diff --git a/Cargo.lock b/Cargo.lock
index 9346961bc16..d7f0fdd62a5 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1787,9 +1787,9 @@ dependencies = [
 
 [[package]]
 name = "gfx-backend-dx12"
-version = "0.5.4"
+version = "0.5.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "69af46bd4fc5479cc7d90bf77ccee0da5dd88d4d27b87986a582f9eec97d02d4"
+checksum = "6e1a979a793023717bcaa7511c8cbb449bab550c093737c98674a659a2bbaf73"
 dependencies = [
  "bitflags",
  "d3d12",
@@ -1815,9 +1815,9 @@ dependencies = [
 
 [[package]]
 name = "gfx-backend-metal"
-version = "0.5.2"
+version = "0.5.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "205f3ca8e74ed814ea2c0206d47d8925077673cab2e21f9b12d48ff781cf87ee"
+checksum = "412a1e0e53e9e325a7c2e0316f1a4e8a14cbe8d8bfb5f030bc3895692f8a8254"
 dependencies = [
  "arrayvec 0.5.1",
  "bitflags",
@@ -1842,9 +1842,9 @@ dependencies = [
 
 [[package]]
 name = "gfx-backend-vulkan"
-version = "0.5.6"
+version = "0.5.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "45ff36feae801fa23d29acd74082603a0145a697a23595757dd4e78828ab33da"
+checksum = "04af900c2597587b35e801e9d3f91fd8078cc06067421a22caa640ad2b1bc53e"
 dependencies = [
  "arrayvec 0.5.1",
  "ash",
@@ -1872,9 +1872,9 @@ dependencies = [
 
 [[package]]
 name = "gfx-hal"
-version = "0.5.0"
+version = "0.5.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bc96180204064c9493e0fe4a9efeb721e0ac59fe8e1906d0c659142a93114fb1"
+checksum = "f13e8fd6aaa8f50146b9519999432e097da43466749d8863c53409322c705339"
 dependencies = [
  "bitflags",
  "raw-window-handle",
@@ -1883,8 +1883,7 @@ dependencies = [
 [[package]]
 name = "gfx-memory"
 version = "0.1.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c2eed6cda674d9cd4d92229102dbd544292124533d236904f987e9afab456137"
+source = "git+https://github.com/gfx-rs/gfx-extras?rev=438353c3f75368c12024ad2fc03cbeb15f351fd9#438353c3f75368c12024ad2fc03cbeb15f351fd9"
 dependencies = [
  "fxhash",
  "gfx-hal",
@@ -6563,7 +6562,7 @@ dependencies = [
 [[package]]
 name = "wgpu-core"
 version = "0.5.0"
-source = "git+https://github.com/gfx-rs/wgpu#ac9587e9ced5b043abad68e260cb8c9e812cffb5"
+source = "git+https://github.com/gfx-rs/wgpu#fc2dd481b2713cd0eda6ffa540faeaf7418fd051"
 dependencies = [
  "arrayvec 0.5.1",
  "bitflags",
@@ -6592,7 +6591,7 @@ dependencies = [
 [[package]]
 name = "wgpu-types"
 version = "0.5.0"
-source = "git+https://github.com/gfx-rs/wgpu#ac9587e9ced5b043abad68e260cb8c9e812cffb5"
+source = "git+https://github.com/gfx-rs/wgpu#fc2dd481b2713cd0eda6ffa540faeaf7418fd051"
 dependencies = [
  "bitflags",
  "peek-poke 0.2.0 (registry+https://github.com/rust-lang/crates.io-index)",
diff --git a/components/script/dom/bindings/trace.rs b/components/script/dom/bindings/trace.rs
index 9d7b385adc3..6208a7bc1ba 100644
--- a/components/script/dom/bindings/trace.rs
+++ b/components/script/dom/bindings/trace.rs
@@ -165,10 +165,10 @@ use tendril::{StrTendril, TendrilSink};
 use time::{Duration, Timespec, Tm};
 use uuid::Uuid;
 use webgpu::{
-    wgpu::command::RawPass, WebGPU, WebGPUAdapter, WebGPUBindGroup, WebGPUBindGroupLayout,
-    WebGPUBuffer, WebGPUCommandBuffer, WebGPUCommandEncoder, WebGPUComputePipeline, WebGPUDevice,
-    WebGPUPipelineLayout, WebGPUQueue, WebGPURenderPipeline, WebGPUSampler, WebGPUShaderModule,
-    WebGPUTexture, WebGPUTextureView,
+    wgpu::command::RawPass, wgpu::id, WebGPU, WebGPUAdapter, WebGPUBindGroup,
+    WebGPUBindGroupLayout, WebGPUBuffer, WebGPUCommandBuffer, WebGPUCommandEncoder,
+    WebGPUComputePipeline, WebGPUDevice, WebGPUPipelineLayout, WebGPUQueue, WebGPURenderPipeline,
+    WebGPUSampler, WebGPUShaderModule, WebGPUTexture, WebGPUTextureView,
 };
 use webrender_api::{DocumentId, ExternalImageId, ImageKey};
 use webxr_api::SwapChainId as WebXRSwapChainId;
@@ -581,7 +581,7 @@ unsafe_no_jsmanaged_fields!(WebGPUContextId);
 unsafe_no_jsmanaged_fields!(WebGPUCommandBuffer);
 unsafe_no_jsmanaged_fields!(WebGPUCommandEncoder);
 unsafe_no_jsmanaged_fields!(WebGPUDevice);
-unsafe_no_jsmanaged_fields!(Option<RawPass>);
+unsafe_no_jsmanaged_fields!(Option<RawPass<id::CommandEncoderId>>);
 unsafe_no_jsmanaged_fields!(GPUBufferState);
 unsafe_no_jsmanaged_fields!(GPUCommandEncoderState);
 unsafe_no_jsmanaged_fields!(WebXRSwapChainId);
diff --git a/components/script/dom/gpuadapter.rs b/components/script/dom/gpuadapter.rs
index 7947a069335..e510afce17d 100644
--- a/components/script/dom/gpuadapter.rs
+++ b/components/script/dom/gpuadapter.rs
@@ -84,7 +84,9 @@ impl GPUAdapterMethods for GPUAdapter {
             extensions: wgt::Extensions::empty(),
             limits: wgt::Limits {
                 max_bind_groups: descriptor.limits.maxBindGroups,
+                ..Default::default()
             },
+            shader_validation: true,
         };
         let id = self
             .global()
diff --git a/components/script/dom/gpubindgroup.rs b/components/script/dom/gpubindgroup.rs
index 4767d28d0f1..8965952051c 100644
--- a/components/script/dom/gpubindgroup.rs
+++ b/components/script/dom/gpubindgroup.rs
@@ -5,34 +5,52 @@
 use crate::dom::bindings::cell::DomRefCell;
 use crate::dom::bindings::codegen::Bindings::GPUBindGroupBinding::GPUBindGroupMethods;
 use crate::dom::bindings::reflector::{reflect_dom_object, Reflector};
-use crate::dom::bindings::root::DomRoot;
+use crate::dom::bindings::root::{Dom, DomRoot};
 use crate::dom::bindings::str::DOMString;
 use crate::dom::globalscope::GlobalScope;
+use crate::dom::gpubindgrouplayout::GPUBindGroupLayout;
 use dom_struct::dom_struct;
 use std::cell::Cell;
-use webgpu::WebGPUBindGroup;
+use webgpu::{WebGPUBindGroup, WebGPUDevice};
 
 #[dom_struct]
 pub struct GPUBindGroup {
     reflector_: Reflector,
     label: DomRefCell<Option<DOMString>>,
     bind_group: WebGPUBindGroup,
+    device: WebGPUDevice,
+    layout: Dom<GPUBindGroupLayout>,
     valid: Cell<bool>,
 }
 
 impl GPUBindGroup {
-    fn new_inherited(bind_group: WebGPUBindGroup, valid: bool) -> Self {
+    fn new_inherited(
+        bind_group: WebGPUBindGroup,
+        device: WebGPUDevice,
+        valid: bool,
+        layout: &GPUBindGroupLayout,
+    ) -> Self {
         Self {
             reflector_: Reflector::new(),
             label: DomRefCell::new(None),
             bind_group,
+            device,
             valid: Cell::new(valid),
+            layout: Dom::from_ref(layout),
         }
     }
 
-    pub fn new(global: &GlobalScope, bind_group: WebGPUBindGroup, valid: bool) -> DomRoot<Self> {
+    pub fn new(
+        global: &GlobalScope,
+        bind_group: WebGPUBindGroup,
+        device: WebGPUDevice,
+        valid: bool,
+        layout: &GPUBindGroupLayout,
+    ) -> DomRoot<Self> {
         reflect_dom_object(
-            Box::new(GPUBindGroup::new_inherited(bind_group, valid)),
+            Box::new(GPUBindGroup::new_inherited(
+                bind_group, device, valid, layout,
+            )),
             global,
         )
     }
diff --git a/components/script/dom/gpubindgrouplayout.rs b/components/script/dom/gpubindgrouplayout.rs
index 8c141256261..216058a89a0 100644
--- a/components/script/dom/gpubindgrouplayout.rs
+++ b/components/script/dom/gpubindgrouplayout.rs
@@ -3,60 +3,40 @@
  * file, You can obtain one at https://mozilla.org/MPL/2.0/. */
 
 use crate::dom::bindings::cell::DomRefCell;
-use crate::dom::bindings::codegen::Bindings::GPUBindGroupLayoutBinding::{
-    GPUBindGroupLayoutEntry, GPUBindGroupLayoutMethods,
-};
+use crate::dom::bindings::codegen::Bindings::GPUBindGroupLayoutBinding::GPUBindGroupLayoutMethods;
 use crate::dom::bindings::reflector::{reflect_dom_object, Reflector};
 use crate::dom::bindings::root::DomRoot;
 use crate::dom::bindings::str::DOMString;
 use crate::dom::globalscope::GlobalScope;
 use dom_struct::dom_struct;
 use std::cell::Cell;
-use webgpu::{WebGPU, WebGPUBindGroupLayout};
+use webgpu::WebGPUBindGroupLayout;
 
 #[dom_struct]
 pub struct GPUBindGroupLayout {
     reflector_: Reflector,
     label: DomRefCell<Option<DOMString>>,
     bind_group_layout: WebGPUBindGroupLayout,
-    #[ignore_malloc_size_of = "defined in webgpu"]
-    bindings: Vec<GPUBindGroupLayoutEntry>,
-    #[ignore_malloc_size_of = "defined in webgpu"]
-    channel: WebGPU,
     valid: Cell<bool>,
 }
 
 impl GPUBindGroupLayout {
-    fn new_inherited(
-        channel: WebGPU,
-        bind_group_layout: WebGPUBindGroupLayout,
-        bindings: Vec<GPUBindGroupLayoutEntry>,
-        valid: bool,
-    ) -> Self {
+    fn new_inherited(bind_group_layout: WebGPUBindGroupLayout, valid: bool) -> Self {
         Self {
             reflector_: Reflector::new(),
-            channel,
             label: DomRefCell::new(None),
             bind_group_layout,
-            bindings,
             valid: Cell::new(valid),
         }
     }
 
     pub fn new(
         global: &GlobalScope,
-        channel: WebGPU,
         bind_group_layout: WebGPUBindGroupLayout,
-        bindings: Vec<GPUBindGroupLayoutEntry>,
         valid: bool,
     ) -> DomRoot<Self> {
         reflect_dom_object(
-            Box::new(GPUBindGroupLayout::new_inherited(
-                channel,
-                bind_group_layout,
-                bindings,
-                valid,
-            )),
+            Box::new(GPUBindGroupLayout::new_inherited(bind_group_layout, valid)),
             global,
         )
     }
@@ -70,10 +50,6 @@ impl GPUBindGroupLayout {
     pub fn id(&self) -> WebGPUBindGroupLayout {
         self.bind_group_layout
     }
-
-    pub fn bindings(&self) -> &[GPUBindGroupLayoutEntry] {
-        &self.bindings
-    }
 }
 
 impl GPUBindGroupLayoutMethods for GPUBindGroupLayout {
diff --git a/components/script/dom/gpubuffer.rs b/components/script/dom/gpubuffer.rs
index 1c756dae15f..a23a30a4b11 100644
--- a/components/script/dom/gpubuffer.rs
+++ b/components/script/dom/gpubuffer.rs
@@ -110,7 +110,7 @@ impl GPUBuffer {
         self.state.borrow()
     }
 
-    pub fn valid(&self) -> bool {
+    pub fn is_valid(&self) -> bool {
         self.valid.get()
     }
 }
diff --git a/components/script/dom/gpucommandencoder.rs b/components/script/dom/gpucommandencoder.rs
index d0ba2dc8624..0615492dec6 100644
--- a/components/script/dom/gpucommandencoder.rs
+++ b/components/script/dom/gpucommandencoder.rs
@@ -240,8 +240,8 @@ impl GPUCommandEncoderMethods for GPUCommandEncoder {
             None => false,
         };
         valid &= (*self.state.borrow() == GPUCommandEncoderState::Open) &&
-            source.valid() &&
-            destination.valid() &
+            source.is_valid() &&
+            destination.is_valid() &
                 !(size & BUFFER_COPY_ALIGN_MASK == 0) &
                 !(source_offset & BUFFER_COPY_ALIGN_MASK == 0) &
                 !(destination_offset & BUFFER_COPY_ALIGN_MASK == 0) &
diff --git a/components/script/dom/gpucomputepassencoder.rs b/components/script/dom/gpucomputepassencoder.rs
index f202cc37038..68fb641bd25 100644
--- a/components/script/dom/gpucomputepassencoder.rs
+++ b/components/script/dom/gpucomputepassencoder.rs
@@ -22,6 +22,7 @@ use webgpu::{
         },
         RawPass,
     },
+    wgpu::id,
     WebGPU, WebGPURequest,
 };
 
@@ -32,7 +33,7 @@ pub struct GPUComputePassEncoder {
     channel: WebGPU,
     label: DomRefCell<Option<DOMString>>,
     #[ignore_malloc_size_of = "defined in wgpu-core"]
-    raw_pass: DomRefCell<Option<RawPass>>,
+    raw_pass: DomRefCell<Option<RawPass<id::CommandEncoderId>>>,
     command_encoder: Dom<GPUCommandEncoder>,
 }
 
diff --git a/components/script/dom/gpucomputepipeline.rs b/components/script/dom/gpucomputepipeline.rs
index e96697c95a2..cd5fac795d7 100644
--- a/components/script/dom/gpucomputepipeline.rs
+++ b/components/script/dom/gpucomputepipeline.rs
@@ -10,6 +10,7 @@ use crate::dom::bindings::root::DomRoot;
 use crate::dom::bindings::str::DOMString;
 use crate::dom::globalscope::GlobalScope;
 use dom_struct::dom_struct;
+use std::cell::Cell;
 use webgpu::WebGPUComputePipeline;
 
 #[dom_struct]
@@ -17,20 +18,26 @@ pub struct GPUComputePipeline {
     reflector_: Reflector,
     label: DomRefCell<Option<DOMString>>,
     compute_pipeline: WebGPUComputePipeline,
+    valid: Cell<bool>,
 }
 
 impl GPUComputePipeline {
-    fn new_inherited(compute_pipeline: WebGPUComputePipeline) -> Self {
+    fn new_inherited(compute_pipeline: WebGPUComputePipeline, valid: bool) -> Self {
         Self {
             reflector_: Reflector::new(),
             label: DomRefCell::new(None),
             compute_pipeline,
+            valid: Cell::new(valid),
         }
     }
 
-    pub fn new(global: &GlobalScope, compute_pipeline: WebGPUComputePipeline) -> DomRoot<Self> {
+    pub fn new(
+        global: &GlobalScope,
+        compute_pipeline: WebGPUComputePipeline,
+        valid: bool,
+    ) -> DomRoot<Self> {
         reflect_dom_object(
-            Box::new(GPUComputePipeline::new_inherited(compute_pipeline)),
+            Box::new(GPUComputePipeline::new_inherited(compute_pipeline, valid)),
             global,
         )
     }
diff --git a/components/script/dom/gpudevice.rs b/components/script/dom/gpudevice.rs
index fbbfa541761..b67c6e696e3 100644
--- a/components/script/dom/gpudevice.rs
+++ b/components/script/dom/gpudevice.rs
@@ -5,10 +5,11 @@
 #![allow(unsafe_code)]
 
 use crate::dom::bindings::cell::DomRefCell;
-use crate::dom::bindings::codegen::Bindings::GPUAdapterBinding::GPULimits;
-use crate::dom::bindings::codegen::Bindings::GPUBindGroupBinding::GPUBindGroupDescriptor;
+use crate::dom::bindings::codegen::Bindings::GPUBindGroupBinding::{
+    GPUBindGroupDescriptor, GPUBindingResource,
+};
 use crate::dom::bindings::codegen::Bindings::GPUBindGroupLayoutBinding::{
-    GPUBindGroupLayoutDescriptor, GPUBindGroupLayoutEntry, GPUBindingType,
+    GPUBindGroupLayoutDescriptor, GPUBindingType,
 };
 use crate::dom::bindings::codegen::Bindings::GPUBufferBinding::GPUBufferDescriptor;
 use crate::dom::bindings::codegen::Bindings::GPUComputePipelineBinding::GPUComputePipelineDescriptor;
@@ -55,12 +56,10 @@ use dom_struct::dom_struct;
 use js::jsapi::{Heap, JSObject};
 use js::jsval::{JSVal, ObjectValue};
 use js::typedarray::{ArrayBuffer, CreateWith};
-use std::collections::{HashMap, HashSet};
+use std::num::NonZeroU64;
 use std::ptr::{self, NonNull};
-use webgpu::wgpu::binding_model::{
-    BindGroupEntry, BindGroupLayoutEntry, BindingResource, BindingType, BufferBinding,
-};
-use webgpu::{self, wgpu, wgt, WebGPU, WebGPURequest};
+use webgpu::wgpu::binding_model::BufferBinding;
+use webgpu::{self, wgt, WebGPU, WebGPUBindings, WebGPURequest};
 
 #[dom_struct]
 pub struct GPUDevice {
@@ -274,149 +273,93 @@ impl GPUDeviceMethods for GPUDevice {
         &self,
         descriptor: &GPUBindGroupLayoutDescriptor,
     ) -> DomRoot<GPUBindGroupLayout> {
-        #[derive(Clone)]
-        struct MaxLimits {
-            max_uniform_buffers_per_shader_stage: i32,
-            max_storage_buffers_per_shader_stage: i32,
-            max_sampled_textures_per_shader_stage: i32,
-            max_storage_textures_per_shader_stage: i32,
-            max_samplers_per_shader_stage: i32,
-        }
-        let mut storeBindings = HashSet::new();
-        // TODO: We should have these limits on device creation
-        let limits = GPULimits::empty();
-
-        let mut validation_map = HashMap::new();
-        let maxLimits = MaxLimits {
-            max_uniform_buffers_per_shader_stage: limits.maxUniformBuffersPerShaderStage as i32,
-            max_storage_buffers_per_shader_stage: limits.maxStorageBuffersPerShaderStage as i32,
-            max_sampled_textures_per_shader_stage: limits.maxSampledTexturesPerShaderStage as i32,
-            max_storage_textures_per_shader_stage: limits.maxStorageTexturesPerShaderStage as i32,
-            max_samplers_per_shader_stage: limits.maxSamplersPerShaderStage as i32,
-        };
-        validation_map.insert(wgt::ShaderStage::VERTEX, maxLimits.clone());
-        validation_map.insert(wgt::ShaderStage::FRAGMENT, maxLimits.clone());
-        validation_map.insert(wgt::ShaderStage::COMPUTE, maxLimits.clone());
-        let mut max_dynamic_uniform_buffers_per_pipeline_layout =
-            limits.maxDynamicUniformBuffersPerPipelineLayout as i32;
-        let mut max_dynamic_storage_buffers_per_pipeline_layout =
-            limits.maxDynamicStorageBuffersPerPipelineLayout as i32;
-        let mut valid = true;
-
-        let bindings = descriptor
+        let entries = descriptor
             .entries
             .iter()
             .map(|bind| {
-                // TODO: binding must be >= 0
-                storeBindings.insert(bind.binding);
                 let visibility = match wgt::ShaderStage::from_bits(bind.visibility) {
                     Some(visibility) => visibility,
-                    None => {
-                        valid = false;
-                        wgt::ShaderStage::from_bits(0).unwrap()
-                    },
+                    None => wgt::ShaderStage::from_bits(0).unwrap(),
                 };
                 let ty = match bind.type_ {
-                    GPUBindingType::Uniform_buffer => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_uniform_buffers_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            max_dynamic_uniform_buffers_per_pipeline_layout -= 1;
-                        };
-                        BindingType::UniformBuffer
+                    GPUBindingType::Uniform_buffer => wgt::BindingType::UniformBuffer {
+                        dynamic: bind.hasDynamicOffset,
+                        min_binding_size: if bind.minBufferBindingSize == 0 {
+                            None
+                        } else {
+                            Some(NonZeroU64::new(bind.minBufferBindingSize).unwrap())
+                        },
                     },
-                    GPUBindingType::Storage_buffer => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_storage_buffers_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            max_dynamic_storage_buffers_per_pipeline_layout -= 1;
-                        };
-                        BindingType::StorageBuffer
+                    GPUBindingType::Storage_buffer => wgt::BindingType::StorageBuffer {
+                        dynamic: bind.hasDynamicOffset,
+                        min_binding_size: if bind.minBufferBindingSize == 0 {
+                            None
+                        } else {
+                            Some(NonZeroU64::new(bind.minBufferBindingSize).unwrap())
+                        },
+                        readonly: false,
                     },
-                    GPUBindingType::Readonly_storage_buffer => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_storage_buffers_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            max_dynamic_storage_buffers_per_pipeline_layout -= 1;
-                        };
-                        BindingType::ReadonlyStorageBuffer
+                    GPUBindingType::Readonly_storage_buffer => wgt::BindingType::StorageBuffer {
+                        dynamic: bind.hasDynamicOffset,
+                        min_binding_size: if bind.minBufferBindingSize == 0 {
+                            None
+                        } else {
+                            Some(NonZeroU64::new(bind.minBufferBindingSize).unwrap())
+                        },
+                        readonly: true,
                     },
-                    GPUBindingType::Sampled_texture => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_sampled_textures_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            valid = false
-                        };
-                        BindingType::SampledTexture
+                    GPUBindingType::Sampled_texture => wgt::BindingType::SampledTexture {
+                        dimension: bind
+                            .viewDimension
+                            .map_or(wgt::TextureViewDimension::D2, |v| {
+                                convert_texture_view_dimension(v)
+                            }),
+                        component_type: if let Some(c) = bind.textureComponentType {
+                            match c {
+                                GPUTextureComponentType::Float => wgt::TextureComponentType::Float,
+                                GPUTextureComponentType::Sint => wgt::TextureComponentType::Sint,
+                                GPUTextureComponentType::Uint => wgt::TextureComponentType::Uint,
+                            }
+                        } else {
+                            wgt::TextureComponentType::Float
+                        },
+                        multisampled: bind.multisampled,
                     },
-                    GPUBindingType::Readonly_storage_texture => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_storage_textures_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            valid = false
-                        };
-                        BindingType::ReadonlyStorageTexture
+                    GPUBindingType::Readonly_storage_texture => wgt::BindingType::StorageTexture {
+                        dimension: bind
+                            .viewDimension
+                            .map_or(wgt::TextureViewDimension::D2, |v| {
+                                convert_texture_view_dimension(v)
+                            }),
+                        format: bind
+                            .storageTextureFormat
+                            .map_or(wgt::TextureFormat::Bgra8UnormSrgb, |f| {
+                                convert_texture_format(f)
+                            }),
+                        readonly: true,
                     },
-                    GPUBindingType::Writeonly_storage_texture => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_storage_textures_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            valid = false
-                        };
-                        BindingType::WriteonlyStorageTexture
+                    GPUBindingType::Writeonly_storage_texture => wgt::BindingType::StorageTexture {
+                        dimension: bind
+                            .viewDimension
+                            .map_or(wgt::TextureViewDimension::D2, |v| {
+                                convert_texture_view_dimension(v)
+                            }),
+                        format: bind
+                            .storageTextureFormat
+                            .map_or(wgt::TextureFormat::Bgra8UnormSrgb, |f| {
+                                convert_texture_format(f)
+                            }),
+                        readonly: true,
                     },
-                    GPUBindingType::Sampler => {
-                        if let Some(limit) = validation_map.get_mut(&visibility) {
-                            limit.max_samplers_per_shader_stage -= 1;
-                        }
-                        if bind.hasDynamicOffset {
-                            valid = false
-                        };
-                        BindingType::Sampler
+                    GPUBindingType::Sampler => wgt::BindingType::Sampler { comparison: false },
+                    GPUBindingType::Comparison_sampler => {
+                        wgt::BindingType::Sampler { comparison: true }
                     },
                 };
 
-                BindGroupLayoutEntry {
-                    binding: bind.binding,
-                    visibility,
-                    ty,
-                    has_dynamic_offset: bind.hasDynamicOffset,
-                    multisampled: bind.multisampled,
-                    texture_component_type: match bind.textureComponentType {
-                        GPUTextureComponentType::Float => wgt::TextureComponentType::Float,
-                        GPUTextureComponentType::Sint => wgt::TextureComponentType::Sint,
-                        GPUTextureComponentType::Uint => wgt::TextureComponentType::Uint,
-                    },
-                    storage_texture_format: match bind.storageTextureFormat {
-                        Some(s) => convert_texture_format(s),
-                        None => wgt::TextureFormat::Bgra8UnormSrgb,
-                    },
-                    view_dimension: convert_texture_view_dimension(bind.viewDimension),
-                }
+                wgt::BindGroupLayoutEntry::new(bind.binding, visibility, ty)
             })
-            .collect::<Vec<BindGroupLayoutEntry>>();
-
-        // bindings are unique
-        valid &= storeBindings.len() == bindings.len();
-
-        // Ensure that values do not exceed the max limit for each ShaderStage.
-        valid &= validation_map.values().all(|stage| {
-            stage.max_uniform_buffers_per_shader_stage >= 0 &&
-                stage.max_storage_buffers_per_shader_stage >= 0 &&
-                stage.max_sampled_textures_per_shader_stage >= 0 &&
-                stage.max_storage_textures_per_shader_stage >= 0 &&
-                stage.max_samplers_per_shader_stage >= 0
-        });
-
-        // DynamicValues does not exceed the max limit for the pipeline
-        valid &= max_dynamic_uniform_buffers_per_pipeline_layout >= 0 &&
-            max_dynamic_storage_buffers_per_pipeline_layout >= 0;
+            .collect::<Vec<_>>();
 
         let bind_group_layout_id = self
             .global()
@@ -428,28 +371,13 @@ impl GPUDeviceMethods for GPUDevice {
             .send(WebGPURequest::CreateBindGroupLayout {
                 device_id: self.device.0,
                 bind_group_layout_id,
-                bindings: bindings.clone(),
+                entries,
             })
             .expect("Failed to create WebGPU BindGroupLayout");
 
         let bgl = webgpu::WebGPUBindGroupLayout(bind_group_layout_id);
 
-        let binds = descriptor
-            .entries
-            .iter()
-            .map(|bind| GPUBindGroupLayoutEntry {
-                binding: bind.binding,
-                hasDynamicOffset: bind.hasDynamicOffset,
-                multisampled: bind.multisampled,
-                type_: bind.type_,
-                visibility: bind.visibility,
-                viewDimension: bind.viewDimension,
-                textureComponentType: bind.textureComponentType,
-                storageTextureFormat: bind.storageTextureFormat,
-            })
-            .collect::<Vec<_>>();
-
-        GPUBindGroupLayout::new(&self.global(), self.channel.clone(), bgl, binds, valid)
+        GPUBindGroupLayout::new(&self.global(), bgl, true)
     }
 
     /// https://gpuweb.github.io/gpuweb/#dom-gpudevice-createpipelinelayout
@@ -457,47 +385,13 @@ impl GPUDeviceMethods for GPUDevice {
         &self,
         descriptor: &GPUPipelineLayoutDescriptor,
     ) -> DomRoot<GPUPipelineLayout> {
-        // TODO: We should have these limits on device creation
-        let limits = GPULimits::empty();
-        let mut bind_group_layouts = Vec::new();
         let mut bgl_ids = Vec::new();
-        let mut max_dynamic_uniform_buffers_per_pipeline_layout =
-            limits.maxDynamicUniformBuffersPerPipelineLayout as i32;
-        let mut max_dynamic_storage_buffers_per_pipeline_layout =
-            limits.maxDynamicStorageBuffersPerPipelineLayout as i32;
         descriptor.bindGroupLayouts.iter().for_each(|each| {
             if each.is_valid() {
-                let id = each.id();
-                bind_group_layouts.push(id);
-                bgl_ids.push(id.0);
+                bgl_ids.push(each.id().0);
             }
-            each.bindings().iter().for_each(|bind| {
-                match bind.type_ {
-                    GPUBindingType::Uniform_buffer => {
-                        if bind.hasDynamicOffset {
-                            max_dynamic_uniform_buffers_per_pipeline_layout -= 1;
-                        };
-                    },
-                    GPUBindingType::Storage_buffer => {
-                        if bind.hasDynamicOffset {
-                            max_dynamic_storage_buffers_per_pipeline_layout -= 1;
-                        };
-                    },
-                    GPUBindingType::Readonly_storage_buffer => {
-                        if bind.hasDynamicOffset {
-                            max_dynamic_storage_buffers_per_pipeline_layout -= 1;
-                        };
-                    },
-                    _ => {},
-                };
-            });
         });
 
-        let valid = descriptor.bindGroupLayouts.len() <= limits.maxBindGroups as usize &&
-            descriptor.bindGroupLayouts.len() == bind_group_layouts.len() &&
-            max_dynamic_uniform_buffers_per_pipeline_layout >= 0 &&
-            max_dynamic_storage_buffers_per_pipeline_layout >= 0;
-
         let pipeline_layout_id = self
             .global()
             .wgpu_id_hub()
@@ -513,50 +407,41 @@ impl GPUDeviceMethods for GPUDevice {
             .expect("Failed to create WebGPU PipelineLayout");
 
         let pipeline_layout = webgpu::WebGPUPipelineLayout(pipeline_layout_id);
-        GPUPipelineLayout::new(&self.global(), bind_group_layouts, pipeline_layout, valid)
+        GPUPipelineLayout::new(&self.global(), pipeline_layout, true)
     }
 
     /// https://gpuweb.github.io/gpuweb/#dom-gpudevice-createbindgroup
     fn CreateBindGroup(&self, descriptor: &GPUBindGroupDescriptor) -> DomRoot<GPUBindGroup> {
-        let alignment: u64 = 256;
-        let mut valid = descriptor.layout.bindings().len() == descriptor.entries.len();
-
-        valid &= descriptor.entries.iter().all(|bind| {
-            let buffer_size = bind.resource.buffer.size();
-            let resource_size = bind.resource.size.unwrap_or(buffer_size);
-            let length = bind.resource.offset.checked_add(resource_size);
-            let usage = wgt::BufferUsage::from_bits(bind.resource.buffer.usage()).unwrap();
-
-            length.is_some() &&
-            buffer_size >= length.unwrap() && // check buffer OOB
-            bind.resource.offset % alignment == 0 && // check alignment
-            bind.resource.offset < buffer_size && // on Vulkan offset must be less than size of buffer
-            descriptor.layout.bindings().iter().any(|layout_bind| {
-                let ty = match layout_bind.type_ {
-                    GPUBindingType::Storage_buffer  => wgt::BufferUsage::STORAGE,
-                    // GPUBindingType::Readonly_storage_buffer  => BufferUsage::STORAGE_READ,
-                    GPUBindingType::Uniform_buffer => wgt::BufferUsage::UNIFORM,
-                    _ => unimplemented!(),
-                };
-                // binding must be present in layout
-                layout_bind.binding == bind.binding &&
-                // binding must contain one buffer of its type
-                usage.contains(ty)
-            })
-        });
-
-        let bindings = descriptor
+        let mut valid = descriptor.layout.is_valid();
+        let entries = descriptor
             .entries
             .iter()
-            .map(|bind| BindGroupEntry {
-                binding: bind.binding,
-                resource: BindingResource::Buffer(BufferBinding {
-                    buffer: bind.resource.buffer.id().0,
-                    offset: bind.resource.offset,
-                    size: wgt::BufferSize(
-                        bind.resource.size.unwrap_or(bind.resource.buffer.size()),
-                    ),
-                }),
+            .map(|bind| {
+                (
+                    bind.binding,
+                    match bind.resource {
+                        GPUBindingResource::GPUSampler(ref s) => {
+                            valid &= s.is_valid();
+                            WebGPUBindings::Sampler(s.id().0)
+                        },
+                        GPUBindingResource::GPUTextureView(ref t) => {
+                            valid &= t.is_valid();
+                            WebGPUBindings::TextureView(t.id().0)
+                        },
+                        GPUBindingResource::GPUBufferBindings(ref b) => {
+                            valid &= b.buffer.is_valid();
+                            WebGPUBindings::Buffer(BufferBinding {
+                                buffer_id: b.buffer.id().0,
+                                offset: b.offset,
+                                size: if let Some(s) = b.size {
+                                    wgt::BufferSize(s)
+                                } else {
+                                    wgt::BufferSize::WHOLE
+                                },
+                            })
+                        },
+                    },
+                )
             })
             .collect::<Vec<_>>();
 
@@ -571,12 +456,18 @@ impl GPUDeviceMethods for GPUDevice {
                 device_id: self.device.0,
                 bind_group_id,
                 bind_group_layout_id: descriptor.layout.id().0,
-                bindings,
+                entries,
             })
             .expect("Failed to create WebGPU BindGroup");
 
         let bind_group = webgpu::WebGPUBindGroup(bind_group_id);
-        GPUBindGroup::new(&self.global(), bind_group, valid)
+        GPUBindGroup::new(
+            &self.global(),
+            bind_group,
+            self.device,
+            valid,
+            &*descriptor.layout,
+        )
     }
 
     /// https://gpuweb.github.io/gpuweb/#dom-gpudevice-createshadermodule
@@ -611,6 +502,7 @@ impl GPUDeviceMethods for GPUDevice {
         &self,
         descriptor: &GPUComputePipelineDescriptor,
     ) -> DomRoot<GPUComputePipeline> {
+        let valid = descriptor.parent.layout.is_valid();
         let pipeline = descriptor.parent.layout.id();
         let program = descriptor.computeStage.module.id();
         let entry_point = descriptor.computeStage.entryPoint.to_string();
@@ -632,7 +524,7 @@ impl GPUDeviceMethods for GPUDevice {
             .expect("Failed to create WebGPU ComputePipeline");
 
         let compute_pipeline = webgpu::WebGPUComputePipeline(compute_pipeline_id);
-        GPUComputePipeline::new(&self.global(), compute_pipeline)
+        GPUComputePipeline::new(&self.global(), compute_pipeline, valid)
     }
 
     /// https://gpuweb.github.io/gpuweb/#dom-gpudevice-createcommandencoder
@@ -740,6 +632,7 @@ impl GPUDeviceMethods for GPUDevice {
             lod_max_clamp: *descriptor.lodMaxClamp,
             compare: descriptor.compare.map(|c| convert_compare_function(c)),
             anisotropy_clamp: None,
+            ..Default::default()
         };
         self.channel
             .0
@@ -752,14 +645,7 @@ impl GPUDeviceMethods for GPUDevice {
 
         let sampler = webgpu::WebGPUSampler(sampler_id);
 
-        GPUSampler::new(
-            &self.global(),
-            self.channel.clone(),
-            self.device,
-            compare_enable,
-            sampler,
-            true,
-        )
+        GPUSampler::new(&self.global(), self.device, compare_enable, sampler, true)
     }
 
     /// https://gpuweb.github.io/gpuweb/#dom-gpudevice-createrenderpipeline
@@ -767,12 +653,7 @@ impl GPUDeviceMethods for GPUDevice {
         &self,
         descriptor: &GPURenderPipelineDescriptor,
     ) -> DomRoot<GPURenderPipeline> {
-        let mut valid = descriptor.parent.layout.is_valid();
-        valid &= descriptor.colorStates.len() <= wgpu::device::MAX_COLOR_TARGETS;
-        if descriptor.alphaToCoverageEnabled {
-            valid &= descriptor.sampleCount > 1;
-        }
-
+        let valid = descriptor.parent.layout.is_valid();
         let vertex_module = descriptor.vertexStage.module.id().0;
         let vertex_entry_point = descriptor.vertexStage.entryPoint.to_string();
         let (fragment_module, fragment_entry_point) = match descriptor.fragmentStage {
@@ -813,10 +694,7 @@ impl GPUDeviceMethods for GPUDevice {
                 color_blend: convert_blend_descriptor(&state.colorBlend),
                 write_mask: match wgt::ColorWrite::from_bits(state.writeMask) {
                     Some(mask) => mask,
-                    None => {
-                        valid = false;
-                        wgt::ColorWrite::empty()
-                    },
+                    None => wgt::ColorWrite::empty(),
                 },
             })
             .collect::<ArrayVec<_>>();
diff --git a/components/script/dom/gpupipelinelayout.rs b/components/script/dom/gpupipelinelayout.rs
index af665ba8c41..a331456b14f 100644
--- a/components/script/dom/gpupipelinelayout.rs
+++ b/components/script/dom/gpupipelinelayout.rs
@@ -10,26 +10,20 @@ use crate::dom::bindings::str::DOMString;
 use crate::dom::globalscope::GlobalScope;
 use dom_struct::dom_struct;
 use std::cell::Cell;
-use webgpu::{WebGPUBindGroupLayout, WebGPUPipelineLayout};
+use webgpu::WebGPUPipelineLayout;
 
 #[dom_struct]
 pub struct GPUPipelineLayout {
     reflector_: Reflector,
-    bind_group_layouts: Vec<WebGPUBindGroupLayout>,
     label: DomRefCell<Option<DOMString>>,
     pipeline_layout: WebGPUPipelineLayout,
     valid: Cell<bool>,
 }
 
 impl GPUPipelineLayout {
-    fn new_inherited(
-        bind_group_layouts: Vec<WebGPUBindGroupLayout>,
-        pipeline_layout: WebGPUPipelineLayout,
-        valid: bool,
-    ) -> Self {
+    fn new_inherited(pipeline_layout: WebGPUPipelineLayout, valid: bool) -> Self {
         Self {
             reflector_: Reflector::new(),
-            bind_group_layouts,
             label: DomRefCell::new(None),
             pipeline_layout,
             valid: Cell::new(valid),
@@ -38,16 +32,11 @@ impl GPUPipelineLayout {
 
     pub fn new(
         global: &GlobalScope,
-        bind_group_layouts: Vec<WebGPUBindGroupLayout>,
         pipeline_layout: WebGPUPipelineLayout,
         valid: bool,
     ) -> DomRoot<Self> {
         reflect_dom_object(
-            Box::new(GPUPipelineLayout::new_inherited(
-                bind_group_layouts,
-                pipeline_layout,
-                valid,
-            )),
+            Box::new(GPUPipelineLayout::new_inherited(pipeline_layout, valid)),
             global,
         )
     }
diff --git a/components/script/dom/gpurenderpassencoder.rs b/components/script/dom/gpurenderpassencoder.rs
index a74d1f86c8e..a2d223b6f5d 100644
--- a/components/script/dom/gpurenderpassencoder.rs
+++ b/components/script/dom/gpurenderpassencoder.rs
@@ -19,6 +19,7 @@ use crate::dom::gpurenderpipeline::GPURenderPipeline;
 use dom_struct::dom_struct;
 use webgpu::{
     wgpu::command::{render_ffi as wgpu_render, RawPass},
+    wgpu::id,
     wgt, WebGPU, WebGPURequest,
 };
 
@@ -29,12 +30,16 @@ pub struct GPURenderPassEncoder {
     channel: WebGPU,
     label: DomRefCell<Option<DOMString>>,
     #[ignore_malloc_size_of = "defined in wgpu-core"]
-    raw_pass: DomRefCell<Option<RawPass>>,
+    raw_pass: DomRefCell<Option<RawPass<id::CommandEncoderId>>>,
     command_encoder: Dom<GPUCommandEncoder>,
 }
 
 impl GPURenderPassEncoder {
-    fn new_inherited(channel: WebGPU, raw_pass: RawPass, parent: &GPUCommandEncoder) -> Self {
+    fn new_inherited(
+        channel: WebGPU,
+        raw_pass: RawPass<id::CommandEncoderId>,
+        parent: &GPUCommandEncoder,
+    ) -> Self {
         Self {
             channel,
             reflector_: Reflector::new(),
@@ -47,7 +52,7 @@ impl GPURenderPassEncoder {
     pub fn new(
         global: &GlobalScope,
         channel: WebGPU,
-        raw_pass: RawPass,
+        raw_pass: RawPass<id::CommandEncoderId>,
         parent: &GPUCommandEncoder,
     ) -> DomRoot<Self> {
         reflect_dom_object(
diff --git a/components/script/dom/gpusampler.rs b/components/script/dom/gpusampler.rs
index db85f60e9b6..97e6c1dc37f 100644
--- a/components/script/dom/gpusampler.rs
+++ b/components/script/dom/gpusampler.rs
@@ -10,13 +10,11 @@ use crate::dom::bindings::str::DOMString;
 use crate::dom::globalscope::GlobalScope;
 use dom_struct::dom_struct;
 use std::cell::Cell;
-use webgpu::{WebGPU, WebGPUDevice, WebGPUSampler};
+use webgpu::{WebGPUDevice, WebGPUSampler};
 
 #[dom_struct]
 pub struct GPUSampler {
     reflector_: Reflector,
-    #[ignore_malloc_size_of = "channels are hard"]
-    channel: WebGPU,
     label: DomRefCell<Option<DOMString>>,
     device: WebGPUDevice,
     compare_enable: bool,
@@ -26,7 +24,6 @@ pub struct GPUSampler {
 
 impl GPUSampler {
     fn new_inherited(
-        channel: WebGPU,
         device: WebGPUDevice,
         compare_enable: bool,
         sampler: WebGPUSampler,
@@ -34,7 +31,6 @@ impl GPUSampler {
     ) -> Self {
         Self {
             reflector_: Reflector::new(),
-            channel,
             label: DomRefCell::new(None),
             valid: Cell::new(valid),
             device,
@@ -45,7 +41,6 @@ impl GPUSampler {
 
     pub fn new(
         global: &GlobalScope,
-        channel: WebGPU,
         device: WebGPUDevice,
         compare_enable: bool,
         sampler: WebGPUSampler,
@@ -53,7 +48,6 @@ impl GPUSampler {
     ) -> DomRoot<Self> {
         reflect_dom_object(
             Box::new(GPUSampler::new_inherited(
-                channel,
                 device,
                 compare_enable,
                 sampler,
@@ -64,6 +58,16 @@ impl GPUSampler {
     }
 }
 
+impl GPUSampler {
+    pub fn id(&self) -> WebGPUSampler {
+        self.sampler
+    }
+
+    pub fn is_valid(&self) -> bool {
+        self.valid.get()
+    }
+}
+
 impl GPUSamplerMethods for GPUSampler {
     /// https://gpuweb.github.io/gpuweb/#dom-gpuobjectbase-label
     fn GetLabel(&self) -> Option<DOMString> {
diff --git a/components/script/dom/gputexture.rs b/components/script/dom/gputexture.rs
index 48060efa666..6bf4892887a 100644
--- a/components/script/dom/gputexture.rs
+++ b/components/script/dom/gputexture.rs
@@ -7,7 +7,7 @@ use crate::dom::bindings::codegen::Bindings::GPUTextureBinding::{
     GPUExtent3DDict, GPUTextureDimension, GPUTextureFormat, GPUTextureMethods,
 };
 use crate::dom::bindings::codegen::Bindings::GPUTextureViewBinding::{
-    GPUTextureAspect, GPUTextureViewDescriptor,
+    GPUTextureAspect, GPUTextureViewDescriptor, GPUTextureViewDimension,
 };
 use crate::dom::bindings::reflector::{reflect_dom_object, DomObject, Reflector};
 use crate::dom::bindings::root::DomRoot;
@@ -122,23 +122,28 @@ impl GPUTextureMethods for GPUTexture {
 
     /// https://gpuweb.github.io/gpuweb/#dom-gputexture-createview
     fn CreateView(&self, descriptor: &GPUTextureViewDescriptor) -> DomRoot<GPUTextureView> {
+        let dimension = if let Some(d) = descriptor.dimension {
+            d
+        } else {
+            match self.dimension {
+                GPUTextureDimension::_1d => GPUTextureViewDimension::_1d,
+                GPUTextureDimension::_2d => {
+                    if self.texture_size.depth > 1 && descriptor.arrayLayerCount == 0 {
+                        GPUTextureViewDimension::_2d_array
+                    } else {
+                        GPUTextureViewDimension::_2d
+                    }
+                },
+                GPUTextureDimension::_3d => GPUTextureViewDimension::_3d,
+            }
+        };
+
+        let format = descriptor.format.unwrap_or(self.format);
+
         let desc = wgt::TextureViewDescriptor {
             label: Default::default(),
-            format: convert_texture_format(descriptor.format.unwrap_or(self.format)),
-            dimension: match descriptor.dimension {
-                Some(d) => convert_texture_view_dimension(d),
-                None => match self.dimension {
-                    GPUTextureDimension::_1d => wgt::TextureViewDimension::D1,
-                    GPUTextureDimension::_2d => {
-                        if self.texture_size.depth > 1 && descriptor.arrayLayerCount == 0 {
-                            wgt::TextureViewDimension::D2Array
-                        } else {
-                            wgt::TextureViewDimension::D2
-                        }
-                    },
-                    GPUTextureDimension::_3d => wgt::TextureViewDimension::D3,
-                },
-            },
+            format: convert_texture_format(format),
+            dimension: convert_texture_view_dimension(dimension),
             aspect: match descriptor.aspect {
                 GPUTextureAspect::All => wgt::TextureAspect::All,
                 GPUTextureAspect::Stencil_only => wgt::TextureAspect::StencilOnly,
@@ -175,7 +180,7 @@ impl GPUTextureMethods for GPUTexture {
 
         let texture_view = WebGPUTextureView(texture_view_id);
 
-        GPUTextureView::new(&self.global(), texture_view, self.device, true)
+        GPUTextureView::new(&self.global(), texture_view, &self, self.valid.get())
     }
 
     /// https://gpuweb.github.io/gpuweb/#dom-gputexture-destroy
diff --git a/components/script/dom/gputextureview.rs b/components/script/dom/gputextureview.rs
index f3bb4209708..06cb5a0ddbf 100644
--- a/components/script/dom/gputextureview.rs
+++ b/components/script/dom/gputextureview.rs
@@ -5,27 +5,32 @@
 use crate::dom::bindings::cell::DomRefCell;
 use crate::dom::bindings::codegen::Bindings::GPUTextureViewBinding::GPUTextureViewMethods;
 use crate::dom::bindings::reflector::{reflect_dom_object, Reflector};
-use crate::dom::bindings::root::DomRoot;
+use crate::dom::bindings::root::{Dom, DomRoot};
 use crate::dom::bindings::str::DOMString;
 use crate::dom::globalscope::GlobalScope;
+use crate::dom::gputexture::GPUTexture;
 use dom_struct::dom_struct;
 use std::cell::Cell;
-use webgpu::{WebGPUDevice, WebGPUTextureView};
+use webgpu::WebGPUTextureView;
 
 #[dom_struct]
 pub struct GPUTextureView {
     reflector_: Reflector,
     label: DomRefCell<Option<DOMString>>,
     texture_view: WebGPUTextureView,
-    device: WebGPUDevice,
+    texture: Dom<GPUTexture>,
     valid: Cell<bool>,
 }
 
 impl GPUTextureView {
-    fn new_inherited(texture_view: WebGPUTextureView, device: WebGPUDevice, valid: bool) -> Self {
+    fn new_inherited(
+        texture_view: WebGPUTextureView,
+        texture: &GPUTexture,
+        valid: bool,
+    ) -> GPUTextureView {
         Self {
             reflector_: Reflector::new(),
-            device,
+            texture: Dom::from_ref(texture),
             label: DomRefCell::new(None),
             texture_view,
             valid: Cell::new(valid),
@@ -35,11 +40,11 @@ impl GPUTextureView {
     pub fn new(
         global: &GlobalScope,
         texture_view: WebGPUTextureView,
-        device: WebGPUDevice,
+        texture: &GPUTexture,
         valid: bool,
-    ) -> DomRoot<Self> {
+    ) -> DomRoot<GPUTextureView> {
         reflect_dom_object(
-            Box::new(GPUTextureView::new_inherited(texture_view, device, valid)),
+            Box::new(GPUTextureView::new_inherited(texture_view, texture, valid)),
             global,
         )
     }
@@ -49,6 +54,10 @@ impl GPUTextureView {
     pub fn id(&self) -> WebGPUTextureView {
         self.texture_view
     }
+
+    pub fn is_valid(&self) -> bool {
+        self.valid.get()
+    }
 }
 
 impl GPUTextureViewMethods for GPUTextureView {
diff --git a/components/script/dom/webidls/GPUAdapter.webidl b/components/script/dom/webidls/GPUAdapter.webidl
index 8d2e34f4a57..7fc2e8438f5 100644
--- a/components/script/dom/webidls/GPUAdapter.webidl
+++ b/components/script/dom/webidls/GPUAdapter.webidl
@@ -24,12 +24,13 @@ enum GPUExtensionName {
 };
 
 dictionary GPULimits {
-    unsigned long maxBindGroups = 4;
-    unsigned long maxDynamicUniformBuffersPerPipelineLayout = 8;
-    unsigned long maxDynamicStorageBuffersPerPipelineLayout = 4;
-    unsigned long maxSampledTexturesPerShaderStage = 16;
-    unsigned long maxSamplersPerShaderStage = 16;
-    unsigned long maxStorageBuffersPerShaderStage = 4;
-    unsigned long maxStorageTexturesPerShaderStage = 4;
-    unsigned long maxUniformBuffersPerShaderStage = 12;
+    GPUSize32 maxBindGroups = 4;
+    GPUSize32 maxDynamicUniformBuffersPerPipelineLayout = 8;
+    GPUSize32 maxDynamicStorageBuffersPerPipelineLayout = 4;
+    GPUSize32 maxSampledTexturesPerShaderStage = 16;
+    GPUSize32 maxSamplersPerShaderStage = 16;
+    GPUSize32 maxStorageBuffersPerShaderStage = 4;
+    GPUSize32 maxStorageTexturesPerShaderStage = 4;
+    GPUSize32 maxUniformBuffersPerShaderStage = 12;
+    GPUSize32 maxUniformBufferBindingSize = 16384;
 };
diff --git a/components/script/dom/webidls/GPUBindGroup.webidl b/components/script/dom/webidls/GPUBindGroup.webidl
index f905755d6fb..7bfa023acba 100644
--- a/components/script/dom/webidls/GPUBindGroup.webidl
+++ b/components/script/dom/webidls/GPUBindGroup.webidl
@@ -3,7 +3,7 @@
  * file, You can obtain one at https://mozilla.org/MPL/2.0/. */
 
 // https://gpuweb.github.io/gpuweb/#gpubindgrouplayout
-[Exposed=(Window, DedicatedWorker), Serializable, Pref="dom.webgpu.enabled"]
+[Exposed=(Window, DedicatedWorker), Pref="dom.webgpu.enabled"]
 interface GPUBindGroup {
 };
 GPUBindGroup includes GPUObjectBase;
@@ -13,10 +13,10 @@ dictionary GPUBindGroupDescriptor : GPUObjectDescriptorBase {
     required sequence<GPUBindGroupEntry> entries;
 };
 
-typedef /*(GPUSampler or GPUTextureView or*/ GPUBufferBindings/*)*/ GPUBindingResource;
+typedef (GPUSampler or GPUTextureView or GPUBufferBindings) GPUBindingResource;
 
 dictionary GPUBindGroupEntry {
-    required unsigned long binding;
+    required GPUIndex32 binding;
     required GPUBindingResource resource;
 };
 
diff --git a/components/script/dom/webidls/GPUBindGroupLayout.webidl b/components/script/dom/webidls/GPUBindGroupLayout.webidl
index 8b2d63a84ff..6413afa6032 100644
--- a/components/script/dom/webidls/GPUBindGroupLayout.webidl
+++ b/components/script/dom/webidls/GPUBindGroupLayout.webidl
@@ -16,11 +16,12 @@ dictionary GPUBindGroupLayoutEntry {
     required GPUIndex32 binding;
     required GPUShaderStageFlags visibility;
     required GPUBindingType type;
-    GPUTextureViewDimension viewDimension = "2d";
-    GPUTextureComponentType textureComponentType = "float";
-    GPUTextureFormat storageTextureFormat;
-    boolean multisampled = false;
     boolean hasDynamicOffset = false;
+    GPUSize64 minBufferBindingSize = 0;
+    GPUTextureViewDimension viewDimension;
+    GPUTextureComponentType textureComponentType;
+    boolean multisampled = false;
+    GPUTextureFormat storageTextureFormat;
 };
 
 enum GPUBindingType {
@@ -28,8 +29,8 @@ enum GPUBindingType {
     "storage-buffer",
     "readonly-storage-buffer",
     "sampler",
+    "comparison-sampler",
     "sampled-texture",
     "readonly-storage-texture",
-    "writeonly-storage-texture",
-    //"comparison-sampler",
+    "writeonly-storage-texture"
 };
diff --git a/components/webgpu/identity.rs b/components/webgpu/identity.rs
index 2b40ff5a9bc..9ed149eb152 100644
--- a/components/webgpu/identity.rs
+++ b/components/webgpu/identity.rs
@@ -8,8 +8,8 @@ use wgpu::{
     hub::{GlobalIdentityHandlerFactory, IdentityHandler, IdentityHandlerFactory},
     id::{
         AdapterId, BindGroupId, BindGroupLayoutId, BufferId, CommandBufferId, ComputePipelineId,
-        DeviceId, PipelineLayoutId, RenderPipelineId, SamplerId, ShaderModuleId, SurfaceId,
-        SwapChainId, TextureId, TextureViewId, TypedId,
+        DeviceId, PipelineLayoutId, RenderBundleId, RenderPipelineId, SamplerId, ShaderModuleId,
+        SurfaceId, SwapChainId, TextureId, TextureViewId, TypedId,
     },
 };
 use wgt::Backend;
@@ -31,6 +31,7 @@ pub enum WebGPUMsg {
     FreeSampler(SamplerId),
     FreeSurface(SurfaceId),
     FreeShaderModule(ShaderModuleId),
+    FreeRenderBundle(RenderBundleId),
     Exit,
 }
 
@@ -73,6 +74,7 @@ impl_identity_handler!(BufferId, "buffer", WebGPUMsg::FreeBuffer);
 impl_identity_handler!(BindGroupId, "bind_group", WebGPUMsg::FreeBindGroup);
 impl_identity_handler!(SwapChainId, "swap_chain", WebGPUMsg::FreeSwapChain);
 impl_identity_handler!(ShaderModuleId, "shader_module", WebGPUMsg::FreeShaderModule);
+impl_identity_handler!(RenderBundleId, "render_bundle", WebGPUMsg::FreeRenderBundle);
 impl_identity_handler!(
     RenderPipelineId,
     "render_pipeline",
diff --git a/components/webgpu/lib.rs b/components/webgpu/lib.rs
index 9f04b2bd58b..800a248a4a6 100644
--- a/components/webgpu/lib.rs
+++ b/components/webgpu/lib.rs
@@ -28,9 +28,7 @@ use webrender_traits::{
     WebrenderImageSource,
 };
 use wgpu::{
-    binding_model::{
-        BindGroupDescriptor, BindGroupEntry, BindGroupLayoutDescriptor, BindGroupLayoutEntry,
-    },
+    binding_model::{BindGroupDescriptor, BindGroupEntry, BindingResource, BufferBinding},
     command::{BufferCopyView, TextureCopyView},
     device::HostMap,
     id,
@@ -73,12 +71,12 @@ pub enum WebGPURequest {
         device_id: id::DeviceId,
         bind_group_id: id::BindGroupId,
         bind_group_layout_id: id::BindGroupLayoutId,
-        bindings: Vec<BindGroupEntry>,
+        entries: Vec<(u32, WebGPUBindings)>,
     },
     CreateBindGroupLayout {
         device_id: id::DeviceId,
         bind_group_layout_id: id::BindGroupLayoutId,
-        bindings: Vec<BindGroupLayoutEntry>,
+        entries: Vec<wgt::BindGroupLayoutEntry>,
     },
     CreateBuffer {
         device_id: id::DeviceId,
@@ -195,6 +193,13 @@ pub enum WebGPURequest {
     },
 }
 
+#[derive(Debug, Deserialize, Serialize)]
+pub enum WebGPUBindings {
+    Buffer(BufferBinding),
+    Sampler(id::SamplerId),
+    TextureView(id::TextureViewId),
+}
+
 #[derive(Clone, Debug, Deserialize, Serialize)]
 pub struct WebGPU(pub IpcSender<WebGPURequest>);
 
@@ -334,14 +339,27 @@ impl WGPU {
                     device_id,
                     bind_group_id,
                     bind_group_layout_id,
-                    bindings,
+                    mut entries,
                 } => {
                     let global = &self.global;
+                    let bindings = entries
+                        .drain(..)
+                        .map(|(bind, res)| {
+                            let resource = match res {
+                                WebGPUBindings::Sampler(s) => BindingResource::Sampler(s),
+                                WebGPUBindings::TextureView(t) => BindingResource::TextureView(t),
+                                WebGPUBindings::Buffer(b) => BindingResource::Buffer(b),
+                            };
+                            BindGroupEntry {
+                                binding: bind,
+                                resource,
+                            }
+                        })
+                        .collect::<Vec<_>>();
                     let descriptor = BindGroupDescriptor {
+                        label: None,
                         layout: bind_group_layout_id,
-                        entries: bindings.as_ptr(),
-                        entries_length: bindings.len(),
-                        label: ptr::null(),
+                        bindings: bindings.as_slice(),
                     };
                     let _ = gfx_select!(bind_group_id =>
                         global.device_create_bind_group(device_id, &descriptor, bind_group_id));
@@ -349,13 +367,12 @@ impl WGPU {
                 WebGPURequest::CreateBindGroupLayout {
                     device_id,
                     bind_group_layout_id,
-                    bindings,
+                    entries,
                 } => {
                     let global = &self.global;
-                    let descriptor = BindGroupLayoutDescriptor {
-                        entries: bindings.as_ptr(),
-                        entries_length: bindings.len(),
-                        label: ptr::null(),
+                    let descriptor = wgt::BindGroupLayoutDescriptor {
+                        bindings: entries.as_slice(),
+                        label: None,
                     };
                     let _ = gfx_select!(bind_group_layout_id =>
                         global.device_create_bind_group_layout(device_id, &descriptor, bind_group_layout_id));
@@ -375,18 +392,9 @@ impl WGPU {
                     command_encoder_id,
                 } => {
                     let global = &self.global;
+                    let desc = wgt::CommandEncoderDescriptor { label: ptr::null() };
                     let _ = gfx_select!(command_encoder_id =>
-                        global.device_create_command_encoder(device_id, &Default::default(), command_encoder_id));
-                },
-                WebGPURequest::CreateContext(sender) => {
-                    let id = self
-                        .external_images
-                        .lock()
-                        .expect("Lock poisoned?")
-                        .next_id(WebrenderImageHandlerType::WebGPU);
-                    if let Err(e) = sender.send(id) {
-                        warn!("Failed to send ExternalImageId to new context ({})", e);
-                    };
+                        global.device_create_command_encoder(device_id, &desc, command_encoder_id));
                 },
                 WebGPURequest::CreateComputePipeline {
                     device_id,
@@ -407,6 +415,16 @@ impl WGPU {
                     let _ = gfx_select!(compute_pipeline_id =>
                         global.device_create_compute_pipeline(device_id, &descriptor, compute_pipeline_id));
                 },
+                WebGPURequest::CreateContext(sender) => {
+                    let id = self
+                        .external_images
+                        .lock()
+                        .expect("Lock poisoned?")
+                        .next_id(WebrenderImageHandlerType::WebGPU);
+                    if let Err(e) = sender.send(id) {
+                        warn!("Failed to send ExternalImageId to new context ({})", e);
+                    };
+                },
                 WebGPURequest::CreatePipelineLayout {
                     device_id,
                     pipeline_layout_id,
@@ -631,6 +649,7 @@ impl WGPU {
                 } => {
                     let adapter_id = match self.global.pick_adapter(
                         &options,
+                        wgt::UnsafeExtensions::disallow(),
                         wgpu::instance::AdapterInputs::IdSet(&ids, |id| id.backend()),
                     ) {
                         Some(id) => id,
@@ -749,9 +768,10 @@ impl WGPU {
                     }
 
                     let buffer_size = (size.height as u32 * buffer_stride) as wgt::BufferAddress;
+                    let comm_desc = wgt::CommandEncoderDescriptor { label: ptr::null() };
                     let _ = gfx_select!(encoder_id => global.device_create_command_encoder(
                         device_id,
-                        &wgt::CommandEncoderDescriptor::default(),
+                        &comm_desc,
                         encoder_id
                     ));