On Wed, Nov 11, 2026 at 03:55:55PM +0100, Christoph Hellwig wrote:
> But i wonder if we want a thin wrapper around ->read_iter/_write_iter
> to do this check so that it matches the normal read/write path?
Yes, I had thought about that as well. We could use
vfs_iocb_iter_{read,write}() for that but it's not thin at all and
brings in the LSM into these codepaths where they shouldn't meddle.
So maybe as a follow-up series with a new
vfs_iocb_iter_{read,write}_raw() or whatever. Feel free to make a
suggestion.
