I have some questions about moving mount of the linux kernel implementation.
In do_move_mount of fs/namespace.c, we donít want to move a mount residing in a shared parent.
As the comment and code
// Don't move a mount residing in a shared parent
if( IS_MNT_SHARED(old->mnt_parent) )
in the following link
It seems this may unmount of all the trees under their corresponding peer mounts.
I just want to know why we donít want to unmount all their corresponding peer mounts. Does anyone have some ideas?