neilb@xxxxxxxxxxxxxxx wrote:
mount the filesystem over nfs open a file several directories deep. e.g. exec 3< /mnt/cramfs/my/long/path/name/file unmount/remount the file system on the server: exportfs -avu umount /test mount /test exportfs -av
try to read from the file
And my question is this: what keeps the superblock of the filesystem on the server from going away when unmounted (assuming it was the last instance). I have been unable to locate in the code where any reference counts pertaining to the superblock are incremented as a result of opening a file. Obviously however, something must keep the superblock from going away or the final read would fail. Even if by some means the inode was preserved across mounts the field inode->i_sb would become invalid if the original SB was destroyed. -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/