Use c10 version of half/bfloat16 in executorch (#8320)

Pull Request resolved: https://github.com/pytorch/executorch/pull/7040

Pull Request resolved: https://github.com/pytorch/pytorch/pull/144111

Accomplished by importing relevant files from c10 into
executorch/runtime/core/portable_type/c10, and then using `using` in
the top-level ExecuTorch headers. This approach should keep the
ExecuTorch build hermetic for embedded use cases. In the future, we
should add a CI job to ensure the c10 files stay identical to the
PyTorch ones.
ghstack-source-id: 265190624
@exported-using-ghexport

Differential Revision: [D66106969](https://our.internmc.facebook.com/intern/diff/D66106969/)

Co-authored-by: Github Executorch <github_executorch@arm.com>