| OpSys Spring 2005 - HW2 |
| OpSys Home | Assignment | Requirements | Submitting | Resources |
diskusage
commands
-
diskusage command for Unix and Windows (2 programs)
The objectives of this assignment are:
You are to write two programs, one that runs on the CS FreeBSD machines and the other that will run on a PC running Windows. The Windows program will be graded (run) on a machine running Windows XP, but if you develop on something running Windows 98 or 2000 it will work fine under XP.
Both program will print out the size of any files or directories listed on the
command line. For the Unix version, your program should match the
numbers reported by the du command, this is the actual
disk usage including the space used by the directory file itself. For
the Windows version you just need to compute the file sizes (NOT
DISKUSAGE, and you DO NOT INCLUDE the size of a directory file). Both
need to work recursively, so they report on an entire hierarchy - given
a directory name compute the space used by everything in the directory.
Here is an example of running the program:
< ./diskusage build comporg opsys .bashrc 807424 build 1659904 comporg 389120 opsys 4096 .bashrc
In the above example the directory build and it's contents are
using up 807424 bytes, the file .bashrc is using 4096.
There are many ways to write this program, but in general the simplest approach is to write a recursive function that can determine the disk usage of a file or directory, in the case of a directory the function calls itself recursively on each file/directory contained in the directory.
- Unix system calls and library functions
The Unix program will report on the number of bytes of disk space used by a file or by a directory and everything contained in the directory. Your program can receive any number of command line arguments, each is the name of a file or directory. Your program should compute the total disk space used by the file or directory (and it's contents) and print a single line that includes the number of bytes followed by the file/directory name.
NOTE: Your command must include the size of a directory
itself when dealing with directories. You can verify that your program
reports the correct sizes by comparing to the output of the Unix
du command:
> du -s -B 1 build comporg opsys .bashrc 807424 build 1659904 comporg 389120 opsys 4096 .bashrc
In the above example, the Unix du command is passed
some options that make it act the way your program is suppose to act,
you don't need to support any options at all (everything on the
command line is assumed to be a file or directory name).
To get information about a specific file or directory in Unix you
can use the stat() system call. You can view an on-line
version of the man page here:
freebsd
stat(2) man page
To read a directory you need to use opendir and
readdir, both are documented here:
freebsd
opendir(3) man page.
NOTE: The structure created by readdir is a struct
dirent which may contain many fields, although the POSIX
standard indicates that you should only use the field named
d_name which is a null terminated string containing a
file name.
- Windows (win32 api) system calls and library functions
IMPORTANT! As it turns out, it is very difficult to write
a program that computes disk usage on a Windows machine with an NTFS files system (any file
system that can handle sizes > 2GB). The issue is finding out the
block size (called cluster in the Windows world) so you can
round up file sizes (you don't have something like stat
that will just hand you # blocks and blocksize...). SO - don't worry
about reporting actual disk usage, you should report file sizes
without regard to how much disk space is actually used. For reporting
the size of a directory, simply add up the sizes of the files held in
the directory (and within any subdirectories), don't worry about the size
of the directory file itself... (you do have to handle this in the
Unix version!).
Your Windows program will be compiled using gcc and run in a cygwin shell. It is expected that you can test your code under cygwin (feel free to develop using Visual Studio, but make sure it compiles and runs under cygwin). Most RPI laptops (last 2 years at least) include cygwin installed. If you need to install cygwin yourself, you can get everything (programs and documentation) at www.cygwin.com. The download process involves getting and running an installation program in which you can pick and choose what you want to install. You need to tell it you want gcc!
When programming in cygwin, you can use any Windows editor to edit your files, or you can get a cygwin version of emacs/vi/pcio and edit wihtin the cygwin shell window. The gcc command line is just like when running on the FreeBSD machines, other utilities (like Make) also work the same).
The win32 API includes many functions for messing with files and directories - below are some that you may find useful:
FindFirstFile and
FindNextFile -
together these are sortof like opendir and
readdir, they allow you to get at all the files contained
in a directory.
You will need to look at the structure
WIN32_FIND_DATA, it contains information including
file size.
The entire collection of file and directory handling functions is available here: File Management Funcitions (msdn.microsoft.com)
NOTE: We don't care whether you support big files or not.
Feel free to assume
any file or directory is no larger than 2^32 bytes. The
WIN32_FIND_DATA structure supports file sizes up to 2^64
bytes, you can ignore the upper 32 bits (nFileSizeHigh).
- Project Requirements
The following are the requirements for the project:
Your unix program must compile and run on the CS department FreeBSD
machines (freebsd.remote.cs.rpi.edu).
Your Windows program must compile and run under cygwin running on Windows XP
Your Windows progam must use win32 API functions, you can't
use the version of stat provided with cygwin!
Your program must provide the output described above and nothing else (a script will be running your program and analyzing the output produced, if you include debugging statements or anything else, it will get confused!).
Your program must work for any regular file or directory
Your program can ignore anything that is not a regular file or a directory (ignore links, special files, etc).
Your submission must include a Makefile we can use to build both versions. See the HW1 FAQ for information on building a simple Makefile
Your submission must include a file named README that includes the following information:
Grading: Each of the progams is worth 1/2 of the grade. Each program will be graded according to the formula below. Note that to get full credit we must be able to understand your code (it must be commented!)
| 30% | Proper output. For Unix, computes diskusage correctly (matches the du command). For Windows, correctly computes file and/or directory sizes. |
|---|---|
| 10% | Impossible to crash the program. Basically you need to make sure there is nothing we can do (there are not command line arguments) that can cause your program to crash (for example, SEGV). Check the return value of all system calls (check for errors)!. Keep track of dynamically allocated memory (and free anything that is not needed). |
| 10% | Code quality (comments, organization, how hard is it to understand ?). |
You can get partial credit for any part.
If you code does not compile and run under FreeBSD/Cygwin, you will lose at least 1/2 the relevant points ( partial credit will be awarded based on visual inspection of the code).
- How to Submit
Log in to WebCT at webct.rpi.edu using your RCS id and password. Once you get to MyWebCT click on "Operating Systems", and from there go to the homework drop boxes. Submit your files (individually, zipped or tarred) to the drop box labeled HW2. IMPORTANT: make sure yout submit both programs!
-Resources