[SYCL][Fusion][NoSTL][NFC] Move SYCLModuleInfo to its own header

[jopperm]

Move SYCLModuleInfo class to its own header under the common directory (because it has users from jit-compiler and passes).

This PR is part of a series of changes to remove uses of STL classes in the kernel fusion interface to prevent ABI issues in the future.

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[jopperm]

@intel/llvm-gatekeepers CUDA failure seems unrelated; could you merge this please?

[bader]

@intel/llvm-gatekeepers CUDA failure seems unrelated; could you merge this please?

Something is going on with CUDA testing. I restarted it, but it failed again.
@intel/dpcpp-devops-reviewers, do we expect CUDA tests to pass?

[aelovikov-intel]

@intel/llvm-gatekeepers CUDA failure seems unrelated; could you merge this please?

Something is going on with CUDA testing. I restarted it, but it failed again. @intel/dpcpp-devops-reviewers, do we expect CUDA tests to pass?

https://github.com/intel/llvm/actions/workflows/sycl_linux_precommit_aws.yml looks healthy, I don't think we should merge with the current results.

[bader]

@intel/llvm-gatekeepers CUDA failure seems unrelated; could you merge this please?

Something is going on with CUDA testing. I restarted it, but it failed again. @intel/dpcpp-devops-reviewers, do we expect CUDA tests to pass?

https://github.com/intel/llvm/actions/workflows/sycl_linux_precommit_aws.yml looks healthy, I don't think we should merge with the current results.

Third re-run passed. 🤷

[sommerlukas]

@bader Are you ok with merging this?

The latest run of the CUDA pipeline passed and I don't see how the non-functional changes in this PR could cause any regression.

[bader]

Yes.