On Wed, Nov 6, 2019 at 6:03 PM Darrick J. Wong <darrick.wong@xxxxxxxxxx> wrote:
>
> On Wed, Nov 06, 2019 at 07:58:55AM +0200, Amir Goldstein wrote:
> > For efficient check if file has xattrs.
> >
> > Signed-off-by: Amir Goldstein <amir73il@xxxxxxxxx>
> > ---
> > io/attr.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/io/attr.c b/io/attr.c
> > index b713d017..ba88ef16 100644
> > --- a/io/attr.c
> > +++ b/io/attr.c
> > @@ -37,6 +37,7 @@ static struct xflags {
> > { FS_XFLAG_FILESTREAM, "S", "filestream" },
> > { FS_XFLAG_DAX, "x", "dax" },
> > { FS_XFLAG_COWEXTSIZE, "C", "cowextsize" },
> > + { FS_XFLAG_HASATTR, "X", "has-xattr" },
> > { 0, NULL, NULL }
> > };
> > #define CHATTR_XFLAG_LIST "r"/*p*/"iasAdtPneEfSxC"
>
> /me wonders if this should have /*X*/ commented out the same way we do
> for "p".
Sure. Eric, please let me know if you want a re-submit for this.
>
> Otherwise, the patch looks ok to me...
>
> /me *also* wonders how many filesystems fail to implement this flag but
> support xattrs.
>
> Oh. All of them. Though I assume overlayfs is being patched... :)
>
It doesn't need to be patched. It doesn't have xattr storage of its own.
The ioctl is passed down to the underlying fs (*).
(*) The ioctl is currently blocked on overlayfs directories.
Thanks,
Amir.
