CMSI 387
OPERATING SY S T E M S
Spring 2009
File System Implementation Supplementary Exercise
Due to a combination of time constraints and the need to have sudo capability on a Linux machine, this ex-
ercise is not an official assignment. Get together with a classmate who has sudo access to a machine and
walk through this, in order to get some hands-dirty quality time with the ext2 file system.
The Short Version
Make an ext2 disk image, mount it, put some files
on it, print the user view of the file system (i.e., a
series of ls invocations), dump the disk image to
hex, and identify, at the hex level, the various sec-
tions listed in Step 6 of The Long Version.
The Long Version
1. To create the disk image, you’ll need to learn
how to use the dd (“disk dump”) command.
The following example creates a new file called
image consisting of 1024 default-size blocks,
and initializes its contents with zeroes:
dd if=/dev/zero of=image count=1024
2. You should now have a file that is equivalent to
a brand-new, unformatted disk. “Format” it be
installing an empty ext2 file system on it:
mke2fs image
3. Mount the disk image — this is what requires
sudo access:
mount -o loop -o nosuid -o nodev image mountpoint
…where mountpoint is the directory under
which you’d like to mount image. You can use
df to verify that your command worked. To
unmount the disk image, do:
umount mountpoint
Again, df will tell you if all went well.
4. Create the following items within that mounted
file system:
a. A non-empty text file at the top-level direc-
tory of the file system
b. A directory at the top-level directory of the
file system
c. A second non-empty text file inside that
subdirectory (give it different content so
you can differentiate the two files)
d. A symbolic link inside that subdirectory to
the text file in the top-level directory
e. A hard link from the top-level directory to
the text file in the subdirectory
5. Run a series of ls commands on the now-
populated file system, and note the output.
Feel free to use various ls switches (e.g., –F, –l,
–a, –i, etc.) to see as much interesting informa-
tion as possible.
6. Dump the disk image file to hex using hexdump
–C, then try to identify these items:
a. The disk image’s superblock
b. The directory entries for the files, links, and
directories that you created
c. Where applicable, the inodes for the items
that you created
d. Where applicable, the data blocks occupied
by these items
Get a feel for that and enjoy the hacker buzz :)
