[WebGPU] Register DataTransfer to Env

[qjia7]

This pull request adds a C API for WebGPU data transfer, enabling tensor copying between CPU and GPU devices via the WebGPU execution provider. The main changes introduce a wrapper implementation for data transfer, integrate it with the plugin execution provider factory, and expose a creation function for use by the ONNX Runtime core.

[fs-eire]

The Environment::data_transfer_mgr_ is different from InferenceSession::data_transfer_mgr_. They are the same type, but the one inside Environment should not depend on any session instance. The changes in this PR brings dependency on a specific session for Environment and I believe this is not what we want.

There is an existing method CreateAndRegisterInternalEps in class Environment, which should have already called RegisterExecutionProviderLibrary on WebGPU. Why the data transfer is not correctly registered - I can take a look at it.

[fs-eire]

Please add override CreateDataTransfer:

onnxruntime/onnxruntime/core/session/plugin_ep/ep_factory_internal_impl.h

Lines 56 to 59 in c30905d

	virtual OrtStatus* CreateDataTransfer(_Outptr_result_maybenull_ OrtDataTransferImpl** data_transfer) noexcept {
	*data_transfer = nullptr;
	return nullptr; // Default implementation does nothing
	}

in class WebGpuEpFactory (file in https://github.com/Microsoft/onnxruntime/blob/main/onnxruntime/core/session/plugin_ep/ep_factory_webgpu.h) for the purpose of this change.

[qjia7]

Please add override CreateDataTransfer:

onnxruntime/onnxruntime/core/session/plugin_ep/ep_factory_internal_impl.h

Lines 56 to 59 in c30905d

	virtual OrtStatus* CreateDataTransfer(_Outptr_result_maybenull_ OrtDataTransferImpl** data_transfer) noexcept {
	*data_transfer = nullptr;
	return nullptr; // Default implementation does nothing
	}

in class WebGpuEpFactory (file in https://github.com/Microsoft/onnxruntime/blob/main/onnxruntime/core/session/plugin_ep/ep_factory_webgpu.h) for the purpose of this change.

WebGPU DataTransfer requires a BufferManager.
For graph capture, the BufferManager is tied to the execution provider instance not session-independent. That's the problem.

[qjia7]

Please add override CreateDataTransfer:

onnxruntime/onnxruntime/core/session/plugin_ep/ep_factory_internal_impl.h

Lines 56 to 59 in c30905d

	virtual OrtStatus* CreateDataTransfer(_Outptr_result_maybenull_ OrtDataTransferImpl** data_transfer) noexcept {
	*data_transfer = nullptr;
	return nullptr; // Default implementation does nothing
	}

in class WebGpuEpFactory (file in https://github.com/Microsoft/onnxruntime/blob/main/onnxruntime/core/session/plugin_ep/ep_factory_webgpu.h) for the purpose of this change.

Done. Use the context 0's buffer manager. Will create one if not exist.

[qjia7]

@fs-eire @guschmue The webgpu related failures have been fixed. Others are not related with my changes. Please take a look, thanks.

[Copilot]

Pull request overview

This pull request adds WebGPU data transfer functionality to the ONNX Runtime core, enabling tensor copying between CPU and GPU devices via the WebGPU execution provider. The implementation provides a C API wrapper with lazy initialization that determines the WebGPU context from the tensors during the first copy operation.

Key Changes:

• Adds CreateDataTransfer method to WebGpuEpFactory for registering data transfer with the environment
• Implements WebGpuDataTransferImpl wrapper that bridges C API and C++ internal data transfer implementation
• Introduces lazy initialization of WebGPU context based on tensor device information

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 8 comments.

Show a summary per file

File	Description
onnxruntime/core/session/plugin_ep/ep_factory_webgpu.h	Declares CreateDataTransfer override method in WebGpuEpFactory
onnxruntime/core/session/plugin_ep/ep_factory_webgpu.cc	Implements CreateDataTransfer by calling WebGPU provider's C API function
onnxruntime/core/providers/webgpu/webgpu_provider_factory_creator.h	Declares C API function OrtWebGpuCreateDataTransfer() for creating data transfer instances
onnxruntime/core/providers/webgpu/webgpu_provider_factory.cc	Implements data transfer wrapper with lazy context initialization, helper functions, and vendor ID filtering
onnxruntime/core/providers/webgpu/webgpu_context.h	Adds HasContext method to WebGpuContextFactory for checking context existence
onnxruntime/core/providers/webgpu/webgpu_context.cc	Implements HasContext method with thread-safe context lookup

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[fs-eire]

I understand that the code are designed for situations that assume different context ID may be used for the same WebGpuDataTransferImpl instance.

This causes the lock behavior to be complicated. for now, I think it is safe to assume context ID won't change: it should always be 0.

based on this, code may be simplified:

• use patterns like this to avoid unnecessary lock operation:

  if (impl.data_transfer_ == nullptr) {
    std::lock_guard<std::mutex> lock(impl.init_mutex_);
    if (impl.data_transfer_ == nullptr) {
      ...
  
      impl.data_transfer_ = ...;
    }
  }

• always create new context:

  context_ptr = &webgpu::WebGpuContextFactory::CreateContext(params.context_config);

  and in ReleaseImpl release it

  static void ReleaseImpl(OrtDataTransferImpl* this_ptr) noexcept {
    WebGpuDataTransferImpl* p_impl = static_cast<WebGpuDataTransferImpl*>(this_ptr);
    int context_id = p_impl->context_id_;
    bool data_transfer_initialized = false;
    {
      std::lock_guard<std::mutex> lock(p_impl->init_mutex_);
      data_transfer_initialized = p_impl->data_transfer_ == nullptr;
    }
    delete p_impl;
    if (data_transfer_initialized) {
      WebGpuContextFactory::ReleaseContext(context_id);
    }
  }

[qjia7]

One new issue appears with this change.

1. OrtEnv::~OrtEnv() (destructor called first)
   Calls webgpu::CleanupWebGpuContexts()
   This clears all WebGPU contexts from WebGpuContextFactory

2. Environment::~Environment() (destructor called later)
   Destroys data_transfer_mgr_ member
   This destroys all registered IDataTransfer instances including WebGpuDataTransferImpl
   WebGpuDataTransferImpl::ReleaseImpl() (called during destruction)

3. Tries to call WebGpuContextFactory::ReleaseContext(context_id)
   But the context was already cleared in step 1
   ReleaseContext had an ORT_ENFORCE that threw an error ❌

So I removed webgpu::CleanupWebGpuContexts() from OrtEnv::~OrtEnv() in commit b988a12. Is it the right way? Or delay it to Environment::UnregisterExecutionProviderLibrary after data_transfer_mgr_.UnregisterDataTransfer ?

[qjia7]

Just explicitly destroy the Environment first. Then call webgpu::CleanupWebGpuContexts() in OrtEnv::~OrtEnv().