d_type and bind mounts

d_type and bind mounts

[Tavian Barnes]

It seems that readdir()/getdents() fill d_type from the underlying
filesystem, not respecting bind mounts of non-directories:

$ touch mount_point
$ sudo mount --bind /dev/null mount_point
$ find -name mount_point -type c
$ find -name mount_point -type f
./mount_point

(Requires a fairly recent GNU findutils to reproduce, older ones
always call stat().)  I've seen similar discussions about d_ino being
for the underlying file, not the mount point, which people have said
is technically a POSIX violation but also unlikely to be fixed.  Is
the same true of d_type?  And is there some workaround a program could
use to get the actual type without the overhead of a whole stat()
call?  For example, a way to tell whether a directory entry is a mount
point?

For reference, the relevant findutils bug is
https://savannah.gnu.org/bugs/?54913.

(Please keep me cc'd, I'm not subscribed to lkml or fsdevel.  Thanks!)

-- 
Tavian Barnes

Re: d_type and bind mounts

[Dave Chinner]

On Mon, Feb 11, 2019 at 02:59:29PM -0500, Tavian Barnes wrote:
> It seems that readdir()/getdents() fill d_type from the underlying
> filesystem, not respecting bind mounts of non-directories:
> 
> $ touch mount_point
> $ sudo mount --bind /dev/null mount_point
> $ find -name mount_point -type c
> $ find -name mount_point -type f
> ./mount_point
> 
> (Requires a fairly recent GNU findutils to reproduce, older ones
> always call stat().)  I've seen similar discussions about d_ino being
> for the underlying file, not the mount point, which people have said

Find is using readdir to look up the directory entry, so what it
finds is the underlying filesystem dirent that is completely unaware
that the dentry cache has an overlayed mountpoint above the
filesystem.  d_type is stored in the underlying directory, so it is
reported from the on-disk information in the underlying filesystem,
not the dentry cache that is aware of the mount point at that
location.

> is technically a POSIX violation but also unlikely to be fixed.  Is
> the same true of d_type?  And is there some workaround a program could
> use to get the actual type without the overhead of a whole stat()
> call? For example, a way to tell whether a directory entry is a mount
> point?

Not that I know of, but that doesn't mean there isn't one. In
general, though, if you are looking for something overlaid on the
underlying dirent (like a mount point) then you have to trigger a
path walk to find whatever is overlaid on that location.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com