diff --git a/dom/webgpu/Adapter.cpp b/dom/webgpu/Adapter.cpp
index 015ad734dbfe..2954ef523add 100644
--- a/dom/webgpu/Adapter.cpp
+++ b/dom/webgpu/Adapter.cpp
@@ -18,7 +18,7 @@ GPU_IMPL_CYCLE_COLLECTION(Adapter, mParent, mBridge)
 GPU_IMPL_JS_WRAP(Adapter)
 
 Adapter::Adapter(Instance* const aParent, RawId aId)
-    : ChildOf(aParent), mBridge(aParent->GetBridge()), mId(aId) {}
+    : ChildOf(aParent), mBridge(aParent->mBridge), mId(aId) {}
 
 Adapter::~Adapter() { Cleanup(); }
 
@@ -29,8 +29,6 @@ void Adapter::Cleanup() {
   }
 }
 
-WebGPUChild* Adapter::GetBridge() const { return mBridge; }
-
 already_AddRefed<dom::Promise> Adapter::RequestDevice(
     const dom::GPUDeviceDescriptor& aDesc, ErrorResult& aRv) {
   RefPtr<dom::Promise> promise = dom::Promise::Create(GetParentObject(), aRv);
diff --git a/dom/webgpu/Adapter.h b/dom/webgpu/Adapter.h
index 942e17a0b4d1..37853e57efbd 100644
--- a/dom/webgpu/Adapter.h
+++ b/dom/webgpu/Adapter.h
@@ -29,19 +29,19 @@ class Adapter final : public ObjectBase, public ChildOf<Instance> {
   GPU_DECL_CYCLE_COLLECTION(Adapter)
   GPU_DECL_JS_WRAP(Adapter)
 
+  const RefPtr<WebGPUChild> mBridge;
+
  private:
   Adapter() = delete;
   ~Adapter();
   void Cleanup();
 
-  const RefPtr<WebGPUChild> mBridge;
   const RawId mId;
   const nsString mName;
 
  public:
   explicit Adapter(Instance* const aParent, RawId aId);
   void GetName(nsString& out) const { out = mName; }
-  WebGPUChild* GetBridge() const;
 
   already_AddRefed<dom::Promise> RequestDevice(
       const dom::GPUDeviceDescriptor& aDesc, ErrorResult& aRv);
diff --git a/dom/webgpu/BindGroup.cpp b/dom/webgpu/BindGroup.cpp
index f147859607eb..c3b2d4237ed7 100644
--- a/dom/webgpu/BindGroup.cpp
+++ b/dom/webgpu/BindGroup.cpp
@@ -14,7 +14,20 @@ namespace webgpu {
 GPU_IMPL_CYCLE_COLLECTION(BindGroup, mParent)
 GPU_IMPL_JS_WRAP(BindGroup)
 
-void BindGroup::Cleanup() {}
+BindGroup::BindGroup(Device* const aParent, RawId aId)
+    : ChildOf(aParent), mId(aId) {}
+
+BindGroup::~BindGroup() { Cleanup(); }
+
+void BindGroup::Cleanup() {
+  if (mValid && mParent) {
+    mValid = false;
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyBindGroup(mId);
+    }
+  }
+}
 
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/BindGroup.h b/dom/webgpu/BindGroup.h
index 926b1164e70d..8ebc2f3754a1 100644
--- a/dom/webgpu/BindGroup.h
+++ b/dom/webgpu/BindGroup.h
@@ -8,6 +8,7 @@
 
 #include "nsWrapperCache.h"
 #include "ObjectModel.h"
+#include "mozilla/webgpu/WebGPUTypes.h"
 
 namespace mozilla {
 namespace webgpu {
@@ -19,9 +20,12 @@ class BindGroup final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(BindGroup)
   GPU_DECL_JS_WRAP(BindGroup)
 
+  BindGroup(Device* const aParent, RawId aId);
+
+  const RawId mId;
+
  private:
-  BindGroup() = delete;
-  ~BindGroup() = default;
+  ~BindGroup();
   void Cleanup();
 };
 
diff --git a/dom/webgpu/BindGroupLayout.cpp b/dom/webgpu/BindGroupLayout.cpp
index ff9a574032a0..a083ae1f8506 100644
--- a/dom/webgpu/BindGroupLayout.cpp
+++ b/dom/webgpu/BindGroupLayout.cpp
@@ -14,7 +14,20 @@ namespace webgpu {
 GPU_IMPL_CYCLE_COLLECTION(BindGroupLayout, mParent)
 GPU_IMPL_JS_WRAP(BindGroupLayout)
 
-void BindGroupLayout::Cleanup() {}
+BindGroupLayout::BindGroupLayout(Device* const aParent, RawId aId)
+    : ChildOf(aParent), mId(aId) {}
+
+BindGroupLayout::~BindGroupLayout() { Cleanup(); }
+
+void BindGroupLayout::Cleanup() {
+  if (mValid && mParent) {
+    mValid = false;
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyBindGroupLayout(mId);
+    }
+  }
+}
 
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/BindGroupLayout.h b/dom/webgpu/BindGroupLayout.h
index 4b6ab7e85027..c9d2e2e52b31 100644
--- a/dom/webgpu/BindGroupLayout.h
+++ b/dom/webgpu/BindGroupLayout.h
@@ -8,6 +8,7 @@
 
 #include "nsWrapperCache.h"
 #include "ObjectModel.h"
+#include "mozilla/webgpu/WebGPUTypes.h"
 
 namespace mozilla {
 namespace webgpu {
@@ -19,9 +20,12 @@ class BindGroupLayout final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(BindGroupLayout)
   GPU_DECL_JS_WRAP(BindGroupLayout)
 
+  BindGroupLayout(Device* const aParent, RawId aId);
+
+  const RawId mId;
+
  private:
-  BindGroupLayout() = delete;
-  ~BindGroupLayout() = default;
+  ~BindGroupLayout();
   void Cleanup();
 };
 
diff --git a/dom/webgpu/Buffer.cpp b/dom/webgpu/Buffer.cpp
index faeac93ad310..13fd668585fa 100644
--- a/dom/webgpu/Buffer.cpp
+++ b/dom/webgpu/Buffer.cpp
@@ -51,7 +51,10 @@ Buffer::~Buffer() {
 
 void Buffer::Cleanup() {
   if (mParent) {
-    mParent->DestroyBuffer(mId);
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyBuffer(mId);
+    }
   }
   mMapping.reset();
 }
diff --git a/dom/webgpu/Buffer.h b/dom/webgpu/Buffer.h
index 7a9d94ebf8a5..65597c7e577c 100644
--- a/dom/webgpu/Buffer.h
+++ b/dom/webgpu/Buffer.h
@@ -34,11 +34,12 @@ class Buffer final : public ObjectBase, public ChildOf<Device> {
   Buffer(Device* const aParent, RawId aId, BufferAddress aSize);
   void InitMapping(ipc::Shmem&& aShmem, JSObject* aArrayBuffer);
 
+  const RawId mId;
+
  private:
   virtual ~Buffer();
   void Cleanup();
 
-  const RawId mId;
   // Note: we can't map a buffer with the size that don't fit into `size_t`
   // (which may be smaller than `BufferAddress`), but general not all buffers
   // are mapped.
diff --git a/dom/webgpu/CommandBuffer.cpp b/dom/webgpu/CommandBuffer.cpp
index a111af0572f3..2b8f5a27adfa 100644
--- a/dom/webgpu/CommandBuffer.cpp
+++ b/dom/webgpu/CommandBuffer.cpp
@@ -26,7 +26,10 @@ CommandBuffer::~CommandBuffer() { Cleanup(); }
 void CommandBuffer::Cleanup() {
   if (mValid && mParent) {
     mValid = false;
-    mParent->DestroyCommandBuffer(mId);
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyCommandBuffer(mId);
+    }
   }
 }
 
diff --git a/dom/webgpu/CommandEncoder.cpp b/dom/webgpu/CommandEncoder.cpp
index 2be1bfb43541..3bec54e849a9 100644
--- a/dom/webgpu/CommandEncoder.cpp
+++ b/dom/webgpu/CommandEncoder.cpp
@@ -7,6 +7,7 @@
 #include "CommandEncoder.h"
 
 #include "CommandBuffer.h"
+#include "Buffer.h"
 #include "ComputePassEncoder.h"
 #include "Device.h"
 
@@ -18,14 +19,29 @@ GPU_IMPL_JS_WRAP(CommandEncoder)
 
 CommandEncoder::CommandEncoder(Device* const aParent,
                                WebGPUChild* const aBridge, RawId aId)
-    : ChildOf(aParent), mBridge(aBridge), mId(aId) {}
+    : ChildOf(aParent), mId(aId), mBridge(aBridge) {}
 
 CommandEncoder::~CommandEncoder() { Cleanup(); }
 
 void CommandEncoder::Cleanup() {
   if (mValid && mParent) {
     mValid = false;
-    mParent->DestroyCommandEncoder(mId);
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyCommandEncoder(mId);
+    }
+  }
+}
+
+void CommandEncoder::CopyBufferToBuffer(const Buffer& aSource,
+                                        BufferAddress aSourceOffset,
+                                        const Buffer& aDestination,
+                                        BufferAddress aDestinationOffset,
+                                        BufferAddress aSize) {
+  if (mValid) {
+    mBridge->SendCommandEncoderCopyBufferToBuffer(
+        mId, aSource.mId, aSourceOffset, aDestination.mId, aDestinationOffset,
+        aSize);
   }
 }
 
diff --git a/dom/webgpu/CommandEncoder.h b/dom/webgpu/CommandEncoder.h
index 448789c74f46..6d0ead3e6d35 100644
--- a/dom/webgpu/CommandEncoder.h
+++ b/dom/webgpu/CommandEncoder.h
@@ -38,20 +38,26 @@ class CommandEncoder final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_JS_WRAP(CommandEncoder)
 
   CommandEncoder(Device* const aParent, WebGPUChild* const aBridge, RawId aId);
-  RawId GetId() const { return mId; }
-  void EndComputePass(Span<const uint8_t> aData, ErrorResult& aRv);
 
-  already_AddRefed<ComputePassEncoder> BeginComputePass(
-      const dom::GPUComputePassDescriptor& aDesc);
-  already_AddRefed<CommandBuffer> Finish(
-      const dom::GPUCommandBufferDescriptor& aDesc);
+  const RawId mId;
 
  private:
   ~CommandEncoder();
   void Cleanup();
 
   const RefPtr<WebGPUChild> mBridge;
-  const RawId mId;
+
+ public:
+  void EndComputePass(Span<const uint8_t> aData, ErrorResult& aRv);
+
+  void CopyBufferToBuffer(const Buffer& aSource, BufferAddress aSourceOffset,
+                          const Buffer& aDestination,
+                          BufferAddress aDestinationOffset,
+                          BufferAddress aSize);
+  already_AddRefed<ComputePassEncoder> BeginComputePass(
+      const dom::GPUComputePassDescriptor& aDesc);
+  already_AddRefed<CommandBuffer> Finish(
+      const dom::GPUCommandBufferDescriptor& aDesc);
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/ComputePassEncoder.cpp b/dom/webgpu/ComputePassEncoder.cpp
index 6624339620ab..7459fad68a6a 100644
--- a/dom/webgpu/ComputePassEncoder.cpp
+++ b/dom/webgpu/ComputePassEncoder.cpp
@@ -5,6 +5,8 @@
 
 #include "mozilla/dom/WebGPUBinding.h"
 #include "ComputePassEncoder.h"
+#include "BindGroup.h"
+#include "ComputePipeline.h"
 
 #include "mozilla/webgpu/ffi/wgpu.h"
 
@@ -29,7 +31,7 @@ ffi::WGPURawPass BeginComputePass(RawId aEncoderId,
 
 ComputePassEncoder::ComputePassEncoder(
     CommandEncoder* const aParent, const dom::GPUComputePassDescriptor& aDesc)
-    : ChildOf(aParent), mRaw(BeginComputePass(aParent->GetId(), aDesc)) {}
+    : ChildOf(aParent), mRaw(BeginComputePass(aParent->mId, aDesc)) {}
 
 ComputePassEncoder::~ComputePassEncoder() {
   if (mValid) {
@@ -38,6 +40,22 @@ ComputePassEncoder::~ComputePassEncoder() {
   }
 }
 
+void ComputePassEncoder::SetBindGroup(
+    uint32_t aSlot, const BindGroup& aBindGroup,
+    const dom::Sequence<uint32_t>& aDynamicOffsets) {
+  if (mValid) {
+    ffi::wgpu_compute_pass_set_bind_group(&mRaw, aSlot, aBindGroup.mId,
+                                          aDynamicOffsets.Elements(),
+                                          aDynamicOffsets.Length());
+  }
+}
+
+void ComputePassEncoder::SetPipeline(const ComputePipeline& aPipeline) {
+  if (mValid) {
+    ffi::wgpu_compute_pass_set_pipeline(&mRaw, aPipeline.mId);
+  }
+}
+
 void ComputePassEncoder::Dispatch(uint32_t x, uint32_t y, uint32_t z) {
   if (mValid) {
     ffi::wgpu_compute_pass_dispatch(&mRaw, x, y, z);
diff --git a/dom/webgpu/ComputePassEncoder.h b/dom/webgpu/ComputePassEncoder.h
index 28a783f8bebc..461c4f8fb8e2 100644
--- a/dom/webgpu/ComputePassEncoder.h
+++ b/dom/webgpu/ComputePassEncoder.h
@@ -37,6 +37,9 @@ class ComputePassEncoder final : public ProgrammablePassEncoder,
   ffi::WGPURawPass mRaw;
 
  public:
+  void SetBindGroup(uint32_t aSlot, const BindGroup& aBindGroup,
+                    const dom::Sequence<uint32_t>& aDynamicOffsets) override;
+  void SetPipeline(const ComputePipeline& aPipeline);
   void Dispatch(uint32_t x, uint32_t y, uint32_t z);
   void EndPass(ErrorResult& aRv);
 };
diff --git a/dom/webgpu/ComputePipeline.cpp b/dom/webgpu/ComputePipeline.cpp
index 65e9151c0dc8..07af34d92b20 100644
--- a/dom/webgpu/ComputePipeline.cpp
+++ b/dom/webgpu/ComputePipeline.cpp
@@ -13,5 +13,20 @@ namespace webgpu {
 GPU_IMPL_CYCLE_COLLECTION(ComputePipeline, mParent)
 GPU_IMPL_JS_WRAP(ComputePipeline)
 
+ComputePipeline::ComputePipeline(Device* const aParent, RawId aId)
+    : ChildOf(aParent), mId(aId) {}
+
+ComputePipeline::~ComputePipeline() { Cleanup(); }
+
+void ComputePipeline::Cleanup() {
+  if (mValid && mParent) {
+    mValid = false;
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyComputePipeline(mId);
+    }
+  }
+}
+
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/ComputePipeline.h b/dom/webgpu/ComputePipeline.h
index c9d9fae81cc9..02726c5980f2 100644
--- a/dom/webgpu/ComputePipeline.h
+++ b/dom/webgpu/ComputePipeline.h
@@ -19,10 +19,13 @@ class ComputePipeline final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(ComputePipeline)
   GPU_DECL_JS_WRAP(ComputePipeline)
 
+  ComputePipeline(Device* const aParent, RawId aId);
+
+  const RawId mId;
+
  private:
-  ComputePipeline() = delete;
-  ~ComputePipeline() = default;
-  void Cleanup() {}
+  ~ComputePipeline();
+  void Cleanup();
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/Device.cpp b/dom/webgpu/Device.cpp
index a8bca0703374..54714dbbaa94 100644
--- a/dom/webgpu/Device.cpp
+++ b/dom/webgpu/Device.cpp
@@ -38,9 +38,9 @@ JSObject* Device::CreateExternalArrayBuffer(JSContext* aCx, size_t aSize,
 
 Device::Device(Adapter* const aParent, RawId aId)
     : DOMEventTargetHelper(aParent->GetParentObject()),
-      mBridge(aParent->GetBridge()),
+      mBridge(aParent->mBridge),
       mId(aId),
-      mQueue(new Queue(this, aParent->GetBridge(), aId)) {}
+      mQueue(new Queue(this, aParent->mBridge, aId)) {}
 
 Device::~Device() { Cleanup(); }
 
@@ -129,12 +129,6 @@ void Device::UnmapBuffer(RawId aId, UniquePtr<ipc::Shmem> aShmem) {
   mBridge->SendDeviceUnmapBuffer(mId, aId, std::move(*aShmem));
 }
 
-void Device::DestroyBuffer(RawId aId) {
-  if (mBridge && mBridge->IsOpen()) {
-    mBridge->DestroyBuffer(aId);
-  }
-}
-
 already_AddRefed<CommandEncoder> Device::CreateCommandEncoder(
     const dom::GPUCommandEncoderDescriptor& aDesc) {
   RawId id = mBridge->DeviceCreateCommandEncoder(mId, aDesc);
@@ -142,16 +136,37 @@ already_AddRefed<CommandEncoder> Device::CreateCommandEncoder(
   return encoder.forget();
 }
 
-void Device::DestroyCommandEncoder(RawId aId) {
-  if (mBridge && mBridge->IsOpen()) {
-    mBridge->DestroyCommandEncoder(aId);
-  }
+already_AddRefed<BindGroupLayout> Device::CreateBindGroupLayout(
+    const dom::GPUBindGroupLayoutDescriptor& aDesc) {
+  RawId id = mBridge->DeviceCreateBindGroupLayout(mId, aDesc);
+  RefPtr<BindGroupLayout> object = new BindGroupLayout(this, id);
+  return object.forget();
+}
+already_AddRefed<PipelineLayout> Device::CreatePipelineLayout(
+    const dom::GPUPipelineLayoutDescriptor& aDesc) {
+  RawId id = mBridge->DeviceCreatePipelineLayout(mId, aDesc);
+  RefPtr<PipelineLayout> object = new PipelineLayout(this, id);
+  return object.forget();
+}
+already_AddRefed<BindGroup> Device::CreateBindGroup(
+    const dom::GPUBindGroupDescriptor& aDesc) {
+  RawId id = mBridge->DeviceCreateBindGroup(mId, aDesc);
+  RefPtr<BindGroup> object = new BindGroup(this, id);
+  return object.forget();
 }
 
-void Device::DestroyCommandBuffer(RawId aId) {
-  if (mBridge && mBridge->IsOpen()) {
-    mBridge->DestroyCommandBuffer(aId);
-  }
+already_AddRefed<ShaderModule> Device::CreateShaderModule(
+    const dom::GPUShaderModuleDescriptor& aDesc) {
+  RawId id = mBridge->DeviceCreateShaderModule(mId, aDesc);
+  RefPtr<ShaderModule> object = new ShaderModule(this, id);
+  return object.forget();
+}
+
+already_AddRefed<ComputePipeline> Device::CreateComputePipeline(
+    const dom::GPUComputePipelineDescriptor& aDesc) {
+  RawId id = mBridge->DeviceCreateComputePipeline(mId, aDesc);
+  RefPtr<ComputePipeline> object = new ComputePipeline(this, id);
+  return object.forget();
 }
 
 }  // namespace webgpu
diff --git a/dom/webgpu/Device.h b/dom/webgpu/Device.h
index fcc7cae4f27d..935de12cb016 100644
--- a/dom/webgpu/Device.h
+++ b/dom/webgpu/Device.h
@@ -79,15 +79,13 @@ class Device final : public DOMEventTargetHelper {
   RefPtr<MappingPromise> MapBufferForReadAsync(RawId aId, size_t aSize,
                                                ErrorResult& aRv);
   void UnmapBuffer(RawId aId, UniquePtr<ipc::Shmem> aShmem);
-  void DestroyBuffer(RawId aId);
-  void DestroyCommandEncoder(RawId aId);
-  void DestroyCommandBuffer(RawId aId);
+
+  const RefPtr<WebGPUChild> mBridge;
 
  private:
   ~Device();
   void Cleanup();
 
-  const RefPtr<WebGPUChild> mBridge;
   const RawId mId;
   bool mValid = true;
   nsString mLabel;
@@ -105,6 +103,19 @@ class Device final : public DOMEventTargetHelper {
 
   already_AddRefed<CommandEncoder> CreateCommandEncoder(
       const dom::GPUCommandEncoderDescriptor& aDesc);
+
+  already_AddRefed<BindGroupLayout> CreateBindGroupLayout(
+      const dom::GPUBindGroupLayoutDescriptor& aDesc);
+  already_AddRefed<PipelineLayout> CreatePipelineLayout(
+      const dom::GPUPipelineLayoutDescriptor& aDesc);
+  already_AddRefed<BindGroup> CreateBindGroup(
+      const dom::GPUBindGroupDescriptor& aDesc);
+
+  already_AddRefed<ShaderModule> CreateShaderModule(
+      const dom::GPUShaderModuleDescriptor& aDesc);
+  already_AddRefed<ComputePipeline> CreateComputePipeline(
+      const dom::GPUComputePipelineDescriptor& aDesc);
+
   // IMPL_EVENT_HANDLER(uncapturederror)
 };
 
diff --git a/dom/webgpu/Instance.cpp b/dom/webgpu/Instance.cpp
index 33e874f9c842..de3b916e91db 100644
--- a/dom/webgpu/Instance.cpp
+++ b/dom/webgpu/Instance.cpp
@@ -15,7 +15,7 @@
 namespace mozilla {
 namespace webgpu {
 
-GPU_IMPL_CYCLE_COLLECTION(Instance, mOwner, mBridge)
+GPU_IMPL_CYCLE_COLLECTION(Instance, mBridge, mOwner)
 
 /*static*/
 already_AddRefed<Instance> Instance::Create(nsIGlobalObject* aOwner) {
@@ -34,7 +34,7 @@ already_AddRefed<Instance> Instance::Create(nsIGlobalObject* aOwner) {
 }
 
 Instance::Instance(nsIGlobalObject* aOwner, WebGPUChild* aBridge)
-    : mOwner(aOwner), mBridge(aBridge) {}
+    : mBridge(aBridge), mOwner(aOwner) {}
 
 Instance::~Instance() { Cleanup(); }
 
@@ -45,8 +45,6 @@ JSObject* Instance::WrapObject(JSContext* cx,
   return dom::GPU_Binding::Wrap(cx, this, givenProto);
 }
 
-WebGPUChild* Instance::GetBridge() const { return mBridge; }
-
 already_AddRefed<dom::Promise> Instance::RequestAdapter(
     const dom::GPURequestAdapterOptions& aOptions, ErrorResult& aRv) {
   RefPtr<dom::Promise> promise = dom::Promise::Create(mOwner, aRv);
diff --git a/dom/webgpu/Instance.h b/dom/webgpu/Instance.h
index a90285af930a..91e61cdb1885 100644
--- a/dom/webgpu/Instance.h
+++ b/dom/webgpu/Instance.h
@@ -27,22 +27,23 @@ class Instance final : public nsWrapperCache {
   GPU_DECL_CYCLE_COLLECTION(Instance)
   GPU_DECL_JS_WRAP(Instance)
 
+  nsIGlobalObject* GetParentObject() const { return mOwner; }
+
   static already_AddRefed<Instance> Create(nsIGlobalObject* aOwner);
 
   already_AddRefed<dom::Promise> RequestAdapter(
       const dom::GPURequestAdapterOptions& aOptions, ErrorResult& aRv);
 
+  const RefPtr<WebGPUChild> mBridge;
+
  private:
   explicit Instance(nsIGlobalObject* aOwner, WebGPUChild* aBridge);
   virtual ~Instance();
   void Cleanup();
 
   nsCOMPtr<nsIGlobalObject> mOwner;
-  const RefPtr<WebGPUChild> mBridge;
 
  public:
-  nsIGlobalObject* GetParentObject() const { return mOwner; }
-  WebGPUChild* GetBridge() const;
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/PipelineLayout.cpp b/dom/webgpu/PipelineLayout.cpp
index 118e773ee681..f590702e5c28 100644
--- a/dom/webgpu/PipelineLayout.cpp
+++ b/dom/webgpu/PipelineLayout.cpp
@@ -14,7 +14,20 @@ namespace webgpu {
 GPU_IMPL_CYCLE_COLLECTION(PipelineLayout, mParent)
 GPU_IMPL_JS_WRAP(PipelineLayout)
 
-PipelineLayout::~PipelineLayout() = default;
+PipelineLayout::PipelineLayout(Device* const aParent, RawId aId)
+    : ChildOf(aParent), mId(aId) {}
+
+PipelineLayout::~PipelineLayout() { Cleanup(); }
+
+void PipelineLayout::Cleanup() {
+  if (mValid && mParent) {
+    mValid = false;
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyPipelineLayout(mId);
+    }
+  }
+}
 
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/PipelineLayout.h b/dom/webgpu/PipelineLayout.h
index e88dc3b11815..10746c32d6b6 100644
--- a/dom/webgpu/PipelineLayout.h
+++ b/dom/webgpu/PipelineLayout.h
@@ -8,6 +8,7 @@
 
 #include "nsWrapperCache.h"
 #include "ObjectModel.h"
+#include "mozilla/webgpu/WebGPUTypes.h"
 
 namespace mozilla {
 namespace webgpu {
@@ -19,10 +20,13 @@ class PipelineLayout final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(PipelineLayout)
   GPU_DECL_JS_WRAP(PipelineLayout)
 
+  PipelineLayout(Device* const aParent, RawId aId);
+
+  const RawId mId;
+
  private:
-  PipelineLayout() = delete;
   virtual ~PipelineLayout();
-  void Cleanup() {}
+  void Cleanup();
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/ProgrammablePassEncoder.cpp b/dom/webgpu/ProgrammablePassEncoder.cpp
index b8331aa8f7be..6811dc1de267 100644
--- a/dom/webgpu/ProgrammablePassEncoder.cpp
+++ b/dom/webgpu/ProgrammablePassEncoder.cpp
@@ -4,6 +4,7 @@
  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
 
 #include "ProgrammablePassEncoder.h"
+#include "BindGroup.h"
 
 namespace mozilla {
 namespace webgpu {
diff --git a/dom/webgpu/ProgrammablePassEncoder.h b/dom/webgpu/ProgrammablePassEncoder.h
index 5d28395060fb..5e78cb89cd2e 100644
--- a/dom/webgpu/ProgrammablePassEncoder.h
+++ b/dom/webgpu/ProgrammablePassEncoder.h
@@ -13,8 +13,6 @@ namespace mozilla {
 namespace dom {
 template <typename T>
 class Sequence;
-template <typename T>
-class Optional;
 }  // namespace dom
 namespace webgpu {
 
@@ -34,6 +32,12 @@ class ProgrammablePassEncoder : public nsISupports, public ObjectBase {
  protected:
   ~ProgrammablePassEncoder() = default;
   void Cleanup() {}
+
+ public:
+  // TODO: remove all of this, see
+  // https://bugzilla.mozilla.org/show_bug.cgi?id=1611024
+  virtual void SetBindGroup(uint32_t aSlot, const BindGroup& aBindGroup,
+                            const dom::Sequence<uint32_t>& aDynamicOffsets) = 0;
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/RenderBundleEncoder.cpp b/dom/webgpu/RenderBundleEncoder.cpp
index 95f649a37db4..6a2679e87036 100644
--- a/dom/webgpu/RenderBundleEncoder.cpp
+++ b/dom/webgpu/RenderBundleEncoder.cpp
@@ -20,5 +20,13 @@ NS_IMPL_CYCLE_COLLECTION_TRACE_BEGIN_INHERITED(RenderBundleEncoder,
 NS_IMPL_CYCLE_COLLECTION_TRACE_END
 GPU_IMPL_JS_WRAP(RenderBundleEncoder)
 
+void RenderBundleEncoder::SetBindGroup(
+    uint32_t aSlot, const BindGroup& aBindGroup,
+    const dom::Sequence<uint32_t>& aDynamicOffsets) {
+  if (mValid) {
+    MOZ_CRASH("TODO");
+  }
+}
+
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/RenderBundleEncoder.h b/dom/webgpu/RenderBundleEncoder.h
index 0944dce33920..1bf9118edf81 100644
--- a/dom/webgpu/RenderBundleEncoder.h
+++ b/dom/webgpu/RenderBundleEncoder.h
@@ -29,6 +29,8 @@ class RenderBundleEncoder final : public RenderEncoderBase,
   ~RenderBundleEncoder() = default;
 
  public:
+  void SetBindGroup(uint32_t aSlot, const BindGroup& aBindGroup,
+                    const dom::Sequence<uint32_t>& aDynamicOffsets) override;
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/RenderPassEncoder.cpp b/dom/webgpu/RenderPassEncoder.cpp
index ed1aeb6d083d..24d3724644dd 100644
--- a/dom/webgpu/RenderPassEncoder.cpp
+++ b/dom/webgpu/RenderPassEncoder.cpp
@@ -21,5 +21,13 @@ GPU_IMPL_JS_WRAP(RenderPassEncoder)
 
 RenderPassEncoder::~RenderPassEncoder() = default;
 
+void RenderPassEncoder::SetBindGroup(
+    uint32_t aSlot, const BindGroup& aBindGroup,
+    const dom::Sequence<uint32_t>& aDynamicOffsets) {
+  if (mValid) {
+    MOZ_CRASH("TODO");
+  }
+}
+
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/RenderPassEncoder.h b/dom/webgpu/RenderPassEncoder.h
index 38b5ef2c3af0..1530933a8443 100644
--- a/dom/webgpu/RenderPassEncoder.h
+++ b/dom/webgpu/RenderPassEncoder.h
@@ -39,6 +39,8 @@ class RenderPassEncoder final : public RenderEncoderBase,
   void Cleanup() {}
 
  public:
+  void SetBindGroup(uint32_t aSlot, const BindGroup& aBindGroup,
+                    const dom::Sequence<uint32_t>& aDynamicOffsets) override;
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/Sampler.h b/dom/webgpu/Sampler.h
index bf283ab20360..e54289ea9b19 100644
--- a/dom/webgpu/Sampler.h
+++ b/dom/webgpu/Sampler.h
@@ -19,6 +19,8 @@ class Sampler final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(Sampler)
   GPU_DECL_JS_WRAP(Sampler)
 
+  const RawId mId;
+
  private:
   Sampler() = delete;
   virtual ~Sampler();
diff --git a/dom/webgpu/ShaderModule.cpp b/dom/webgpu/ShaderModule.cpp
index 1b176968dba2..7c30e91a3f3a 100644
--- a/dom/webgpu/ShaderModule.cpp
+++ b/dom/webgpu/ShaderModule.cpp
@@ -14,7 +14,20 @@ namespace webgpu {
 GPU_IMPL_CYCLE_COLLECTION(ShaderModule, mParent)
 GPU_IMPL_JS_WRAP(ShaderModule)
 
-ShaderModule::~ShaderModule() = default;
+ShaderModule::ShaderModule(Device* const aParent, RawId aId)
+    : ChildOf(aParent), mId(aId) {}
+
+ShaderModule::~ShaderModule() { Cleanup(); }
+
+void ShaderModule::Cleanup() {
+  if (mValid && mParent) {
+    mValid = false;
+    WebGPUChild* bridge = mParent->mBridge;
+    if (bridge && bridge->IsOpen()) {
+      bridge->DestroyShaderModule(mId);
+    }
+  }
+}
 
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/ShaderModule.h b/dom/webgpu/ShaderModule.h
index 7af690d57f9d..d64fce322e57 100644
--- a/dom/webgpu/ShaderModule.h
+++ b/dom/webgpu/ShaderModule.h
@@ -19,10 +19,13 @@ class ShaderModule final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(ShaderModule)
   GPU_DECL_JS_WRAP(ShaderModule)
 
+  ShaderModule(Device* const aParent, RawId aId);
+
+  const RawId mId;
+
  private:
-  ShaderModule() = delete;
   virtual ~ShaderModule();
-  void Cleanup() {}
+  void Cleanup();
 };
 
 }  // namespace webgpu
diff --git a/dom/webgpu/TextureView.h b/dom/webgpu/TextureView.h
index db0fcc9f5d60..483830b9f647 100644
--- a/dom/webgpu/TextureView.h
+++ b/dom/webgpu/TextureView.h
@@ -19,6 +19,8 @@ class TextureView final : public ObjectBase, public ChildOf<Device> {
   GPU_DECL_CYCLE_COLLECTION(TextureView)
   GPU_DECL_JS_WRAP(TextureView)
 
+  const RawId mId;
+
  private:
   TextureView() = delete;
   virtual ~TextureView();
diff --git a/dom/webgpu/ffi/moz.build b/dom/webgpu/ffi/moz.build
index 985abd023ece..b290474e4565 100644
--- a/dom/webgpu/ffi/moz.build
+++ b/dom/webgpu/ffi/moz.build
@@ -15,7 +15,9 @@ if CONFIG['COMPILE_ENVIRONMENT']:
     GeneratedFile('wgpu_ffi_generated.h',
                   script='/layout/style/RunCbindgen.py',
                   entry_point='generate',
-                  inputs=['/gfx/wgpu/wgpu-remote'])
+                  inputs=[
+                    '/gfx/wgpu/wgpu-remote',
+                  ])
 
     EXPORTS.mozilla.webgpu.ffi += [
         '!wgpu_ffi_generated.h',
diff --git a/dom/webgpu/ipc/PWebGPU.ipdl b/dom/webgpu/ipc/PWebGPU.ipdl
index bfe194c42519..aa64da9788f3 100644
--- a/dom/webgpu/ipc/PWebGPU.ipdl
+++ b/dom/webgpu/ipc/PWebGPU.ipdl
@@ -6,11 +6,17 @@
  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
 
 using RawId from "mozilla/webgpu/WebGPUTypes.h";
+using BufferAddress from "mozilla/webgpu/WebGPUTypes.h";
+using SerialBindGroupLayoutDescriptor from "mozilla/webgpu/WebGPUTypes.h";
+using SerialPipelineLayoutDescriptor from "mozilla/webgpu/WebGPUTypes.h";
+using SerialBindGroupDescriptor from "mozilla/webgpu/WebGPUTypes.h";
+using SerialComputePipelineDescriptor from "mozilla/webgpu/WebGPUTypes.h";
 using dom::GPURequestAdapterOptions from "mozilla/dom/WebGPUBinding.h";
 using dom::GPUDeviceDescriptor from "mozilla/dom/WebGPUBinding.h";
 using dom::GPUBufferDescriptor from "mozilla/dom/WebGPUBinding.h";
 using dom::GPUCommandEncoderDescriptor from "mozilla/dom/WebGPUBinding.h";
 using dom::GPUCommandBufferDescriptor from "mozilla/dom/WebGPUBinding.h";
+using dom::GPUPipelineLayoutDescriptor from "mozilla/dom/WebGPUBinding.h";
 
 include "mozilla/webgpu/WebGPUSerialize.h";
 include protocol PCompositorBridge;
@@ -37,11 +43,22 @@ parent:
   async BufferMapRead(RawId selfId, Shmem shmem) returns (Shmem sm);
   async BufferDestroy(RawId selfId);
   async DeviceCreateCommandEncoder(RawId selfId, GPUCommandEncoderDescriptor desc, RawId newId);
+  async CommandEncoderCopyBufferToBuffer(RawId selfId, RawId sourceId, BufferAddress sourceOffset, RawId destinationId, BufferAddress destinationOffset, BufferAddress size);
   async CommandEncoderRunComputePass(RawId selfId, Shmem shmem);
   async CommandEncoderFinish(RawId selfId, GPUCommandBufferDescriptor desc);
   async CommandEncoderDestroy(RawId selfId);
   async CommandBufferDestroy(RawId selfId);
   async QueueSubmit(RawId selfId, RawId[] commandBuffers);
+  async DeviceCreateBindGroupLayout(RawId selfId, SerialBindGroupLayoutDescriptor desc, RawId newId);
+  async BindGroupLayoutDestroy(RawId selfId);
+  async DeviceCreatePipelineLayout(RawId selfId, SerialPipelineLayoutDescriptor desc, RawId newId);
+  async PipelineLayoutDestroy(RawId selfId);
+  async DeviceCreateBindGroup(RawId selfId, SerialBindGroupDescriptor desc, RawId newId);
+  async BindGroupDestroy(RawId selfId);
+  async DeviceCreateShaderModule(RawId selfId, uint32_t[] data, RawId newId);
+  async ShaderModuleDestroy(RawId selfId);
+  async DeviceCreateComputePipeline(RawId selfId, SerialComputePipelineDescriptor desc, RawId newId);
+  async ComputePipelineDestroy(RawId selfId);
   async Shutdown();
 
 child:
diff --git a/dom/webgpu/ipc/WebGPUChild.cpp b/dom/webgpu/ipc/WebGPUChild.cpp
index 7d3cb98d5399..f931a942d86b 100644
--- a/dom/webgpu/ipc/WebGPUChild.cpp
+++ b/dom/webgpu/ipc/WebGPUChild.cpp
@@ -110,6 +110,100 @@ RawId WebGPUChild::CommandEncoderFinish(
   return aSelfId;
 }
 
+RawId WebGPUChild::DeviceCreateBindGroupLayout(
+    RawId aSelfId, const dom::GPUBindGroupLayoutDescriptor& aDesc) {
+  RawId id = ffi::wgpu_client_make_bind_group_layout_id(mClient, aSelfId);
+  nsTArray<ffi::WGPUBindGroupLayoutBinding> bindings(aDesc.mBindings.Length());
+  for (const auto& binding : aDesc.mBindings) {
+    ffi::WGPUBindGroupLayoutBinding b = {};
+    b.binding = binding.mBinding;
+    b.visibility = binding.mVisibility;
+    b.ty = ffi::WGPUBindingType(binding.mType);
+    Unused << binding.mTextureComponentType;  // TODO
+    b.texture_dimension =
+        ffi::WGPUTextureViewDimension(binding.mTextureDimension);
+    b.multisampled = binding.mMultisampled;
+    b.dynamic = binding.mDynamic;
+    bindings.AppendElement(b);
+  }
+  SerialBindGroupLayoutDescriptor desc = {std::move(bindings)};
+  if (!SendDeviceCreateBindGroupLayout(aSelfId, desc, id)) {
+    MOZ_CRASH("IPC failure");
+  }
+  return id;
+}
+
+RawId WebGPUChild::DeviceCreatePipelineLayout(
+    RawId aSelfId, const dom::GPUPipelineLayoutDescriptor& aDesc) {
+  RawId id = ffi::wgpu_client_make_pipeline_layout_id(mClient, aSelfId);
+  SerialPipelineLayoutDescriptor desc = {};
+  for (const auto& layouts : aDesc.mBindGroupLayouts) {
+    desc.mBindGroupLayouts.AppendElement(layouts->mId);
+  }
+  if (!SendDeviceCreatePipelineLayout(aSelfId, desc, id)) {
+    MOZ_CRASH("IPC failure");
+  }
+  return id;
+}
+
+RawId WebGPUChild::DeviceCreateBindGroup(
+    RawId aSelfId, const dom::GPUBindGroupDescriptor& aDesc) {
+  RawId id = ffi::wgpu_client_make_bind_group_id(mClient, aSelfId);
+  SerialBindGroupDescriptor desc = {};
+  desc.mLayout = aDesc.mLayout->mId;
+  for (const auto& binding : aDesc.mBindings) {
+    SerialBindGroupBinding bd = {};
+    bd.mBinding = binding.mBinding;
+    if (binding.mResource.IsGPUBufferBinding()) {
+      bd.mType = SerialBindGroupBindingType::Buffer;
+      const auto& bufBinding = binding.mResource.GetAsGPUBufferBinding();
+      bd.mValue = bufBinding.mBuffer->mId;
+      bd.mBufferOffset = bufBinding.mOffset;
+      bd.mBufferSize =
+          bufBinding.mSize.WasPassed() ? bufBinding.mSize.Value() : 0;
+    }
+    if (binding.mResource.IsGPUTextureView()) {
+      bd.mType = SerialBindGroupBindingType::Texture;
+      bd.mValue = binding.mResource.GetAsGPUTextureView()->mId;
+    }
+    if (binding.mResource.IsGPUSampler()) {
+      bd.mType = SerialBindGroupBindingType::Sampler;
+      bd.mValue = binding.mResource.GetAsGPUSampler()->mId;
+    }
+    desc.mBindings.AppendElement(bd);
+  }
+  if (!SendDeviceCreateBindGroup(aSelfId, desc, id)) {
+    MOZ_CRASH("IPC failure");
+  }
+  return id;
+}
+
+RawId WebGPUChild::DeviceCreateShaderModule(
+    RawId aSelfId, const dom::GPUShaderModuleDescriptor& aDesc) {
+  RawId id = ffi::wgpu_client_make_shader_module_id(mClient, aSelfId);
+  MOZ_ASSERT(aDesc.mCode.IsUint32Array());
+  const auto& code = aDesc.mCode.GetAsUint32Array();
+  code.ComputeLengthAndData();
+  nsTArray<uint32_t> data(code.Length());
+  data.AppendElements(code.Data(), code.Length());
+  if (!SendDeviceCreateShaderModule(aSelfId, data, id)) {
+    MOZ_CRASH("IPC failure");
+  }
+  return id;
+}
+RawId WebGPUChild::DeviceCreateComputePipeline(
+    RawId aSelfId, const dom::GPUComputePipelineDescriptor& aDesc) {
+  RawId id = ffi::wgpu_client_make_compute_pipeline_id(mClient, aSelfId);
+  SerialProgrammableStageDescriptor stage = {};
+  stage.mModule = aDesc.mComputeStage.mModule->mId;
+  stage.mEntryPoint = aDesc.mComputeStage.mEntryPoint;
+  SerialComputePipelineDescriptor desc = {aDesc.mLayout->mId, stage};
+  if (!SendDeviceCreateComputePipeline(aSelfId, desc, id)) {
+    MOZ_CRASH("IPC failure");
+  }
+  return id;
+}
+
 void WebGPUChild::QueueSubmit(RawId aSelfId,
                               const nsTArray<RawId>& aCommandBufferIds) {
   SendQueueSubmit(aSelfId, aCommandBufferIds);
@@ -134,6 +228,26 @@ void WebGPUChild::DestroyCommandBuffer(RawId aId) {
   SendCommandBufferDestroy(aId);
   ffi::wgpu_client_kill_encoder_id(mClient, aId);
 }
+void WebGPUChild::DestroyBindGroupLayout(RawId aId) {
+  SendBindGroupLayoutDestroy(aId);
+  ffi::wgpu_client_kill_bind_group_layout_id(mClient, aId);
+}
+void WebGPUChild::DestroyPipelineLayout(RawId aId) {
+  SendPipelineLayoutDestroy(aId);
+  ffi::wgpu_client_kill_pipeline_layout_id(mClient, aId);
+}
+void WebGPUChild::DestroyBindGroup(RawId aId) {
+  SendBindGroupDestroy(aId);
+  ffi::wgpu_client_kill_bind_group_id(mClient, aId);
+}
+void WebGPUChild::DestroyShaderModule(RawId aId) {
+  SendShaderModuleDestroy(aId);
+  ffi::wgpu_client_kill_shader_module_id(mClient, aId);
+}
+void WebGPUChild::DestroyComputePipeline(RawId aId) {
+  SendComputePipelineDestroy(aId);
+  ffi::wgpu_client_kill_compute_pipeline_id(mClient, aId);
+}
 
 }  // namespace webgpu
 }  // namespace mozilla
diff --git a/dom/webgpu/ipc/WebGPUChild.h b/dom/webgpu/ipc/WebGPUChild.h
index d76843febc70..5ad9593c4183 100644
--- a/dom/webgpu/ipc/WebGPUChild.h
+++ b/dom/webgpu/ipc/WebGPUChild.h
@@ -44,12 +44,28 @@ class WebGPUChild final : public PWebGPUChild {
       RawId aSelfId, const dom::GPUCommandEncoderDescriptor& aDesc);
   RawId CommandEncoderFinish(RawId aSelfId,
                              const dom::GPUCommandBufferDescriptor& aDesc);
+  RawId DeviceCreateBindGroupLayout(
+      RawId aSelfId, const dom::GPUBindGroupLayoutDescriptor& aDesc);
+  RawId DeviceCreatePipelineLayout(
+      RawId aSelfId, const dom::GPUPipelineLayoutDescriptor& aDesc);
+  RawId DeviceCreateBindGroup(RawId aSelfId,
+                              const dom::GPUBindGroupDescriptor& aDesc);
+  RawId DeviceCreateShaderModule(RawId aSelfId,
+                                 const dom::GPUShaderModuleDescriptor& aDesc);
+  RawId DeviceCreateComputePipeline(
+      RawId aSelfId, const dom::GPUComputePipelineDescriptor& aDesc);
+
   void QueueSubmit(RawId aSelfId, const nsTArray<RawId>& aCommandBufferIds);
 
   void DestroyAdapter(RawId aId);
   void DestroyBuffer(RawId aId);
   void DestroyCommandEncoder(RawId aId);
   void DestroyCommandBuffer(RawId aId);
+  void DestroyBindGroupLayout(RawId aId);
+  void DestroyPipelineLayout(RawId aId);
+  void DestroyBindGroup(RawId aId);
+  void DestroyShaderModule(RawId aId);
+  void DestroyComputePipeline(RawId aId);
 
  private:
   virtual ~WebGPUChild();
diff --git a/dom/webgpu/ipc/WebGPUParent.cpp b/dom/webgpu/ipc/WebGPUParent.cpp
index 265d8fd6953b..6f964c7447a8 100644
--- a/dom/webgpu/ipc/WebGPUParent.cpp
+++ b/dom/webgpu/ipc/WebGPUParent.cpp
@@ -46,6 +46,9 @@ ipc::IPCResult WebGPUParent::RecvInstanceRequestAdapter(
 ipc::IPCResult WebGPUParent::RecvAdapterRequestDevice(
     RawId aSelfId, const dom::GPUDeviceDescriptor& aDesc, RawId aNewId) {
   ffi::WGPUDeviceDescriptor desc = {};
+  desc.limits.max_bind_groups = aDesc.mLimits.WasPassed()
+                                    ? aDesc.mLimits.Value().mMaxBindGroups
+                                    : WGPUDEFAULT_BIND_GROUPS;
   Unused << aDesc;  // no useful fields
   // TODO: fill up the descriptor
   ffi::wgpu_server_adapter_request_device(mContext, aSelfId, &desc, aNewId);
@@ -67,6 +70,11 @@ ipc::IPCResult WebGPUParent::RecvDeviceCreateBuffer(
   ffi::WGPUBufferDescriptor desc = {};
   desc.usage = aDesc.mUsage;
   desc.size = aDesc.mSize;
+  // tweak: imply STORAGE_READ. This is yet to be figured out by the spec,
+  // see https://github.com/gpuweb/gpuweb/issues/541
+  if (desc.usage & WGPUBufferUsage_STORAGE) {
+    desc.usage |= WGPUBufferUsage_STORAGE_READ;
+  }
   ffi::wgpu_server_device_create_buffer(mContext, aSelfId, &desc, aNewId);
   return IPC_OK();
 }
@@ -121,6 +129,16 @@ ipc::IPCResult WebGPUParent::RecvDeviceCreateCommandEncoder(
   return IPC_OK();
 }
 
+ipc::IPCResult WebGPUParent::RecvCommandEncoderCopyBufferToBuffer(
+    RawId aSelfId, RawId aSourceId, BufferAddress aSourceOffset,
+    RawId aDestinationId, BufferAddress aDestinationOffset,
+    BufferAddress aSize) {
+  ffi::wgpu_server_encoder_copy_buffer_to_buffer(mContext, aSelfId, aSourceId,
+                                                 aSourceOffset, aDestinationId,
+                                                 aDestinationOffset, aSize);
+  return IPC_OK();
+}
+
 ipc::IPCResult WebGPUParent::RecvCommandEncoderRunComputePass(RawId aSelfId,
                                                               Shmem&& shmem) {
   ffi::wgpu_server_encode_compute_pass(mContext, aSelfId, shmem.get<uint8_t>(),
@@ -153,6 +171,107 @@ ipc::IPCResult WebGPUParent::RecvQueueSubmit(
   return IPC_OK();
 }
 
+ipc::IPCResult WebGPUParent::RecvDeviceCreateBindGroupLayout(
+    RawId aSelfId, const SerialBindGroupLayoutDescriptor& aDesc, RawId aNewId) {
+  ffi::WGPUBindGroupLayoutDescriptor desc = {};
+  desc.bindings = aDesc.mBindings.Elements();
+  desc.bindings_length = aDesc.mBindings.Length();
+  ffi::wgpu_server_device_create_bind_group_layout(mContext, aSelfId, &desc,
+                                                   aNewId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvBindGroupLayoutDestroy(RawId aSelfId) {
+  ffi::wgpu_server_bind_group_layout_destroy(mContext, aSelfId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvDeviceCreatePipelineLayout(
+    RawId aSelfId, const SerialPipelineLayoutDescriptor& aDesc, RawId aNewId) {
+  ffi::WGPUPipelineLayoutDescriptor desc = {};
+  desc.bind_group_layouts = aDesc.mBindGroupLayouts.Elements();
+  desc.bind_group_layouts_length = aDesc.mBindGroupLayouts.Length();
+  ffi::wgpu_server_device_create_pipeline_layout(mContext, aSelfId, &desc,
+                                                 aNewId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvPipelineLayoutDestroy(RawId aSelfId) {
+  ffi::wgpu_server_pipeline_layout_destroy(mContext, aSelfId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvDeviceCreateBindGroup(
+    RawId aSelfId, const SerialBindGroupDescriptor& aDesc, RawId aNewId) {
+  nsTArray<ffi::WGPUBindGroupBinding> ffiBindings(aDesc.mBindings.Length());
+  for (const auto& binding : aDesc.mBindings) {
+    ffi::WGPUBindGroupBinding bgb = {};
+    bgb.binding = binding.mBinding;
+    switch (binding.mType) {
+      case SerialBindGroupBindingType::Buffer:
+        bgb.resource.tag = ffi::WGPUBindingResource_Buffer;
+        bgb.resource.buffer._0.buffer = binding.mValue;
+        bgb.resource.buffer._0.offset = binding.mBufferOffset;
+        bgb.resource.buffer._0.size = binding.mBufferSize;
+        break;
+      case SerialBindGroupBindingType::Texture:
+        bgb.resource.tag = ffi::WGPUBindingResource_TextureView;
+        bgb.resource.texture_view._0 = binding.mValue;
+        break;
+      case SerialBindGroupBindingType::Sampler:
+        bgb.resource.tag = ffi::WGPUBindingResource_Sampler;
+        bgb.resource.sampler._0 = binding.mValue;
+        break;
+      default:
+        MOZ_CRASH("unreachable");
+    }
+    ffiBindings.AppendElement(bgb);
+  }
+  ffi::WGPUBindGroupDescriptor desc = {};
+  desc.layout = aDesc.mLayout;
+  desc.bindings = ffiBindings.Elements();
+  desc.bindings_length = ffiBindings.Length();
+  ffi::wgpu_server_device_create_bind_group(mContext, aSelfId, &desc, aNewId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvBindGroupDestroy(RawId aSelfId) {
+  ffi::wgpu_server_bind_group_destroy(mContext, aSelfId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvDeviceCreateShaderModule(
+    RawId aSelfId, const nsTArray<uint32_t>& aData, RawId aNewId) {
+  ffi::WGPUShaderModuleDescriptor desc = {};
+  desc.code.bytes = aData.Elements();
+  desc.code.length = aData.Length();
+  ffi::wgpu_server_device_create_shader_module(mContext, aSelfId, &desc,
+                                               aNewId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvShaderModuleDestroy(RawId aSelfId) {
+  ffi::wgpu_server_shader_module_destroy(mContext, aSelfId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvDeviceCreateComputePipeline(
+    RawId aSelfId, const SerialComputePipelineDescriptor& aDesc, RawId aNewId) {
+  NS_LossyConvertUTF16toASCII entryPoint(aDesc.mComputeStage.mEntryPoint);
+  ffi::WGPUComputePipelineDescriptor desc = {};
+  desc.layout = aDesc.mLayout;
+  desc.compute_stage.module = aDesc.mComputeStage.mModule;
+  desc.compute_stage.entry_point = entryPoint.get();
+  ffi::wgpu_server_device_create_compute_pipeline(mContext, aSelfId, &desc,
+                                                  aNewId);
+  return IPC_OK();
+}
+
+ipc::IPCResult WebGPUParent::RecvComputePipelineDestroy(RawId aSelfId) {
+  ffi::wgpu_server_compute_pipeline_destroy(mContext, aSelfId);
+  return IPC_OK();
+}
+
 ipc::IPCResult WebGPUParent::RecvShutdown() {
   mTimer.Stop();
   ffi::wgpu_server_poll_all_devices(mContext, true);
diff --git a/dom/webgpu/ipc/WebGPUParent.h b/dom/webgpu/ipc/WebGPUParent.h
index d1eb3a7630a9..9637c57a9e7d 100644
--- a/dom/webgpu/ipc/WebGPUParent.h
+++ b/dom/webgpu/ipc/WebGPUParent.h
@@ -42,6 +42,10 @@ class WebGPUParent final : public PWebGPUParent {
   ipc::IPCResult RecvDeviceCreateCommandEncoder(
       RawId aSelfId, const dom::GPUCommandEncoderDescriptor& aDesc,
       RawId aNewId);
+  ipc::IPCResult RecvCommandEncoderCopyBufferToBuffer(
+      RawId aSelfId, RawId aSourceId, BufferAddress aSourceOffset,
+      RawId aDestinationId, BufferAddress aDestinationOffset,
+      BufferAddress aSize);
   ipc::IPCResult RecvCommandEncoderRunComputePass(RawId aSelfId, Shmem&& shmem);
   ipc::IPCResult RecvCommandEncoderFinish(
       RawId aSelfId, const dom::GPUCommandBufferDescriptor& aDesc);
@@ -49,6 +53,24 @@ class WebGPUParent final : public PWebGPUParent {
   ipc::IPCResult RecvCommandBufferDestroy(RawId aSelfId);
   ipc::IPCResult RecvQueueSubmit(RawId aSelfId,
                                  const nsTArray<RawId>& aCommandBuffers);
+  ipc::IPCResult RecvDeviceCreateBindGroupLayout(
+      RawId aSelfId, const SerialBindGroupLayoutDescriptor& aDesc,
+      RawId aNewId);
+  ipc::IPCResult RecvBindGroupLayoutDestroy(RawId aSelfId);
+  ipc::IPCResult RecvDeviceCreatePipelineLayout(
+      RawId aSelfId, const SerialPipelineLayoutDescriptor& aDesc, RawId aNewId);
+  ipc::IPCResult RecvPipelineLayoutDestroy(RawId aSelfId);
+  ipc::IPCResult RecvDeviceCreateBindGroup(
+      RawId aSelfId, const SerialBindGroupDescriptor& aDesc, RawId aNewId);
+  ipc::IPCResult RecvBindGroupDestroy(RawId aSelfId);
+  ipc::IPCResult RecvDeviceCreateShaderModule(RawId aSelfId,
+                                              const nsTArray<uint32_t>& aData,
+                                              RawId aNewId);
+  ipc::IPCResult RecvShaderModuleDestroy(RawId aSelfId);
+  ipc::IPCResult RecvDeviceCreateComputePipeline(
+      RawId aSelfId, const SerialComputePipelineDescriptor& aDesc,
+      RawId aNewId);
+  ipc::IPCResult RecvComputePipelineDestroy(RawId aSelfId);
   ipc::IPCResult RecvShutdown();
 
  private:
diff --git a/dom/webgpu/ipc/WebGPUSerialize.h b/dom/webgpu/ipc/WebGPUSerialize.h
index 54c33b6024b1..9ab173093741 100644
--- a/dom/webgpu/ipc/WebGPUSerialize.h
+++ b/dom/webgpu/ipc/WebGPUSerialize.h
@@ -9,14 +9,17 @@
 #include "WebGPUTypes.h"
 #include "ipc/IPCMessageUtils.h"
 #include "mozilla/dom/WebGPUBinding.h"
+#include "mozilla/webgpu/ffi/wgpu.h"
 
 namespace IPC {
 
-#define DEFINE_IPC_SERIALIZER_ENUM(something)                    \
-  template <>                                                    \
-  struct ParamTraits<something>                                  \
-      : public ContiguousEnumSerializer<something, something(0), \
-                                        something::EndGuard_> {}
+#define DEFINE_IPC_SERIALIZER_ENUM_GUARD(something, guard) \
+  template <>                                              \
+  struct ParamTraits<something>                            \
+      : public ContiguousEnumSerializer<something, something(0), guard> {}
+
+#define DEFINE_IPC_SERIALIZER_ENUM(something) \
+  DEFINE_IPC_SERIALIZER_ENUM_GUARD(something, something::EndGuard_)
 
 DEFINE_IPC_SERIALIZER_ENUM(mozilla::dom::GPUPowerPreference);
 
@@ -32,7 +35,35 @@ DEFINE_IPC_SERIALIZER_WITH_FIELDS(mozilla::dom::GPUBufferDescriptor, mSize,
 DEFINE_IPC_SERIALIZER_WITHOUT_FIELDS(mozilla::dom::GPUCommandEncoderDescriptor);
 DEFINE_IPC_SERIALIZER_WITHOUT_FIELDS(mozilla::dom::GPUCommandBufferDescriptor);
 
+// TODO: have the enum sentinels generated by `cbindgen` when it supports them
+// see https://github.com/eqrion/cbindgen/issues/458
+DEFINE_IPC_SERIALIZER_ENUM_GUARD(
+    mozilla::webgpu::ffi::WGPUBindingType,
+    mozilla::webgpu::ffi::WGPUBindingType_StorageTexture);
+DEFINE_IPC_SERIALIZER_ENUM_GUARD(
+    mozilla::webgpu::ffi::WGPUTextureViewDimension,
+    mozilla::webgpu::ffi::WGPUTextureViewDimension_D3);
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(
+    mozilla::webgpu::ffi::WGPUBindGroupLayoutBinding, binding, visibility, ty,
+    texture_dimension, multisampled, dynamic);
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(
+    mozilla::webgpu::SerialBindGroupLayoutDescriptor, mBindings);
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(
+    mozilla::webgpu::SerialPipelineLayoutDescriptor, mBindGroupLayouts);
+DEFINE_IPC_SERIALIZER_ENUM(mozilla::webgpu::SerialBindGroupBindingType);
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(mozilla::webgpu::SerialBindGroupBinding,
+                                  mBinding, mType, mValue, mBufferOffset,
+                                  mBufferSize);
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(mozilla::webgpu::SerialBindGroupDescriptor,
+                                  mLayout, mBindings);
+
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(
+    mozilla::webgpu::SerialProgrammableStageDescriptor, mModule, mEntryPoint);
+DEFINE_IPC_SERIALIZER_WITH_FIELDS(
+    mozilla::webgpu::SerialComputePipelineDescriptor, mLayout, mComputeStage);
+
 #undef DEFINE_IPC_SERIALIZER_ENUM
+#undef DEFINE_IPC_SERIALIZER_ENUM_GUARD
 
 }  // namespace IPC
 #endif  // WEBGPU_SERIALIZE_H_
diff --git a/dom/webgpu/ipc/WebGPUTypes.h b/dom/webgpu/ipc/WebGPUTypes.h
index b253fcc2af4f..6d221642e562 100644
--- a/dom/webgpu/ipc/WebGPUTypes.h
+++ b/dom/webgpu/ipc/WebGPUTypes.h
@@ -7,15 +7,56 @@
 #define WEBGPU_TYPES_H_
 
 #include <cstdint>
+#include "nsTArray.h"
 
 namespace mozilla {
 namespace webgpu {
+namespace ffi {
+struct WGPUBindGroupLayoutBinding;
+}  // namespace ffi
 
 typedef uint64_t RawId;
 typedef uint64_t BufferAddress;
 
-}  // namespace webgpu
+struct SerialBindGroupLayoutDescriptor {
+  nsTArray<ffi::WGPUBindGroupLayoutBinding> mBindings;
+};
 
+struct SerialPipelineLayoutDescriptor {
+  nsTArray<RawId> mBindGroupLayouts;
+};
+
+enum class SerialBindGroupBindingType : uint8_t {
+  Buffer,
+  Texture,
+  Sampler,
+  EndGuard_
+};
+
+struct SerialBindGroupBinding {
+  uint32_t mBinding;
+  SerialBindGroupBindingType mType;
+  RawId mValue;
+  BufferAddress mBufferOffset;
+  BufferAddress mBufferSize;
+};
+
+struct SerialBindGroupDescriptor {
+  RawId mLayout;
+  nsTArray<SerialBindGroupBinding> mBindings;
+};
+
+struct SerialProgrammableStageDescriptor {
+  RawId mModule;
+  nsString mEntryPoint;
+};
+
+struct SerialComputePipelineDescriptor {
+  RawId mLayout;
+  SerialProgrammableStageDescriptor mComputeStage;
+};
+
+}  // namespace webgpu
 }  // namespace mozilla
 
 #endif  // WEBGPU_TYPES_H_
diff --git a/dom/webidl/WebGPU.webidl b/dom/webidl/WebGPU.webidl
index e296b3a0831c..6d6f1c7fb262 100644
--- a/dom/webidl/WebGPU.webidl
+++ b/dom/webidl/WebGPU.webidl
@@ -115,12 +115,12 @@ interface GPUDevice {
     //GPUTexture createTexture(GPUTextureDescriptor descriptor);
     //GPUSampler createSampler(optional GPUSamplerDescriptor descriptor = {});
 
-    //GPUBindGroupLayout createBindGroupLayout(GPUBindGroupLayoutDescriptor descriptor);
-    //GPUPipelineLayout createPipelineLayout(GPUPipelineLayoutDescriptor descriptor);
-    //GPUBindGroup createBindGroup(GPUBindGroupDescriptor descriptor);
+    GPUBindGroupLayout createBindGroupLayout(GPUBindGroupLayoutDescriptor descriptor);
+    GPUPipelineLayout createPipelineLayout(GPUPipelineLayoutDescriptor descriptor);
+    GPUBindGroup createBindGroup(GPUBindGroupDescriptor descriptor);
 
-    //GPUShaderModule createShaderModule(GPUShaderModuleDescriptor descriptor);
-    //GPUComputePipeline createComputePipeline(GPUComputePipelineDescriptor descriptor);
+    GPUShaderModule createShaderModule(GPUShaderModuleDescriptor descriptor);
+    GPUComputePipeline createComputePipeline(GPUComputePipelineDescriptor descriptor);
     //GPURenderPipeline createRenderPipeline(GPURenderPipelineDescriptor descriptor);
 
     [NewObject]
@@ -634,7 +634,7 @@ dictionary GPUPipelineDescriptorBase : GPUObjectDescriptorBase {
 };
 
 dictionary GPUProgrammableStageDescriptor {
-    required GPUShaderModule module_;
+    required GPUShaderModule module;
     required DOMString entryPoint;
 };
 
@@ -764,7 +764,6 @@ interface GPUCommandEncoder {
     [NewObject]
     GPUComputePassEncoder beginComputePass(optional GPUComputePassDescriptor descriptor = {});
 
-    /*
     void copyBufferToBuffer(
         GPUBuffer source,
         u64 sourceOffset,
@@ -772,6 +771,7 @@ interface GPUCommandEncoder {
         u64 destinationOffset,
         u64 size);
 
+    /*
     void copyBufferToTexture(
         GPUBufferCopyView source,
         GPUTextureCopyView destination,
@@ -805,8 +805,8 @@ GPUCommandEncoder includes GPUObjectBase;
 [Pref="dom.webgpu.enabled",
  Exposed=Window]
 interface GPUProgrammablePassEncoder {
-    //void setBindGroup(u32 index, GPUBindGroup bindGroup,
-    //                  optional sequence<u64> dynamicOffsets);
+    void setBindGroup(unsigned long index, GPUBindGroup bindGroup,
+                      optional sequence<unsigned long> dynamicOffsets = []);
 
     //void pushDebugGroup(DOMString groupLabel);
     //void popDebugGroup();
@@ -856,7 +856,7 @@ dictionary GPUComputePassDescriptor : GPUObjectDescriptorBase {
 [Pref="dom.webgpu.enabled",
  Exposed=Window]
 interface GPUComputePassEncoder : GPUProgrammablePassEncoder {
-    //void setPipeline(GPUComputePipeline pipeline);
+    void setPipeline(GPUComputePipeline pipeline);
     void dispatch(u32 x, optional u32 y = 1, optional u32 z = 1);
     //void dispatchIndirect(GPUBuffer indirectBuffer, u64 indirectOffset);
 
diff --git a/gfx/wgpu/ffi/wgpu.h b/gfx/wgpu/ffi/wgpu.h
index 56ef02c1c7a7..521c9f3cf45b 100644
--- a/gfx/wgpu/ffi/wgpu.h
+++ b/gfx/wgpu/ffi/wgpu.h
@@ -413,6 +413,8 @@ typedef WGPURawPass *WGPUComputePassId;
 
 typedef const char *WGPURawString;
 
+typedef uint32_t WGPUDynamicOffset;
+
 typedef uint64_t WGPUId_ComputePipeline_Dummy;
 
 typedef WGPUId_ComputePipeline_Dummy WGPUComputePipelineId;
@@ -502,8 +504,8 @@ typedef uint32_t WGPUBufferUsage;
 #define WGPUBufferUsage_VERTEX 32
 #define WGPUBufferUsage_UNIFORM 64
 #define WGPUBufferUsage_STORAGE 128
-#define WGPUBufferUsage_STORAGE_READ 256
-#define WGPUBufferUsage_INDIRECT 512
+#define WGPUBufferUsage_INDIRECT 256
+#define WGPUBufferUsage_STORAGE_READ 512
 #define WGPUBufferUsage_NONE 0
 
 typedef struct {
@@ -788,7 +790,7 @@ void wgpu_compute_pass_push_debug_group(WGPURawPass *_pass, WGPURawString _label
 void wgpu_compute_pass_set_bind_group(WGPURawPass *pass,
                                       uint32_t index,
                                       WGPUBindGroupId bind_group_id,
-                                      const WGPUBufferAddress *offsets,
+                                      const WGPUDynamicOffset *offsets,
                                       uintptr_t offset_length);
 
 void wgpu_compute_pass_set_pipeline(WGPURawPass *pass, WGPUComputePipelineId pipeline_id);
@@ -888,7 +890,7 @@ void wgpu_render_pass_push_debug_group(WGPURawRenderPass *_pass, WGPURawString _
 void wgpu_render_pass_set_bind_group(WGPURawRenderPass *pass,
                                      uint32_t index,
                                      WGPUBindGroupId bind_group_id,
-                                     const WGPUBufferAddress *offsets,
+                                     const WGPUDynamicOffset *offsets,
                                      uintptr_t offset_length);
 
 void wgpu_render_pass_set_blend_color(WGPURawRenderPass *pass, const WGPUColor *color);
diff --git a/gfx/wgpu/wgpu-core/src/command/bind.rs b/gfx/wgpu/wgpu-core/src/command/bind.rs
index 443d034a24b8..33737ccff684 100644
--- a/gfx/wgpu/wgpu-core/src/command/bind.rs
+++ b/gfx/wgpu/wgpu-core/src/command/bind.rs
@@ -6,7 +6,7 @@ use crate::{
     binding_model::BindGroup,
     hub::GfxBackend,
     id::{BindGroupId, BindGroupLayoutId, PipelineLayoutId},
-    BufferAddress,
+    DynamicOffset,
     Stored,
 };
 
@@ -26,7 +26,7 @@ pub struct BindGroupPair {
 #[derive(Debug)]
 pub enum LayoutChange<'a> {
     Unchanged,
-    Match(BindGroupId, &'a [BufferAddress]),
+    Match(BindGroupId, &'a [DynamicOffset]),
     Mismatch,
 }
 
@@ -41,7 +41,7 @@ pub struct FollowUpIter<'a> {
     iter: slice::Iter<'a, BindGroupEntry>,
 }
 impl<'a> Iterator for FollowUpIter<'a> {
-    type Item = (BindGroupId, &'a [BufferAddress]);
+    type Item = (BindGroupId, &'a [DynamicOffset]);
     fn next(&mut self) -> Option<Self::Item> {
         self.iter
             .next()
@@ -58,7 +58,7 @@ impl<'a> Iterator for FollowUpIter<'a> {
 pub struct BindGroupEntry {
     expected_layout_id: Option<BindGroupLayoutId>,
     provided: Option<BindGroupPair>,
-    dynamic_offsets: Vec<BufferAddress>,
+    dynamic_offsets: Vec<DynamicOffset>,
 }
 
 impl BindGroupEntry {
@@ -66,7 +66,7 @@ impl BindGroupEntry {
         &mut self,
         bind_group_id: BindGroupId,
         bind_group: &BindGroup<B>,
-        offsets: &[BufferAddress],
+        offsets: &[DynamicOffset],
     ) -> Provision {
         debug_assert_eq!(B::VARIANT, bind_group_id.backend());
 
@@ -167,7 +167,7 @@ impl Binder {
         index: usize,
         bind_group_id: BindGroupId,
         bind_group: &BindGroup<B>,
-        offsets: &[BufferAddress],
+        offsets: &[DynamicOffset],
     ) -> Option<(PipelineLayoutId, FollowUpIter<'a>)> {
         log::trace!("\tBinding [{}] = group {:?}", index, bind_group_id);
         debug_assert_eq!(B::VARIANT, bind_group_id.backend());
diff --git a/gfx/wgpu/wgpu-core/src/command/compute.rs b/gfx/wgpu/wgpu-core/src/command/compute.rs
index b8575370d7a9..0372884c807b 100644
--- a/gfx/wgpu/wgpu-core/src/command/compute.rs
+++ b/gfx/wgpu/wgpu-core/src/command/compute.rs
@@ -13,6 +13,7 @@ use crate::{
     id,
     resource::BufferUsage,
     BufferAddress,
+    DynamicOffset,
 };
 
 use hal::command::CommandBuffer as _;
@@ -27,7 +28,7 @@ enum ComputeCommand {
         index: u8,
         num_dynamic_offsets: u8,
         bind_group_id: id::BindGroupId,
-        phantom_offsets: PhantomSlice<BufferAddress>,
+        phantom_offsets: PhantomSlice<DynamicOffset>,
     },
     SetPipeline(id::ComputePipelineId),
     Dispatch([u32; 3]),
@@ -95,7 +96,7 @@ impl<F> Global<F> {
                     if cfg!(debug_assertions) {
                         for off in offsets {
                             assert_eq!(
-                                *off % BIND_BUFFER_ALIGNMENT,
+                                *off as BufferAddress % BIND_BUFFER_ALIGNMENT,
                                 0,
                                 "Misaligned dynamic buffer offset: {} does not align with {}",
                                 off,
@@ -222,6 +223,7 @@ pub mod compute_ffi {
     use crate::{
         id,
         BufferAddress,
+        DynamicOffset,
         RawString,
     };
     use std::{convert::TryInto, slice};
@@ -231,7 +233,7 @@ pub mod compute_ffi {
         pass: &mut RawPass,
         index: u32,
         bind_group_id: id::BindGroupId,
-        offsets: *const BufferAddress,
+        offsets: *const DynamicOffset,
         offset_length: usize,
     ) {
         pass.encode(&ComputeCommand::SetBindGroup {
diff --git a/gfx/wgpu/wgpu-core/src/command/render.rs b/gfx/wgpu/wgpu-core/src/command/render.rs
index f71bc62035a8..fa2e58b35fce 100644
--- a/gfx/wgpu/wgpu-core/src/command/render.rs
+++ b/gfx/wgpu/wgpu-core/src/command/render.rs
@@ -24,6 +24,7 @@ use crate::{
     track::TrackerSet,
     BufferAddress,
     Color,
+    DynamicOffset,
     Stored,
 };
 
@@ -106,7 +107,7 @@ enum RenderCommand {
         index: u8,
         num_dynamic_offsets: u8,
         bind_group_id: id::BindGroupId,
-        phantom_offsets: PhantomSlice<BufferAddress>,
+        phantom_offsets: PhantomSlice<DynamicOffset>,
     },
     SetPipeline(id::RenderPipelineId),
     SetIndexBuffer {
@@ -798,7 +799,7 @@ impl<F> Global<F> {
                     if cfg!(debug_assertions) {
                         for off in offsets {
                             assert_eq!(
-                                *off % BIND_BUFFER_ALIGNMENT,
+                                *off as BufferAddress % BIND_BUFFER_ALIGNMENT,
                                 0,
                                 "Misaligned dynamic buffer offset: {} does not align with {}",
                                 off,
@@ -1124,6 +1125,7 @@ pub mod render_ffi {
         id,
         BufferAddress,
         Color,
+        DynamicOffset,
         RawString,
     };
     use std::{convert::TryInto, slice};
@@ -1133,7 +1135,7 @@ pub mod render_ffi {
         pass: &mut RawRenderPass,
         index: u32,
         bind_group_id: id::BindGroupId,
-        offsets: *const BufferAddress,
+        offsets: *const DynamicOffset,
         offset_length: usize,
     ) {
         pass.raw.encode(&RenderCommand::SetBindGroup {
diff --git a/gfx/wgpu/wgpu-core/src/device/mod.rs b/gfx/wgpu/wgpu-core/src/device/mod.rs
index d412e89c6b75..3409e6b49f65 100644
--- a/gfx/wgpu/wgpu-core/src/device/mod.rs
+++ b/gfx/wgpu/wgpu-core/src/device/mod.rs
@@ -498,6 +498,8 @@ impl<F: IdentityFilter<id::BufferId>> Global<F> {
         let hub = B::hub(self);
         let mut token = Token::root();
 
+        log::info!("Create buffer {:?} with ID {:?}", desc, id_in);
+
         let (device_guard, mut token) = hub.devices.read(&mut token);
         let device = &device_guard[device_id];
         let buffer = device.create_buffer(device_id, desc);
@@ -884,14 +886,16 @@ impl<F: IdentityFilter<id::BindGroupLayoutId>> Global<F> {
             .map(|b| (b.binding, b))
             .collect();
 
-        {
-            let (bind_group_layout_guard, _) = hub.bind_group_layouts.read(&mut token);
-            let bind_group_layout =
-                bind_group_layout_guard
-                    .iter(device_id.backend())
-                    .find(|(_, bgl)| bgl.bindings == bindings_map);
+        // TODO: deduplicate the bind group layouts at some level.
+        // We can't do it right here, because in the remote scenario
+        // the client need to know if the same ID can be used, or not.
+        if false {
+            let (bgl_guard, _) = hub.bind_group_layouts.read(&mut token);
+            let bind_group_layout_id = bgl_guard
+                .iter(device_id.backend())
+                .find(|(_, bgl)| bgl.bindings == bindings_map);
 
-            if let Some((id, _)) = bind_group_layout {
+            if let Some((id, _)) = bind_group_layout_id {
                 return id;
             }
         }
@@ -925,6 +929,13 @@ impl<F: IdentityFilter<id::BindGroupLayoutId>> Global<F> {
         hub.bind_group_layouts
             .register_identity(id_in, layout, &mut token)
     }
+
+    pub fn bind_group_layout_destroy<B: GfxBackend>(&self, bind_group_layout_id: id::BindGroupLayoutId) {
+        let hub = B::hub(self);
+        let mut token = Token::root();
+        //TODO: track usage by GPU
+        hub.bind_group_layouts.unregister(bind_group_layout_id, &mut token);
+    }
 }
 
 impl<F: IdentityFilter<id::PipelineLayoutId>> Global<F> {
@@ -967,6 +978,13 @@ impl<F: IdentityFilter<id::PipelineLayoutId>> Global<F> {
         hub.pipeline_layouts
             .register_identity(id_in, layout, &mut token)
     }
+
+    pub fn pipeline_layout_destroy<B: GfxBackend>(&self, pipeline_layout_id: id::PipelineLayoutId) {
+        let hub = B::hub(self);
+        let mut token = Token::root();
+        //TODO: track usage by GPU
+        hub.pipeline_layouts.unregister(pipeline_layout_id, &mut token);
+    }
 }
 
 impl<F: IdentityFilter<id::BindGroupId>> Global<F> {
@@ -1200,6 +1218,13 @@ impl<F: IdentityFilter<id::ShaderModuleId>> Global<F> {
         hub.shader_modules
             .register_identity(id_in, shader, &mut token)
     }
+
+    pub fn shader_module_destroy<B: GfxBackend>(&self, shader_module_id: id::ShaderModuleId) {
+        let hub = B::hub(self);
+        let mut token = Token::root();
+        //TODO: track usage by GPU
+        hub.shader_modules.unregister(shader_module_id, &mut token);
+    }
 }
 
 impl<F: IdentityFilter<id::CommandEncoderId>> Global<F> {
@@ -1718,6 +1743,13 @@ impl<F: IdentityFilter<id::RenderPipelineId>> Global<F> {
         hub.render_pipelines
             .register_identity(id_in, pipeline, &mut token)
     }
+
+    pub fn render_pipeline_destroy<B: GfxBackend>(&self, render_pipeline_id: id::RenderPipelineId) {
+        let hub = B::hub(self);
+        let mut token = Token::root();
+        //TODO: track usage by GPU
+        hub.render_pipelines.unregister(render_pipeline_id, &mut token);
+    }
 }
 
 impl<F: IdentityFilter<id::ComputePipelineId>> Global<F> {
@@ -1773,6 +1805,13 @@ impl<F: IdentityFilter<id::ComputePipelineId>> Global<F> {
         hub.compute_pipelines
             .register_identity(id_in, pipeline, &mut token)
     }
+
+    pub fn compute_pipeline_destroy<B: GfxBackend>(&self, compute_pipeline_id: id::ComputePipelineId) {
+        let hub = B::hub(self);
+        let mut token = Token::root();
+        //TODO: track usage by GPU
+        hub.compute_pipelines.unregister(compute_pipeline_id, &mut token);
+    }
 }
 
 fn validate_swap_chain_descriptor(
diff --git a/gfx/wgpu/wgpu-core/src/hub.rs b/gfx/wgpu/wgpu-core/src/hub.rs
index 208ed5ad97c2..7c0cfc369b30 100644
--- a/gfx/wgpu/wgpu-core/src/hub.rs
+++ b/gfx/wgpu/wgpu-core/src/hub.rs
@@ -39,7 +39,7 @@ use vec_map::VecMap;
 
 #[cfg(debug_assertions)]
 use std::cell::Cell;
-use std::{marker::PhantomData, ops};
+use std::{fmt::Debug, marker::PhantomData, ops};
 
 
 /// A simple structure to manage identities of objects.
@@ -251,13 +251,13 @@ impl<'a, T> Drop for Token<'a, T> {
 }
 
 
-pub trait IdentityFilter<I> {
-    type Input: Clone;
+pub trait IdentityFilter<I>: Debug {
+    type Input: Clone + Debug;
     fn process(&self, id: Self::Input, backend: Backend) -> I;
     fn free(&self, id: I);
 }
 
-impl<I: TypedId + Clone> IdentityFilter<I> for () {
+impl<I: TypedId + Clone + Debug> IdentityFilter<I> for () {
     type Input = I;
     fn process(&self, id: I, _backend: Backend) -> I {
         //debug_assert_eq!(id.unzip().2, backend);
@@ -266,7 +266,7 @@ impl<I: TypedId + Clone> IdentityFilter<I> for () {
     fn free(&self, _id: I) {}
 }
 
-impl<I: TypedId> IdentityFilter<I> for Mutex<IdentityManager> {
+impl<I: TypedId + Debug> IdentityFilter<I> for Mutex<IdentityManager> {
     type Input = PhantomData<I>;
     fn process(&self, _id: Self::Input, backend: Backend) -> I {
         self.lock().alloc(backend)
diff --git a/gfx/wgpu/wgpu-core/src/lib.rs b/gfx/wgpu/wgpu-core/src/lib.rs
index d6e2fc9208b7..4239b2992ae1 100644
--- a/gfx/wgpu/wgpu-core/src/lib.rs
+++ b/gfx/wgpu/wgpu-core/src/lib.rs
@@ -55,6 +55,7 @@ pub enum Backend {
 }
 
 pub type BufferAddress = u64;
+pub type DynamicOffset = u32;
 pub type RawString = *const c_char;
 
 //TODO: make it private. Currently used for swapchain creation impl.
diff --git a/gfx/wgpu/wgpu-core/src/resource.rs b/gfx/wgpu/wgpu-core/src/resource.rs
index 1456fbb44966..7baa64df9176 100644
--- a/gfx/wgpu/wgpu-core/src/resource.rs
+++ b/gfx/wgpu/wgpu-core/src/resource.rs
@@ -32,8 +32,8 @@ bitflags::bitflags! {
         const VERTEX = 32;
         const UNIFORM = 64;
         const STORAGE = 128;
-        const STORAGE_READ = 256;
-        const INDIRECT = 512;
+        const INDIRECT = 256;
+        const STORAGE_READ = 512;
         const NONE = 0;
         /// The combination of all read-only usages.
         const READ_ALL = Self::MAP_READ.bits | Self::COPY_SRC.bits |
diff --git a/gfx/wgpu/wgpu-remote/src/lib.rs b/gfx/wgpu/wgpu-remote/src/lib.rs
index a979d8991b3b..798db2a7e178 100644
--- a/gfx/wgpu/wgpu-remote/src/lib.rs
+++ b/gfx/wgpu/wgpu-remote/src/lib.rs
@@ -27,6 +27,11 @@ struct IdentityHub {
     devices: IdentityManager,
     buffers: IdentityManager,
     command_buffers: IdentityManager,
+    bind_group_layouts: IdentityManager,
+    pipeline_layouts: IdentityManager,
+    bind_groups: IdentityManager,
+    shader_modules: IdentityManager,
+    compute_pipelines: IdentityManager,
 }
 
 #[derive(Debug, Default)]
@@ -200,3 +205,138 @@ pub unsafe extern "C" fn wgpu_command_encoder_begin_compute_pass(
 pub unsafe extern "C" fn wgpu_compute_pass_destroy(pass: core::command::RawPass) {
     let _ = pass.into_vec();
 }
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_make_bind_group_layout_id(
+    client: &Client,
+    device_id: id::DeviceId,
+) -> id::BindGroupLayoutId {
+    let backend = device_id.backend();
+    client
+        .identities
+        .lock()
+        .select(backend)
+        .bind_group_layouts
+        .alloc(backend)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_kill_bind_group_layout_id(
+    client: &Client,
+    id: id::BindGroupLayoutId,
+) {
+    client
+        .identities
+        .lock()
+        .select(id.backend())
+        .bind_group_layouts
+        .free(id)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_make_pipeline_layout_id(
+    client: &Client,
+    device_id: id::DeviceId,
+) -> id::PipelineLayoutId {
+    let backend = device_id.backend();
+    client
+        .identities
+        .lock()
+        .select(backend)
+        .pipeline_layouts
+        .alloc(backend)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_kill_pipeline_layout_id(
+    client: &Client,
+    id: id::PipelineLayoutId,
+) {
+    client
+        .identities
+        .lock()
+        .select(id.backend())
+        .pipeline_layouts
+        .free(id)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_make_bind_group_id(
+    client: &Client,
+    device_id: id::DeviceId,
+) -> id::BindGroupId {
+    let backend = device_id.backend();
+    client
+        .identities
+        .lock()
+        .select(backend)
+        .bind_groups
+        .alloc(backend)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_kill_bind_group_id(
+    client: &Client,
+    id: id::BindGroupId,
+) {
+    client
+        .identities
+        .lock()
+        .select(id.backend())
+        .bind_groups
+        .free(id)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_make_shader_module_id(
+    client: &Client,
+    device_id: id::DeviceId,
+) -> id::ShaderModuleId {
+    let backend = device_id.backend();
+    client
+        .identities
+        .lock()
+        .select(backend)
+        .shader_modules
+        .alloc(backend)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_kill_shader_module_id(
+    client: &Client,
+    id: id::ShaderModuleId,
+) {
+    client
+        .identities
+        .lock()
+        .select(id.backend())
+        .shader_modules
+        .free(id)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_make_compute_pipeline_id(
+    client: &Client,
+    device_id: id::DeviceId,
+) -> id::ComputePipelineId {
+    let backend = device_id.backend();
+    client
+        .identities
+        .lock()
+        .select(backend)
+        .compute_pipelines
+        .alloc(backend)
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_client_kill_compute_pipeline_id(
+    client: &Client,
+    id: id::ComputePipelineId,
+) {
+    client
+        .identities
+        .lock()
+        .select(id.backend())
+        .compute_pipelines
+        .free(id)
+}
diff --git a/gfx/wgpu/wgpu-remote/src/server.rs b/gfx/wgpu/wgpu-remote/src/server.rs
index 1a1c14f3c571..deaf559cdf52 100644
--- a/gfx/wgpu/wgpu-remote/src/server.rs
+++ b/gfx/wgpu/wgpu-remote/src/server.rs
@@ -127,9 +127,9 @@ pub extern "C" fn wgpu_server_device_create_encoder(
     global: &Global,
     self_id: id::DeviceId,
     desc: &core::command::CommandEncoderDescriptor,
-    encoder_id: id::CommandEncoderId,
+    new_id: id::CommandEncoderId,
 ) {
-    gfx_select!(self_id => global.device_create_command_encoder(self_id, &desc, encoder_id));
+    gfx_select!(self_id => global.device_create_command_encoder(self_id, &desc, new_id));
 }
 
 #[no_mangle]
@@ -157,6 +157,19 @@ pub extern "C" fn wgpu_server_command_buffer_destroy(
     gfx_select!(self_id => global.command_buffer_destroy(self_id));
 }
 
+#[no_mangle]
+pub unsafe extern "C" fn wgpu_server_encoder_copy_buffer_to_buffer(
+    global: &Global,
+    self_id: id::CommandEncoderId,
+    source_id: id::BufferId,
+    source_offset: core::BufferAddress,
+    destination_id: id::BufferId,
+    destination_offset: core::BufferAddress,
+    size: core::BufferAddress,
+) {
+    gfx_select!(self_id => global.command_encoder_copy_buffer_to_buffer(self_id, source_id, source_offset, destination_id, destination_offset, size));
+}
+
 #[no_mangle]
 pub unsafe extern "C" fn wgpu_server_encode_compute_pass(
     global: &Global,
@@ -193,3 +206,93 @@ pub unsafe extern "C" fn wgpu_server_queue_submit(
     let command_buffers = slice::from_raw_parts(command_buffer_ids, command_buffer_id_length);
     gfx_select!(self_id => global.queue_submit(self_id, command_buffers));
 }
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_device_create_bind_group_layout(
+    global: &Global,
+    self_id: id::DeviceId,
+    desc: &core::binding_model::BindGroupLayoutDescriptor,
+    new_id: id::BindGroupLayoutId,
+) {
+    gfx_select!(self_id => global.device_create_bind_group_layout(self_id, desc, new_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_bind_group_layout_destroy(
+    global: &Global,
+    self_id: id::BindGroupLayoutId,
+) {
+    gfx_select!(self_id => global.bind_group_layout_destroy(self_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_device_create_pipeline_layout(
+    global: &Global,
+    self_id: id::DeviceId,
+    desc: &core::binding_model::PipelineLayoutDescriptor,
+    new_id: id::PipelineLayoutId,
+) {
+    gfx_select!(self_id => global.device_create_pipeline_layout(self_id, desc, new_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_pipeline_layout_destroy(
+    global: &Global,
+    self_id: id::PipelineLayoutId,
+) {
+    gfx_select!(self_id => global.pipeline_layout_destroy(self_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_device_create_bind_group(
+    global: &Global,
+    self_id: id::DeviceId,
+    desc: &core::binding_model::BindGroupDescriptor,
+    new_id: id::BindGroupId,
+) {
+    gfx_select!(self_id => global.device_create_bind_group(self_id, desc, new_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_bind_group_destroy(
+    global: &Global,
+    self_id: id::BindGroupId,
+) {
+    gfx_select!(self_id => global.bind_group_destroy(self_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_device_create_shader_module(
+    global: &Global,
+    self_id: id::DeviceId,
+    desc: &core::pipeline::ShaderModuleDescriptor,
+    new_id: id::ShaderModuleId,
+) {
+    gfx_select!(self_id => global.device_create_shader_module(self_id, desc, new_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_shader_module_destroy(
+    global: &Global,
+    self_id: id::ShaderModuleId,
+) {
+    gfx_select!(self_id => global.shader_module_destroy(self_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_device_create_compute_pipeline(
+    global: &Global,
+    self_id: id::DeviceId,
+    desc: &core::pipeline::ComputePipelineDescriptor,
+    new_id: id::ComputePipelineId,
+) {
+    gfx_select!(self_id => global.device_create_compute_pipeline(self_id, desc, new_id));
+}
+
+#[no_mangle]
+pub extern "C" fn wgpu_server_compute_pipeline_destroy(
+    global: &Global,
+    self_id: id::ComputePipelineId,
+) {
+    gfx_select!(self_id => global.compute_pipeline_destroy(self_id));
+}