With my system, with `kernel-2.6.17-1.2142_FC4', `glibc-2.3.6-3', `e2fsprogs-1.37-4', the command: ./mount -nv -t iso9660 -o ro,norock,mode=0400,uid=ID_OF_REGULAR_USER DEVICE DIR yielded the same results: `mode=' is honored in permissions of regular files and ignored in ones of directories, leaving the latter too wide. Agree that should clarify initial bug report this way. And it may be inacceptable to let anyone read (or even search) mounted directories for exactly the same reasons that call for not permitting to read file data. This permission changes as above prove that `mode=' does affect what is passed to system call, and kernel does part of what is desired in response. It is still unknown: does system call (since some version) have additional argument fields to specify the desired permissions of mounted directories? If yes, not passing them is a bug of `mount'. (If kernel / system calls documentation does not require that explicitly, it is also bug of this documentation.) If no such fields even exist, this is solely bug in kernel. - To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
