On 8/31/23 09:40, Christoph Hellwig wrote: > On Wed, Aug 30, 2023 at 08:15:17PM +0200, Bernd Schubert wrote: >> While adding shared direct IO write locks to fuse Miklos noticed >> that file_remove_privs() needs an exclusive lock. I then >> noticed that btrfs actually has the same issue as I had in my patch, >> it was calling into that function with a shared lock. >> This series adds a new exported function file_needs_remove_privs(), >> which used by the follow up btrfs patch and will be used by the >> DIO code path in fuse as well. If that function returns any mask >> the shared lock needs to be dropped and replaced by the exclusive >> variant. > > FYI, xfs and ext4 use a simple IS_NOSEC check for this. That has > a lot more false positives, but also is a much faster check in the > fast path. It might be worth benchmarking which way to go, but > we should be consistent across file systems for the behavior. > Thanks, interesting! It is basically the same as my file_needs_remove_privs, as long as IS_NOSEC is set. I can remove the new function and export and to change to that check. Not sure if it is that much faster, but should keep the kernel a bit smaller. Thanks, Bernd
