In this project, you will design and implement a simplifies UNIX like file system in C or C++. This project does not use MIPS or SPIM, so it is completely independent of your other homework assignments. This will count as midterm exam for this course. Part 1
Design a file system that uses i-nodes blocks and data blocks to keep your files. Your file system will use, a i-node structure like Fig 4.33 of your textbook (single link, double link and triple link i-nodes). Your directory, i-node block, data block structure will be similar to Fig 4.34 of the textbook. Your file attributes will include size, last modification date and time, and name of the file. No permissions or owner attributes will be kept. Write a design report that specifies the following
- Define your directory structure and directory entries;
- Define how you keep the free blocks and free i-nodes;
- Define your i-node structure
- Define your superblock that contains crucial information about the file system such as the block size, i-node positions, block positions, etc.
Your report should include the function names of your source code that handles the file system operations listed in the table of Part 3.
Part 2
Write a C/C++ program that creates an empty file system as a 1 MB Linux file. This file will include all the information about your file system including the i-nodes, data blocks, free blocks and i-nodes, directories, data, etc. The sample run of the program will be like
makeFileSystem 4 400 mySystem.dat
where 4 is the block size of the file system in KB for both data blocks and i-node blocks, and 400 is the number of free i-nodes for an empty file system. mySystem.dat is the Linux file that contains all the file system. When you
work on the file system, this file contains all the information for the file system. Note that the size of mySystem.dat will be exactly 1 MB all the time whether it contains any information or not. Part 3
You will write a program that performs file system operation on the file system. The program will work like following
fileSystemOper fileSystem.data operation parameters
where fileSystemOper is your program, fileSystem.data is the file system data file that you have created in Part 2. You will keep modifying the same fileSystem.data file for all your operations. Allowable operations and parameters for these operations are given below in the following table.
Operation | Parameters | Explanation | Example |
list | Path | Lists the contents of the directory shown by path on the screen. | fileSystemOper fileSystem.data list / lists the contents of the root directory. The output will be like ls l |
mkdir rmdir | Path and dir name | Makes or removes a directory | fileSystemOper fileSystem.data mkdir /usr/ysa makes a new directory under the directory ysa if possible. These two works exactly like mkdir and rmdir commands of Linux shell |
dumpe2fs | None | Gives information about the file system. | fileSystemOper fileSystem.data dumpe2fs works like simplified and modified Linux dumpe2fs command. It will list block count, i-node count, free block and i-nodes, number of files and directories, and block size. Different from regular dumpe2fs, this command lists all the occupied i-nodes, blocks and the file names for each of them. |
write | Path and file name | Creates and writes data to the file | fileSystemOper fileSystem.data write /usr/ysa/file linuxFile Creates a file named file under /usr/ysa in your file system, then copies the contents of the Linux file into the new file. This works very similar to Linux copy command. |
read | Path and file name | Reads data from thefile | fileSystemOper fileSystem.data read /usr/ysa/file linuxFile Reads the file named file under /usr/ysa in your file system, then writes this data to the Linux file. This again works very similar to Linux copy command. |
del | Path and file name | Deletes file from the path | fileSystemOper fileSystem.data del /usr/ysa/file Deletes the file named file under /usr/ysa in your file system. This again works very similar to Linux del command. |
Operations below are bonus (20 points) | |||
ln | Source and destination path and file names | Hard linking between2 files | fileSystemOper fileSystem.data ln /usr/ysa/file1 /usr/ysa/file2 Allows more than one filename to refer to the samefileLinux ln command |
lnsym | Source and destination path and file names | Symbolic linking between 2 files | fileSystemOper fileSystem.data lnsym /usr/ysa/file1 /usr/ysa/file2 Linux ln-s command |
fsck | none | Simplified File system check | fileSystemOper fileSystem.data fsck Just print the two table of Fig 4.27 of your textbook for both i-nodes and blocks |
Here is a sequence file system operation commands that you can use to test your file system. Suppose you have a file named linuxFile.data in your Linux current directory.
makeFileSystem 4 400 mySystem.data fileSystemOper fileSystem.data mkdir /usr fileSystemOper fileSystem.data mkdir /usr/ysa fileSystemOper fileSystem.data mkdir /bin/ysa ; Should print error! fileSystemOper fileSystem.data write /usr/ysa/file1 linuxFile.data fileSystemOper fileSystem.data write /usr/file2 linuxFile.data
fileSystemOper fileSystem.data write /file3 linuxFile.data
fileSystemOper fileSystem.data list / ; Should list 1 dir, 1 file fileSystemOper fileSystem.data del /usr/ysa/file1 fileSystemOper fileSystem.data dumpe2fs
fileSystemOper fileSystem.data read /usr/file2 linuxFile2.data
cmp linuxFile2.data linuxFile.data ; Should not print any difference
fileSystemOper fileSystem.data ln /usr/file2 /usr/linkedfile2 fileSystemOper fileSystem.data list /usr
fileSystemOper fileSystem.data write /usr/linkedfile2 linuxFile.data fileSystemOper fileSystem.data dumpe2fs
fileSystemOper fileSystem.data lnsym /usr/file2 /usr/symlinkedfile2 fileSystemOper fileSystem.data list /usr fileSystemOper fileSystem.data del /usr/file2 fileSystemOper fileSystem.data list /usr ; Should see likedfile2 is there, symlinkedFile2 is there but.. fileSystemOper fileSystem.data write /usr/symlinkedfile2 linuxFile.data ; Should print error!
fileSystemOper fileSystem.data dumpe2fs
Notes
- Always be careful about the errors, such as bad block sizes, bad file names, non-existent files or directories, etc.
- Run experiments that uses up all of your i-nodes and data blocks.
- Try to get fragmentation and show your fragmented file system using the dumpe2fs
- Do not use any code from any other source even a single line!
Reviews
There are no reviews yet.