[CUDA] Fix null allocator passed to plugin EP kernel PrePack

[tianleiwu]

Description

When ONNX Runtime is built with the CUDA execution provider as a plugin
(onnxruntime_BUILD_CUDA_EP_AS_PLUGIN=ON), the EP-API op-kernel adapter
(ep::adapter::KernelImpl::PrePackWeightImpl) received a valid OrtAllocator*
from the framework but discarded it and forwarded a null AllocatorPtr{}
into the wrapped kernel's PrePack(). Any CUDA kernel that pre-packs a constant
weight (e.g. MatMulNBits, Conv, GroupQueryAttention, quantized MoE) then
allocated scratch through that null allocator and crashed during session
initialization:

IAllocator::ValidateAllocator(const T&) [with T = std::shared_ptr<onnxruntime::IAllocator>]
allocator != nullptr was false

This surfaced end to end as an ONNX Runtime GenAI og.Model(...) failure on a
gpt-oss-20b (MatMulNBits) model when the CUDA EP was loaded as a plugin: the
trivial init session succeeded, but the first real model session crashed while
pre-packing quantized weights.

Key Changes

File	Change
include/onnxruntime/ep/adapter/op_kernel.h	PrePackWeightImpl now wraps the incoming OrtAllocator* and forwards a valid AllocatorPtr to OpKernel::PrePack instead of a null AllocatorPtr{}.
include/onnxruntime/ep/adapter/allocator.h	Add a non-owning IAllocatorWrappingOrtAllocator(OrtAllocator*) constructor. The framework owns the pre-pack allocator, so the wrapper must not take ownership (an owning Ort::Allocator would release it on destruction). Calls now go through the raw OrtAllocator* function pointers directly, preserving the Reserve→Alloc (version ≥ 18) and GetStats/AllocOnStream (version ≥ 23) fallbacks.
onnxruntime/test/python/transformers/test_cuda_plugin_ep.py	Add test_registration_matmul_nbits_prepack: builds a fp16 MatMulNBits model with a runtime-prepacked (weight_prepacked=0) quantized weight, so weight pre-packing (MatMulNBits::PrePack_B → IAllocator::MakeUniquePtr(alloc, ...)) runs during session creation. This crashed before the fix and now passes.

Motivation and Context

The pre-pack allocator is provided and owned by the framework for the duration
of the PrePack call. The legacy (in-tree) CUDA EP received it correctly; only
the plugin op-kernel adapter dropped it. The non-owning wrapper matches the
lifetime contract used elsewhere in the adapter (e.g. KernelInfoGetAllocator)
and keeps the CUDA-EP-as-plugin build behaviorally identical to the in-tree EP.

Testing

• New test_registration_matmul_nbits_prepack in test_cuda_plugin_ep.py
  passes on a CUDA-EP-as-plugin build (ORT_TEST_CUDA_PLUGIN_EP=1) and skips
  gracefully when the device lacks fpA_intB GEMM support. It re-raises (fails)
  if the allocator != nullptr assertion recurs.
• Verified the model that originally reproduced the crash now creates and runs
  end to end through the CUDA plugin EP (gpt-oss-20b, MatMulNBits), including a
  100-sample MMLU sanity run.
• Existing test_cuda_plugin_ep.py registration tests continue to pass.

[Copilot]

Pull request overview

Fixes a crash when ONNX Runtime is built with the CUDA EP as a plugin by ensuring the framework-provided pre-pack allocator is correctly forwarded through the EP-API kernel adapter into kernel PrePack() implementations.

Changes:

• Forward the incoming OrtAllocator* through KernelImpl::PrePackWeightImpl instead of dropping it and passing a null AllocatorPtr.
• Add a non-owning IAllocatorWrappingOrtAllocator(OrtAllocator*) path so the adapter can wrap the framework-owned allocator without releasing it.
• Add a Python regression test that constructs a MatMulNBits model intended to trigger weight pre-packing during session creation.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
include/onnxruntime/ep/adapter/op_kernel.h	Fixes allocator plumbing in plugin EP PrePackWeight adapter path.
include/onnxruntime/ep/adapter/allocator.h	Adds non-owning allocator wrapper constructor and routes calls via raw OrtAllocator* function pointers.
onnxruntime/test/python/transformers/test_cuda_plugin_ep.py	Adds regression test for CUDA plugin EP pre-pack allocator forwarding via MatMulNBits pre-pack.

[tianleiwu]

@microsoft-github-policy-service rerun