use ffi::TensorView instead of ffi::Tensor

[cyx-6]

📌 Description

This PR migrates the arguments from ffi::Tensor instead of ffi::TensorView.

• ffi::Tensor: Owning tensor friendly to torch.Tensor.
• ffi::TensorView: Non-owning tensor, equivalent to dltensor*, which is commonly used in JAX/XLA.

So we do such migration for broader support.

🔍 Related Issues

🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull request, please make sure the following items are complete.

✅ Pre-commit Checks

• I have installed pre-commit by running pip install pre-commit (or used your preferred method).
• I have installed the hooks with pre-commit install.
• I have run the hooks manually with pre-commit run --all-files and fixed any reported issues.

If you are unsure about how to set up pre-commit, see the pre-commit documentation.

🧪 Tests

• Tests have been added or updated as needed.
• All tests are passing (unittest, etc.).

Reviewer Notes

[yzh119]

Please create a PR to update the tvm-ffi version in docker container first.

[yzh119]

Queued after #1857

[yzh119]

Queued behind #1860

Also, @cyx-6 please explain the motivation of the change (JAX's non-owner view of tensors)

[yzh119]

/bot run

[flashinfer-bot]

GitLab MR !77 has been created, and the CI pipeline #36450141 is currently running. I'll report back once the pipeline job completes.

[yzh119]

@cyx-6 would you mind double checking whether there are remaining usage of Tensor?

If we have fixed all usage and https://gitlab-master.nvidia.com/dl/flashinfer/flashinfer-ci/-/pipelines/36450141 doesn't introduce new regression, we can merge this PR.

[cyx-6]

/bot run

[flashinfer-bot]

GitLab MR !77 has been updated with latest changes, and the CI pipeline #36453969 is currently running. I'll report back once the pipeline job completes.

[flashinfer-bot]

[SUCCESS] Pipeline #36453969: 11/17 passed