On Sat, Aug 20, 2022 at 07:10:38PM +0100, Al Viro wrote: > This is mostly whack-a-mole stuff - a bunch of places > are passing struct path pointers around, without bothering > to mark them const, even though they are not going to try > and modify the contents of struct path. > > It's a bad practice, since there are invariants along > the lines of "file->f_path stays unchanged open-to-release" > and verifying those can get very unpleasant when you are > forced to take detours down the long call chains that could've > been avoided. > > Patches in that pile are independent from each other > and if anyone wants to grab some of them into subsystem's > tree - just say so; I'll be happy to exclude those from the > vfs.git branch if they go into another tree. > > Currently they are in vfs.git#work.path; individual > patches in followups. > Looks good, Reviewed-by: Christian Brauner (Microsoft) <brauner@xxxxxxxxxx>
