends::all());
-    let surface = unsafe { instance.create_surface(&window) };
+    let surface = unsafe { instance.create_surface(&window) }.unwrap();
     let adapter = instance
         .request_adapter(&wgpu::RequestAdapterOptions {
             power_preference: wgpu::PowerPreference::default(),
diff --git a/wgpu/examples/hello-windows/main.rs b/wgpu/examples/hello-windows/main.rs
index 0dcc87daac..4a93cf6b7f 100644
--- a/wgpu/examples/hello-windows/main.rs
+++ b/wgpu/examples/hello-windows/main.rs
@@ -20,7 +20,7 @@ struct Viewport {
 
 impl ViewportDesc {
     fn new(window: Window, background: wgpu::Color, instance: &wgpu::Instance) -> Self {
-        let surface = unsafe { instance.create_surface(&window) };
+        let surface = unsafe { instance.create_surface(&window) }.unwrap();
         Self {
             window,
             background,
diff --git a/wgpu/examples/mipmap/blit.wgsl b/wgpu/examples/mipmap/blit.wgsl
index c293e51c73..69adbb3ccd 100644
--- a/wgpu/examples/mipmap/blit.wgsl
+++ b/wgpu/examples/mipmap/blit.wgsl
@@ -3,6 +3,24 @@ struct VertexOutput {
     @location(0) tex_coords: vec2<f32>,
 };
 
+// meant to be called with 3 vertex indices: 0, 1, 2
+// draws one large triangle over the clip space like this:
+// (the asterisks represent the clip space bounds)
+//-1,1           1,1
+// ---------------------------------
+// |              *              .
+// |              *           .
+// |              *        .
+// |              *      .
+// |              *    . 
+// |              * .
+// |***************
+// |            . 1,-1 
+// |          .
+// |       .
+// |     .
+// |   .
+// |.
 @vertex
 fn vs_main(@builtin(vertex_index) vertex_index: u32) -> VertexOutput {
     var result: VertexOutput;
diff --git a/wgpu/src/backend/direct.rs b/wgpu/src/backend/direct.rs
index f6ce4e343c..37506bf19a 100644
--- a/wgpu/src/backend/direct.rs
+++ b/wgpu/src/backend/direct.rs
@@ -219,24 +219,30 @@ impl Context {
     pub fn instance_create_surface_from_canvas(
         self: &Arc<Self>,
         canvas: &web_sys::HtmlCanvasElement,
-    ) -> Surface {
-        let id = self.0.create_surface_webgl_canvas(canvas, ());
-        Surface {
+    ) -> Result<Surface, crate::CreateSurfaceError> {
+        let id = self
+            .0
+            .create_surface_webgl_canvas(canvas, ())
+            .map_err(|hal::InstanceError| crate::CreateSurfaceError {})?;
+        Ok(Surface {
             id,
             configured_device: Mutex::default(),
-        }
+        })
     }
 
     #[cfg(all(target_arch = "wasm32", feature = "webgl", not(feature = "emscripten")))]
     pub fn instance_create_surface_from_offscreen_canvas(
         self: &Arc<Self>,
         canvas: &web_sys::OffscreenCanvas,
-    ) -> Surface {
-        let id = self.0.create_surface_webgl_offscreen_canvas(canvas, ());
-        Surface {
+    ) -> Result<Surface, crate::CreateSurfaceError> {
+        let id = self
+            .0
+            .create_surface_webgl_offscreen_canvas(canvas, ())
+            .map_err(|hal::InstanceError| crate::CreateSurfaceError {})?;
+        Ok(Surface {
             id,
             configured_device: Mutex::default(),
-        }
+        })
     }
 
     #[cfg(target_os = "windows")]
@@ -943,13 +949,13 @@ impl crate::Context for Context {
         &self,
         display_handle: raw_window_handle::RawDisplayHandle,
         window_handle: raw_window_handle::RawWindowHandle,
-    ) -> Self::SurfaceId {
-        Surface {
+    ) -> Result<Self::SurfaceId, crate::CreateSurfaceError> {
+        Ok(Surface {
             id: self
                 .0
                 .instance_create_surface(display_handle, window_handle, ()),
             configured_device: Mutex::new(None),
-        }
+        })
     }
 
     fn instance_request_adapter(
diff --git a/wgpu/src/backend/web.rs b/wgpu/src/backend/web.rs
index 93b4a65370..19daf02785 100644
--- a/wgpu/src/backend/web.rs
+++ b/wgpu/src/backend/web.rs
@@ -1026,23 +1026,51 @@ impl Context {
     pub fn instance_create_surface_from_canvas(
         &self,
         canvas: &web_sys::HtmlCanvasElement,
-    ) -> <Self as crate::Context>::SurfaceId {
-        let context: wasm_bindgen::JsValue = match canvas.get_context("webgpu") {
-            Ok(Some(ctx)) => ctx.into(),
-            _ => panic!("expected to get context from canvas"),
-        };
-        create_identified(context.into())
+    ) -> Result<<Self as crate::Context>::SurfaceId, crate::CreateSurfaceError> {
+        self.create_surface_from_context(canvas.get_context("webgpu"))
     }
 
     pub fn instance_create_surface_from_offscreen_canvas(
         &self,
         canvas: &web_sys::OffscreenCanvas,
-    ) -> <Self as crate::Context>::SurfaceId {
-        let context: wasm_bindgen::JsValue = match canvas.get_context("webgpu") {
-            Ok(Some(ctx)) => ctx.into(),
-            _ => panic!("expected to get context from canvas"),
+    ) -> Result<<Self as crate::Context>::SurfaceId, crate::CreateSurfaceError> {
+        self.create_surface_from_context(canvas.get_context("webgpu"))
+    }
+
+    /// Common portion of public `instance_create_surface_from_*` functions.
+    ///
+    /// Note: Analogous code also exists in the WebGL2 backend at
+    /// `wgpu_hal::gles::web::Instance`.
+    fn create_surface_from_context(
+        &self,
+        context_result: Result<Option<js_sys::Object>, wasm_bindgen::JsValue>,
+    ) -> Result<<Self as crate::Context>::SurfaceId, crate::CreateSurfaceError> {
+        let context: js_sys::Object = match context_result {
+            Ok(Some(context)) => context,
+            Ok(None) => {
+                // <https://html.spec.whatwg.org/multipage/canvas.html#dom-canvas-getcontext-dev>
+                // A getContext() call “returns null if contextId is not supported, or if the
+                // canvas has already been initialized with another context type”. Additionally,
+                // “not supported” could include “insufficient GPU resources” or “the GPU process
+                // previously crashed”. So, we must return it as an `Err` since it could occur
+                // for circumstances outside the application author's control.
+                return Err(crate::CreateSurfaceError {});
+            }
+            Err(js_error) => {
+                // <https://html.spec.whatwg.org/multipage/canvas.html#dom-canvas-getcontext>
+                // A thrown exception indicates misuse of the canvas state. Ideally we wouldn't
+                // panic in this case ... TODO
+                panic!("canvas.getContext() threw {js_error:?}")
+            }
         };
-        create_identified(context.into())
+
+        // Not returning this error because it is a type error that shouldn't happen unless
+        // the browser, JS builtin objects, or wasm bindings are misbehaving somehow.
+        let context: web_sys::GpuCanvasContext = context
+            .dyn_into()
+            .expect("canvas context is not a GPUCanvasContext");
+
+        Ok(create_identified(context))
     }
 
     pub fn queue_copy_external_image_to_texture(
@@ -1141,7 +1169,7 @@ impl crate::Context for Context {
         &self,
         _display_handle: raw_window_handle::RawDisplayHandle,
         window_handle: raw_window_handle::RawWindowHandle,
-    ) -> Self::SurfaceId {
+    ) -> Result<Self::SurfaceId, crate::CreateSurfaceError> {
         let canvas_attribute = match window_handle {
             raw_window_handle::RawWindowHandle::Web(web_handle) => web_handle.id,
             _ => panic!("expected valid handle for canvas"),
diff --git a/wgpu/src/lib.rs b/wgpu/src/lib.rs
index fb1cb6987e..066635ff1f 100644
--- a/wgpu/src/lib.rs
+++ b/wgpu/src/lib.rs
@@ -205,7 +205,7 @@ trait Context: Debug + Send + Sized + Sync {
         &self,
         display_handle: raw_window_handle::RawDisplayHandle,
         window_handle: raw_window_handle::RawWindowHandle,
-    ) -> Self::SurfaceId;
+    ) -> Result<Self::SurfaceId, crate::CreateSurfaceError>;
     fn instance_request_adapter(
         &self,
         options: &RequestAdapterOptions<'_>,
@@ -1213,7 +1213,7 @@ pub struct BufferBinding<'a> {
 }
 static_assertions::assert_impl_all!(BufferBinding: Send, Sync);
 
-/// Operation to perform to the output attachment at the start of a renderpass.
+/// Operation to perform to the output attachment at the start of a render pass.
 ///
 /// The render target must be cleared at least once before its content is loaded.
 ///
@@ -1369,8 +1369,8 @@ pub struct TextureViewDescriptor<'a> {
     pub label: Label<'a>,
     /// Format of the texture view. At this time, it must be the same as the underlying format of the texture.
     pub format: Option<TextureFormat>,
-    /// The dimension of the texture view. For 1D textures, this must be `1D`. For 2D textures it must be one of
-    /// `D2`, `D2Array`, `Cube`, and `CubeArray`. For 3D textures it must be `3D`
+    /// The dimension of the texture view. For 1D textures, this must be `D1`. For 2D textures it must be one of
+    /// `D2`, `D2Array`, `Cube`, and `CubeArray`. For 3D textures it must be `D3`
     pub dimension: Option<TextureViewDimension>,
     /// Aspect of the texture. Color textures must be [`TextureAspect::All`].
     pub aspect: TextureAspect,
@@ -1674,13 +1674,13 @@ pub struct RenderBundleEncoderDescriptor<'a> {
     /// Debug label of the render bundle encoder. This will show up in graphics debuggers for easy identification.
     pub label: Label<'a>,
     /// The formats of the color attachments that this render bundle is capable to rendering to. This
-    /// must match the formats of the color attachments in the renderpass this render bundle is executed in.
+    /// must match the formats of the color attachments in the render pass this render bundle is executed in.
     pub color_formats: &'a [Option<TextureFormat>],
     /// Information about the depth attachment that this render bundle is capable to rendering to. This
-    /// must match the format of the depth attachments in the renderpass this render bundle is executed in.
+    /// must match the format of the depth attachments in the render pass this render bundle is executed in.
     pub depth_stencil: Option<RenderBundleDepthStencil>,
     /// Sample count this render bundle is capable of rendering to. This must match the pipelines and
-    /// the renderpasses it is used in.
+    /// the render passes it is used in.
     pub sample_count: u32,
     /// If this render bundle will rendering to multiple array layers in the attachments at the same time.
     pub multiview: Option<NonZeroU32>,
@@ -1845,23 +1845,34 @@ impl Instance {
     ///
     /// # Safety
     ///
-    /// - Raw Window Handle must be a valid object to create a surface upon and
-    ///   must remain valid for the lifetime of the returned surface.
-    /// - If not called on the main thread, metal backend will panic.
+    /// - `raw_window_handle` must be a valid object to create a surface upon.
+    /// - `raw_window_handle` must remain valid until after the returned [`Surface`] is
+    ///   dropped.
+    ///
+    /// # Errors
+    ///
+    /// - On WebGL2: Will return an error if the browser does not support WebGL2,
+    ///   or declines to provide GPU access (such as due to a resource shortage).
+    ///
+    /// # Panics
+    ///
+    /// - On macOS/Metal: will panic if not called on the main thread.
+    /// - On web: will panic if the `raw_window_handle` does not properly refer to a
+    ///   canvas element.
     pub unsafe fn create_surface<
         W: raw_window_handle::HasRawWindowHandle + raw_window_handle::HasRawDisplayHandle,
     >(
         &self,
         window: &W,
-    ) -> Surface {
-        Surface {
+    ) -> Result<Surface, CreateSurfaceError> {
+        Ok(Surface {
             context: Arc::clone(&self.context),
             id: Context::instance_create_surface(
                 &*self.context,
                 raw_window_handle::HasRawDisplayHandle::raw_display_handle(window),
                 raw_window_handle::HasRawWindowHandle::raw_window_handle(window),
-            ),
-        }
+            )?,
+        })
     }
 
     /// Creates a surface from `CoreAnimationLayer`.
@@ -1891,29 +1902,42 @@ impl Instance {
     ///
     /// The `canvas` argument must be a valid `<canvas>` element to
     /// create a surface upon.
+    ///
+    /// # Errors
+    ///
+    /// - On WebGL2: Will return an error if the browser does not support WebGL2,
+    ///   or declines to provide GPU access (such as due to a resource shortage).
     #[cfg(all(target_arch = "wasm32", not(feature = "emscripten")))]
-    pub fn create_surface_from_canvas(&self, canvas: &web_sys::HtmlCanvasElement) -> Surface {
-        Surface {
+    pub fn create_surface_from_canvas(
+        &self,
+        canvas: &web_sys::HtmlCanvasElement,
+    ) -> Result<Surface, CreateSurfaceError> {
+        Ok(Surface {
             context: Arc::clone(&self.context),
-            id: self.context.instance_create_surface_from_canvas(canvas),
-        }
+            id: self.context.instance_create_surface_from_canvas(canvas)?,
+        })
     }
 
     /// Creates a surface from a `web_sys::OffscreenCanvas`.
     ///
     /// The `canvas` argument must be a valid `OffscreenCanvas` object
     /// to create a surface upon.
+    ///
+    /// # Errors
+    ///
+    /// - On WebGL2: Will return an error if the browser does not support WebGL2,
+    ///   or declines to provide GPU access (such as due to a resource shortage).
     #[cfg(all(target_arch = "wasm32", not(feature = "emscripten")))]
     pub fn create_surface_from_offscreen_canvas(
         &self,
         canvas: &web_sys::OffscreenCanvas,
-    ) -> Surface {
-        Surface {
+    ) -> Result<Surface, CreateSurfaceError> {
+        Ok(Surface {
             context: Arc::clone(&self.context),
             id: self
                 .context
-                .instance_create_surface_from_offscreen_canvas(canvas),
-        }
+                .instance_create_surface_from_offscreen_canvas(canvas)?,
+        })
     }
 
     /// Polls all devices.
@@ -2407,6 +2431,22 @@ impl Display for RequestDeviceError {
 
 impl error::Error for RequestDeviceError {}
 
+/// [`Instance::create_surface()`] or a related function failed.
+#[derive(Clone, PartialEq, Eq, Debug)]
+#[non_exhaustive]
+pub struct CreateSurfaceError {
+    // TODO: Report diagnostic clues
+}
+static_assertions::assert_impl_all!(CreateSurfaceError: Send, Sync);
+
+impl Display for CreateSurfaceError {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "Creating a surface failed")
+    }
+}
+
+impl error::Error for CreateSurfaceError {}
+
 /// Error occurred when trying to async map a buffer.
 #[derive(Clone, PartialEq, Eq, Debug)]
 pub struct BufferAsyncError;
@@ -3162,7 +3202,7 @@ impl<'a> RenderPass<'a> {
 
 /// [`Features::MULTI_DRAW_INDIRECT_COUNT`] must be enabled on the device in order to call these functions.
 impl<'a> RenderPass<'a> {
-    /// Disptaches multiple draw calls from the active vertex buffer(s) based on the contents of the `indirect_buffer`.
+    /// Dispatches multiple draw calls from the active vertex buffer(s) based on the contents of the `indirect_buffer`.
     /// The count buffer is read to determine how many draws to issue.
     ///
     /// The indirect buffer must be long enough to account for `max_count` draws, however only `count` will
@@ -4055,14 +4095,14 @@ impl<T> UncapturedErrorHandler for T where T: Fn(Error) + Send + 'static {}
 pub enum Error {
     /// Out of memory error
     OutOfMemory {
-        ///
+        /// Lower level source of the error.
         source: Box<dyn error::Error + Send + 'static>,
     },
     /// Validation error, signifying a bug in code or data
     Validation {
-        ///
+        /// Lower level source of the error.
         source: Box<dyn error::Error + Send + 'static>,
-        ///
+        /// Description of the validation error.
         description: String,
     },
 }
diff --git a/wgpu/tests/buffer_usages.rs b/wgpu/tests/buffer_usages.rs
index db734c108b..7eb3f0e832 100644
--- a/wgpu/tests/buffer_usages.rs
+++ b/wgpu/tests/buffer_usages.rs
@@ -1,11 +1,13 @@
 //! Tests for buffer usages validation.
 
 use crate::common::{fail_if, initialize_test, TestParameters};
+use wasm_bindgen_test::*;
 use wgt::BufferAddress;
 
 const BUFFER_SIZE: BufferAddress = 1234;
 
 #[test]
+#[wasm_bindgen_test]
 fn buffer_usage() {
     fn try_create(enable_mappable_primary_buffers: bool, usages: &[(bool, &[wgpu::BufferUsages])]) {
         let mut parameters = TestParameters::default();
diff --git a/wgpu/tests/clear_texture.rs b/wgpu/tests/clear_texture.rs
index abe86cac43..0e8a75f3b0 100644
--- a/wgpu/tests/clear_texture.rs
+++ b/wgpu/tests/clear_texture.rs
@@ -1,4 +1,5 @@
 use crate::common::{initialize_test, TestParameters, TestingContext};
+use wasm_bindgen_test::*;
 use wgpu::util::align_to;
 
 static TEXTURE_FORMATS_UNCOMPRESSED: &[wgpu::TextureFormat] = &[
@@ -202,9 +203,11 @@ fn single_texture_clear_test(
     size: wgpu::Extent3d,
     dimension: wgpu::TextureDimension,
 ) {
-    println!(
+    log::info!(
         "clearing texture with {:?}, dimension {:?}, size {:?}",
-        format, dimension, size
+        format,
+        dimension,
+        size
     );
 
     let texture = ctx.device.create_texture(&wgpu::TextureDescriptor {
@@ -315,6 +318,7 @@ fn clear_texture_2d_uncompressed() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn clear_texture_d32_s8() {
     initialize_test(
         TestParameters::default()
diff --git a/wgpu/tests/common/isolation.rs b/wgpu/tests/common/isolation.rs
new file mode 100644
index 0000000000..e4c7b6e560
--- /dev/null
+++ b/wgpu/tests/common/isolation.rs
@@ -0,0 +1,43 @@
+use std::sync::atomic::{AtomicBool, Ordering};
+
+/// True if a test is in progress somewhere in the process, false otherwise.
+static TEST_ACTIVE_IN_PROCESS: AtomicBool = AtomicBool::new(false);
+
+const OTHER_TEST_IN_PROGRESS_ERROR: &str = "TEST ISOLATION ERROR:
+
+wgpu's test harness requires that no more than one test is running per process.
+
+The best way to facilitate this is by using cargo-nextest which runs each test in its own process
+and has a very good testing UI:
+
+cargo install cargo-nextest
+cargo nextest run
+
+Alternatively, you can run tests in single threaded mode (much slower).
+
+cargo test -- --test-threads=1
+
+Calling std::process::abort()...
+";
+
+/// When this guard is active, enforces that there is only a single test running in the process
+/// at any one time. If there are multiple processes, creating the guard hard terminates the process.
+pub struct OneTestPerProcessGuard(());
+
+impl OneTestPerProcessGuard {
+    pub fn new() -> Self {
+        let other_tests_in_flight = TEST_ACTIVE_IN_PROCESS.swap(true, Ordering::SeqCst);
+        if other_tests_in_flight {
+            log::error!("{}", OTHER_TEST_IN_PROGRESS_ERROR);
+            // Hard exit to call attention to the error
+            std::process::abort();
+        }
+        OneTestPerProcessGuard(())
+    }
+}
+
+impl Drop for OneTestPerProcessGuard {
+    fn drop(&mut self) {
+        TEST_ACTIVE_IN_PROCESS.store(false, Ordering::SeqCst);
+    }
+}
diff --git a/wgpu/tests/common/mod.rs b/wgpu/tests/common/mod.rs
index dedf80f5a3..37fafed043 100644
--- a/wgpu/tests/common/mod.rs
+++ b/wgpu/tests/common/mod.rs
@@ -3,11 +3,13 @@
 
 use std::panic::{catch_unwind, AssertUnwindSafe};
 
+use wgpu::{Adapter, Device, DownlevelFlags, Instance, Queue, Surface};
 use wgt::{Backends, DeviceDescriptor, DownlevelCapabilities, Features, Limits};
 
-use wgpu::{util, Adapter, Device, DownlevelFlags, Instance, Queue};
-
 pub mod image;
+mod isolation;
+
+const CANVAS_ID: &str = "test-canvas";
 
 async fn initialize_device(
     adapter: &Adapter,
@@ -166,18 +168,17 @@ impl TestParameters {
         self
     }
 }
+
 pub fn initialize_test(parameters: TestParameters, test_function: impl FnOnce(TestingContext)) {
     // We don't actually care if it fails
+    #[cfg(not(target_arch = "wasm32"))]
     let _ = env_logger::try_init();
+    #[cfg(target_arch = "wasm32")]
+    let _ = console_log::init_with_level(log::Level::Info);
 
-    let backend_bits = util::backend_bits_from_env().unwrap_or_else(Backends::all);
-    let instance = Instance::new(backend_bits);
-    let adapter = pollster::block_on(util::initialize_adapter_from_env_or_default(
-        &instance,
-        backend_bits,
-        None,
-    ))
-    .expect("could not find suitable adapter on the system");
+    let _test_guard = isolation::OneTestPerProcessGuard::new();
+
+    let (adapter, _) = initialize_adapter();
 
     let adapter_info = adapter.get_info();
     let adapter_lowercase_name = adapter_info.name.to_lowercase();
@@ -187,19 +188,19 @@ pub fn initialize_test(parameters: TestParameters, test_function: impl FnOnce(Te
 
     let missing_features = parameters.required_features - adapter_features;
     if !missing_features.is_empty() {
-        println!("TEST SKIPPED: MISSING FEATURES {:?}", missing_features);
+        log::info!("TEST SKIPPED: MISSING FEATURES {:?}", missing_features);
         return;
     }
 
     if !parameters.required_limits.check_limits(&adapter_limits) {
-        println!("TEST SKIPPED: LIMIT TOO LOW");
+        log::info!("TEST SKIPPED: LIMIT TOO LOW");
         return;
     }
 
     let missing_downlevel_flags =
         parameters.required_downlevel_properties.flags - adapter_downlevel_capabilities.flags;
     if !missing_downlevel_flags.is_empty() {
-        println!(
+        log::info!(
             "TEST SKIPPED: MISSING DOWNLEVEL FLAGS {:?}",
             missing_downlevel_flags
         );
@@ -209,7 +210,7 @@ pub fn initialize_test(parameters: TestParameters, test_function: impl FnOnce(Te
     if adapter_downlevel_capabilities.shader_model
         < parameters.required_downlevel_properties.shader_model
     {
-        println!(
+        log::info!(
             "TEST SKIPPED: LOW SHADER MODEL {:?}",
             adapter_downlevel_capabilities.shader_model
         );
@@ -273,7 +274,7 @@ pub fn initialize_test(parameters: TestParameters, test_function: impl FnOnce(Te
     });
 
     if let Some((reason, true)) = expected_failure_reason {
-        println!("EXPECTED TEST FAILURE SKIPPED: {:?}", reason);
+        log::info!("EXPECTED TEST FAILURE SKIPPED: {:?}", reason);
         return;
     }
 
@@ -301,9 +302,10 @@ pub fn initialize_test(parameters: TestParameters, test_function: impl FnOnce(Te
         // We got the conditions we expected
         if let Some((expected_reason, _)) = expected_failure_reason {
             // Print out reason for the failure
-            println!(
+            log::info!(
                 "GOT EXPECTED TEST FAILURE DUE TO {}: {:?}",
-                failure_cause, expected_reason
+                failure_cause,
+                expected_reason
             );
         }
     } else if let Some((reason, _)) = expected_failure_reason {
@@ -314,6 +316,72 @@ pub fn initialize_test(parameters: TestParameters, test_function: impl FnOnce(Te
     }
 }
 
+fn initialize_adapter() -> (Adapter, SurfaceGuard) {
+    let backend_bits = wgpu::util::backend_bits_from_env().unwrap_or_else(Backends::all);
+    let instance = Instance::new(backend_bits);
+    let compatible_surface;
+
+    #[cfg(not(all(target_arch = "wasm32", feature = "webgl")))]
+    {
+        compatible_surface = None;
+    }
+    #[cfg(all(target_arch = "wasm32", feature = "webgl"))]
+    {
+        // On wasm, append a canvas to the document body for initializing the adapter
+        let canvas = create_html_canvas();
+
+        let surface = instance
+            .create_surface_from_canvas(&canvas)
+            .expect("could not create surface from canvas");
+
+        compatible_surface = Some(surface);
+    }
+
+    let compatible_surface: Option<&Surface> = compatible_surface.as_ref();
+    let adapter = pollster::block_on(wgpu::util::initialize_adapter_from_env_or_default(
+        &instance,
+        backend_bits,
+        compatible_surface,
+    ))
+    .expect("could not find suitable adapter on the system");
+
+    (adapter, SurfaceGuard)
+}
+
+struct SurfaceGuard;
+
+#[cfg(all(target_arch = "wasm32", feature = "webgl"))]
+impl Drop for SurfaceGuard {
+    fn drop(&mut self) {
+        delete_html_canvas();
+    }
+}
+
+#[cfg(all(target_arch = "wasm32", feature = "webgl"))]
+fn create_html_canvas() -> web_sys::HtmlCanvasElement {
+    use wasm_bindgen::JsCast;
+
+    web_sys::window()
+        .and_then(|win| win.document())
+        .and_then(|doc| {
+            let body = doc.body().unwrap();
+            let canvas = doc.create_element("Canvas").unwrap();
+            canvas.set_id(CANVAS_ID);
+            body.append_child(&canvas).unwrap();
+            canvas.dyn_into::<web_sys::HtmlCanvasElement>().ok()
+        })
+        .expect("couldn't append canvas to document body")
+}
+
+#[cfg(all(target_arch = "wasm32", feature = "webgl"))]
+fn delete_html_canvas() {
+    if let Some(document) = web_sys::window().and_then(|win| win.document()) {
+        if let Some(element) = document.get_element_by_id(CANVAS_ID) {
+            element.remove();
+        }
+    };
+}
+
 // Run some code in an error scope and assert that validation fails.
 pub fn fail<T>(device: &wgpu::Device, callback: impl FnOnce() -> T) -> T {
     device.push_error_scope(wgpu::ErrorFilter::Validation);
diff --git a/wgpu/tests/device.rs b/wgpu/tests/device.rs
index 2763a6a21a..69da6b9f3e 100644
--- a/wgpu/tests/device.rs
+++ b/wgpu/tests/device.rs
@@ -1,6 +1,9 @@
+use wasm_bindgen_test::*;
+
 use crate::common::{initialize_test, TestParameters};
 
 #[test]
+#[wasm_bindgen_test]
 fn device_initialization() {
     initialize_test(TestParameters::default(), |_ctx| {
         // intentionally empty
diff --git a/wgpu/tests/encoder.rs b/wgpu/tests/encoder.rs
index f722f8e34b..119f2cad8d 100644
--- a/wgpu/tests/encoder.rs
+++ b/wgpu/tests/encoder.rs
@@ -1,6 +1,8 @@
 use crate::common::{initialize_test, TestParameters};
+use wasm_bindgen_test::*;
 
 #[test]
+#[wasm_bindgen_test]
 fn drop_encoder() {
     initialize_test(TestParameters::default(), |ctx| {
         let encoder = ctx
diff --git a/wgpu/tests/instance.rs b/wgpu/tests/instance.rs
index c5de0480bd..1ead17c57e 100644
--- a/wgpu/tests/instance.rs
+++ b/wgpu/tests/instance.rs
@@ -1,4 +1,7 @@
+use wasm_bindgen_test::*;
+
 #[test]
+#[wasm_bindgen_test]
 fn initialize() {
     let _ = wgpu::Instance::new(
         wgpu::util::backend_bits_from_env().unwrap_or_else(wgpu::Backends::all),
diff --git a/wgpu/tests/poll.rs b/wgpu/tests/poll.rs
index 6113436d0b..77cac2a35f 100644
--- a/wgpu/tests/poll.rs
+++ b/wgpu/tests/poll.rs
@@ -7,6 +7,7 @@ use wgpu::{
 };
 
 use crate::common::{initialize_test, TestParameters, TestingContext};
+use wasm_bindgen_test::*;
 
 fn generate_dummy_work(ctx: &TestingContext) -> CommandBuffer {
     let buffer = ctx.device.create_buffer(&BufferDescriptor {
@@ -53,6 +54,7 @@ fn generate_dummy_work(ctx: &TestingContext) -> CommandBuffer {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn wait() {
     initialize_test(TestParameters::default().skip(), |ctx| {
         let cmd_buf = generate_dummy_work(&ctx);
@@ -63,6 +65,7 @@ fn wait() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn double_wait() {
     initialize_test(TestParameters::default().skip(), |ctx| {
         let cmd_buf = generate_dummy_work(&ctx);
@@ -74,6 +77,7 @@ fn double_wait() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn wait_on_submission() {
     initialize_test(TestParameters::default().skip(), |ctx| {
         let cmd_buf = generate_dummy_work(&ctx);
@@ -84,6 +88,7 @@ fn wait_on_submission() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn double_wait_on_submission() {
     initialize_test(TestParameters::default().skip(), |ctx| {
         let cmd_buf = generate_dummy_work(&ctx);
@@ -95,6 +100,7 @@ fn double_wait_on_submission() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn wait_out_of_order() {
     initialize_test(TestParameters::default().skip(), |ctx| {
         let cmd_buf1 = generate_dummy_work(&ctx);
diff --git a/wgpu/tests/queue_transfer.rs b/wgpu/tests/queue_transfer.rs
index 7724c291cd..90310ea6e7 100644
--- a/wgpu/tests/queue_transfer.rs
+++ b/wgpu/tests/queue_transfer.rs
@@ -3,8 +3,10 @@
 use std::num::NonZeroU32;
 
 use crate::common::{fail, initialize_test, TestParameters};
+use wasm_bindgen_test::*;
 
 #[test]
+#[wasm_bindgen_test]
 fn queue_write_texture_overflow() {
     initialize_test(TestParameters::default(), |ctx| {
         let texture = ctx.device.create_texture(&wgpu::TextureDescriptor {
diff --git a/wgpu/tests/resource_descriptor_accessor.rs b/wgpu/tests/resource_descriptor_accessor.rs
index f43a996a40..cbb3aac422 100644
--- a/wgpu/tests/resource_descriptor_accessor.rs
+++ b/wgpu/tests/resource_descriptor_accessor.rs
@@ -1,7 +1,9 @@
 use crate::common::{initialize_test, TestParameters};
+use wasm_bindgen_test::*;
 
 /// Buffer's size and usage can be read back.
 #[test]
+#[wasm_bindgen_test]
 fn buffer_size_and_usage() {
     initialize_test(TestParameters::default(), |ctx| {
         let buffer = ctx.device.create_buffer(&wgpu::BufferDescriptor {
diff --git a/wgpu/tests/resource_error.rs b/wgpu/tests/resource_error.rs
index 81d50e5800..716c646a7d 100644
--- a/wgpu/tests/resource_error.rs
+++ b/wgpu/tests/resource_error.rs
@@ -1,6 +1,8 @@
 use crate::common::{fail, initialize_test, valid, TestParameters};
+use wasm_bindgen_test::*;
 
 #[test]
+#[wasm_bindgen_test]
 fn bad_buffer() {
     // Create a buffer with bad parameters and call a few methods.
     // Validation should fail but there should be not panic.
@@ -24,6 +26,7 @@ fn bad_buffer() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn bad_texture() {
     // Create a texture with bad parameters and call a few methods.
     // Validation should fail but there should be not panic.
diff --git a/wgpu/tests/root.rs b/wgpu/tests/root.rs
index e721e3a7a7..a3132eda81 100644
--- a/wgpu/tests/root.rs
+++ b/wgpu/tests/root.rs
@@ -1,3 +1,5 @@
+use wasm_bindgen_test::wasm_bindgen_test_configure;
+
 // All files containing tests
 mod common;
 
@@ -20,3 +22,5 @@ mod transfer;
 mod vertex_indices;
 mod write_texture;
 mod zero_init_texture_after_discard;
+
+wasm_bindgen_test_configure!(run_in_browser);
diff --git a/wgpu/tests/shader/numeric_builtins.rs b/wgpu/tests/shader/numeric_builtins.rs
index 83b278cfbf..c215054cc3 100644
--- a/wgpu/tests/shader/numeric_builtins.rs
+++ b/wgpu/tests/shader/numeric_builtins.rs
@@ -1,3 +1,4 @@
+use wasm_bindgen_test::*;
 use wgpu::{DownlevelFlags, Limits};
 
 use crate::{
@@ -40,6 +41,7 @@ fn create_numeric_builtin_test() -> Vec<ShaderTest> {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn numeric_builtins() {
     initialize_test(
         TestParameters::default()
diff --git a/wgpu/tests/shader/struct_layout.rs b/wgpu/tests/shader/struct_layout.rs
index 2250143b5f..c0bba4d1ed 100644
--- a/wgpu/tests/shader/struct_layout.rs
+++ b/wgpu/tests/shader/struct_layout.rs
@@ -1,5 +1,6 @@
 use std::fmt::Write;
 
+use wasm_bindgen_test::*;
 use wgpu::{Backends, DownlevelFlags, Features, Limits};
 
 use crate::{
@@ -177,6 +178,7 @@ fn create_struct_layout_tests(storage_type: InputStorageType) -> Vec<ShaderTest>
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn uniform_input() {
     initialize_test(
         TestParameters::default()
@@ -193,6 +195,7 @@ fn uniform_input() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn storage_input() {
     initialize_test(
         TestParameters::default()
@@ -209,6 +212,7 @@ fn storage_input() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn push_constant_input() {
     initialize_test(
         TestParameters::default()
diff --git a/wgpu/tests/shader_primitive_index/mod.rs b/wgpu/tests/shader_primitive_index/mod.rs
index 5e6c6b1b70..a7f68936f2 100644
--- a/wgpu/tests/shader_primitive_index/mod.rs
+++ b/wgpu/tests/shader_primitive_index/mod.rs
@@ -1,5 +1,6 @@
 use crate::common::{initialize_test, TestParameters, TestingContext};
 use std::num::NonZeroU32;
+use wasm_bindgen_test::*;
 use wgpu::util::{align_to, DeviceExt};
 
 //
@@ -37,6 +38,7 @@ use wgpu::util::{align_to, DeviceExt};
 // buffer [3, 4, 5, 0, 1, 2]. This also swaps the resulting pixel colors.
 //
 #[test]
+#[wasm_bindgen_test]
 fn draw() {
     //
     //   +-----+-----+
@@ -61,6 +63,7 @@ fn draw() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn draw_indexed() {
     //
     //   +-----+-----+
diff --git a/wgpu/tests/texture_bounds.rs b/wgpu/tests/texture_bounds.rs
index 2796815198..635dac8480 100644
--- a/wgpu/tests/texture_bounds.rs
+++ b/wgpu/tests/texture_bounds.rs
@@ -1,30 +1,32 @@
 //! Tests for texture copy bounds checks.
 
-use crate::common::{initialize_test, TestParameters};
+use crate::common::{fail_if, initialize_test, TestParameters};
 use std::num::NonZeroU32;
+use wasm_bindgen_test::*;
 
 #[test]
+#[wasm_bindgen_test]
 fn bad_copy_origin() {
     fn try_origin(origin: wgpu::Origin3d, size: wgpu::Extent3d, should_panic: bool) {
-        let mut parameters = TestParameters::default();
-        if should_panic {
-            parameters = parameters.failure();
-        }
+        let parameters = TestParameters::default();
 
         initialize_test(parameters, |ctx| {
             let texture = ctx.device.create_texture(&TEXTURE_DESCRIPTOR);
             let data = vec![255; BUFFER_SIZE as usize];
-            ctx.queue.write_texture(
-                wgpu::ImageCopyTexture {
-                    texture: &texture,
-                    mip_level: 0,
-                    origin,
-                    aspect: wgpu::TextureAspect::All,
-                },
-                &data,
-                BUFFER_COPY_LAYOUT,
-                size,
-            );
+
+            fail_if(&ctx.device, should_panic, || {
+                ctx.queue.write_texture(
+                    wgpu::ImageCopyTexture {
+                        texture: &texture,
+                        mip_level: 0,
+                        origin,
+                        aspect: wgpu::TextureAspect::All,
+                    },
+                    &data,
+                    BUFFER_COPY_LAYOUT,
+                    size,
+                )
+            });
         });
     }
 
diff --git a/wgpu/tests/vertex_indices/mod.rs b/wgpu/tests/vertex_indices/mod.rs
index 177b857448..2c739e8b3d 100644
--- a/wgpu/tests/vertex_indices/mod.rs
+++ b/wgpu/tests/vertex_indices/mod.rs
@@ -1,5 +1,6 @@
 use std::num::NonZeroU64;
 
+use wasm_bindgen_test::*;
 use wgpu::util::DeviceExt;
 
 use crate::common::{initialize_test, TestParameters, TestingContext};
@@ -131,6 +132,7 @@ fn pulling_common(
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn draw() {
     initialize_test(TestParameters::default().test_features_limits(), |ctx| {
         pulling_common(ctx, &[0, 1, 2, 3, 4, 5], |cmb| {
@@ -140,6 +142,7 @@ fn draw() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn draw_vertex_offset() {
     initialize_test(
         TestParameters::default()
@@ -155,6 +158,7 @@ fn draw_vertex_offset() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn draw_instanced() {
     initialize_test(TestParameters::default().test_features_limits(), |ctx| {
         pulling_common(ctx, &[0, 1, 2, 3, 4, 5], |cmb| {
@@ -164,6 +168,7 @@ fn draw_instanced() {
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn draw_instanced_offset() {
     initialize_test(
         TestParameters::default()
diff --git a/wgpu/tests/write_texture.rs b/wgpu/tests/write_texture.rs
index 742e97d818..9d351ed70c 100644
--- a/wgpu/tests/write_texture.rs
+++ b/wgpu/tests/write_texture.rs
@@ -3,8 +3,10 @@
 use crate::common::{initialize_test, TestParameters};
 
 use std::num::NonZeroU32;
+use wasm_bindgen_test::*;
 
 #[test]
+#[wasm_bindgen_test]
 fn write_texture_subset() {
     let size = 256;
     let parameters = TestParameters::default().backend_failure(wgpu::Backends::DX12);
diff --git a/wgpu/tests/zero_init_texture_after_discard.rs b/wgpu/tests/zero_init_texture_after_discard.rs
index 6043d0388f..3c584eb7e4 100644
--- a/wgpu/tests/zero_init_texture_after_discard.rs
+++ b/wgpu/tests/zero_init_texture_after_discard.rs
@@ -1,9 +1,11 @@
 use std::num::NonZeroU32;
 
 use crate::common::{initialize_test, TestParameters};
+use wasm_bindgen_test::*;
 
 // Checks if discarding a color target resets its init state, causing a zero read of this texture when copied in after submit of the encoder.
 #[test]
+#[wasm_bindgen_test]
 fn discarding_color_target_resets_texture_init_state_check_visible_on_copy_after_submit() {
     initialize_test(TestParameters::default(), |ctx| {
         let (texture, readback_buffer) =
@@ -39,6 +41,7 @@ fn discarding_color_target_resets_texture_init_state_check_visible_on_copy_after
 
 // Checks if discarding a color target resets its init state, causing a zero read of this texture when copied in the same encoder to a buffer.
 #[test]
+#[wasm_bindgen_test]
 fn discarding_color_target_resets_texture_init_state_check_visible_on_copy_in_same_encoder() {
     initialize_test(TestParameters::default(), |ctx| {
         let (texture, readback_buffer) =
@@ -67,6 +70,7 @@ fn discarding_color_target_resets_texture_init_state_check_visible_on_copy_in_sa
 }
 
 #[test]
+#[wasm_bindgen_test]
 #[allow(clippy::single_element_loop)]
 fn discarding_depth_target_resets_texture_init_state_check_visible_on_copy_in_same_encoder() {
     initialize_test(
@@ -109,6 +113,7 @@ fn discarding_depth_target_resets_texture_init_state_check_visible_on_copy_in_sa
 }
 
 #[test]
+#[wasm_bindgen_test]
 fn discarding_either_depth_or_stencil_aspect() {
     initialize_test(TestParameters::default(), |ctx| {
         let (texture, _) = create_white_texture_and_readback_buffer(