On Thu, Oct 28, 2021 at 1:46 AM Daniel Borkmann <daniel@xxxxxxxxxxxxx> wrote:
>
> [ Adding Miklos & Greg to Cc for review given e0e0be8a8355 ("libfs: support RENAME_NOREPLACE in
> simple_rename()"). If you have a chance, would be great if you could take a look, thanks! ]
>
> On 10/21/21 5:15 PM, Lorenz Bauer wrote:
> > Allow atomic exchange via RENAME_EXCHANGE when using simple_rename.
> > This affects binderfs, ramfs, hubetlbfs and bpffs. There isn't much
> > to do except update the various *time fields.
> >
> > Signed-off-by: Lorenz Bauer <lmb@xxxxxxxxxxxxxx>
> > ---
> > fs/libfs.c | 6 +++++-
> > 1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/fs/libfs.c b/fs/libfs.c
> > index 51b4de3b3447..93c03d593749 100644
> > --- a/fs/libfs.c
> > +++ b/fs/libfs.c
> > @@ -455,9 +455,12 @@ int simple_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
> > struct inode *inode = d_inode(old_dentry);
> > int they_are_dirs = d_is_dir(old_dentry);
> >
> > - if (flags & ~RENAME_NOREPLACE)
> > + if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE))
> > return -EINVAL;
> >
> > + if (flags & RENAME_EXCHANGE)
> > + goto done;
> > +
This is not sufficient. RENAME_EXCHANGE can swap a dir and a
non-dir, in which case the parent nlink counters need to be fixed up.
See shmem_exchange(). My suggestion is to move that function to
libfs.c:simple_rename_exchange().
Thanks,
Miklos
