Stage 2 : Understanding the Filesystem (2 Hours)

The eXpOS package that you had downloaded in the previous stage consists mainly of a machine simulator. The machine is called the eXperimental String Machine (XSM) and consists of a processor, memory and disk. Some support tools that help you to program the machine are also provided.

One important point to note about the system is that the machine is a bare, and comes with no software in it (except for a boot ROM). Hence, the only way to insert some software code into the system is to prepare the code "outside" (that is, in your Linux/Unix system) and insert your code into the machine. The support tools provided along with the package are precisely designed to help you with this task.

The package comes with three major support tools - two compilers and a disk interface tool called XFS-Interface. The compilers allow you to write high level code and translate it into the XSM machine code. We will look at them in later stages. The XFS-Inteface tool helps you to transfer files between your Linux/Unix system and the XSM machines disk.

XSM machine's disk contains 512 blocks, each capable of storing 512 words. When files are stored in the disk, some format has to be followed so that one can figure out where in the disk are the blocks of a file located.

XSM disk is formatted to what is known as theeXpFS file system format. The format specifies how data as well as meta-data for each file stored in the disk must be organized. The XFS interface tool allows you to load data files (and executable files as well) from your Linux/Unix system into the XSM disk in accordance with the eXpFS format.

The eXpFS format specifies that each data/executable file can span across at most four data blocks, and that the index to these blocks along with the name and the size of the file must be stored in a pre-define area of the disk called the Inode table. (The inode table is stored in disk blocks 3 and 4). There are also other pre-defined areas of the disk that stores meta data about the disk (see description of the root file and the disk free list for more details). When you use XFS interface to load a file from your Linux/Unix system to the XSM disk, the interface tool will correctly fill all the required meta data information as stipulated by the eXpFS format.

The eXperimental Filesystem (eXpFS) is a simulated filesystem. A UNIX file named "disk.xfs" simulates the hard disk of the XSM machine. Building eXpOS begins with understanding the underlying filesystem (eXpFS) and its interface (xfs-interface) to the host (UNIX) environment. The xfs-interface is used for transferring files between your linux system and the xsm disk.

In this stage, you will create a text file and load it to the XFS disk using xfs-interface.

Run the XFS Interface

cd $HOME/myexpos/xfs-interface
./xfs-interface

Start by formatting the disk to the eXpOS file system format in the XFS interface using fdisk command.

The fdisk< command converts the raw disk into the filesystem format recognised by the eXpOS operating system. It initialises the disk data structures such as disk free list, inode table, user table and root file.

Type the following commands in the xfs-interface prompt.

# fdisk
# exit

You will be back in the UNIX shell and a file named disk.xfs is created in the location $HOME/myexpos/xfs-interface/. This UNIX file simulates the hard disk of the XSM machine. The disk is formatted to eXperimental File System (eXpFS) (see eXpFS Specification).
The XSM machine's disk is a sequence of 512 blocks, each block capable of holding 512 words (see Disk Organization). The second block of the formatted disk contains a disk free list which is explained below.

The Disk Free List in XFS is a data structure which keeps track of used and unused blocks in the disk. An unused block is indicated by 0 and a used block is indicated by 1. Check the contents of the Disk Free List after formatting the disk. Use the df command to view the Disk Free List (stored in disk block number 2). The output will be as follows:

0    -   1
1    -   1
2    -   1
3    -   1
4    -   1
5    -   1
6    -   1
7    -   1
8    -   1
9    -   1
10   -   1
11   -   1
12   -   1
13   -   1
14   -   1
15   -   1
16   -   1
17   -   1
18   -   1
19   -   1
20   -   1
21   -   1
22   -   1
23   -   1
24   -   1
25   -   1
.
.
No of Free Blocks = 443
Total No of Blocks = 512

The first 69 blocks (blocks 0 to 68) are reserved for stroing various OS data structures and routines as well as Idle code, INIT program, etc (see Disk Organization). Hence the Disk Free List entries for these are marked as 1 (used) and the remaining entries for blocks 69 to 511 are 0 (unused).

Create a file in your UNIX machine with sample data. A sample data file is given below:

There is a place where the sidewalk ends
And before the street begins,
And there the grass grows soft and white,
And there the sun burns crimson bright,
And there the moon-bird rests from his flight
To cool in the peppermint wind.

Save the file as $HOME/myexpos/sample.dat

Load this data file $HOME/myexpos/sample.dat to the XFS disk from your UNIX machine. This can be done by the following commands:

cd $HOME/myexpos/xfs-interface
./xfs-interface

This will take you to the xfs-interface prompt. Type the following commands.

# load --data $HOME/myexpos/sample.dat

This will load the file to the XFS disk and the following updations happen in disk data structures :

1) A disk block will be allocated for the file (as sample.dat contains less than 512 words) and corresponding to this allocated block (here block 69 - this is because the 1^st free block is allocated by the allocator), an entry will be marked as 1 (used) in the Disk Free List.

2) An entry in the InodeTable will be created for this file. Inode Table contains information such as the file type, file name, file size, userid, permission and the block numbers of the data of data files loaded through xfs-interface is the root. Userid is the index of the user entry in the User Table.The userid of root is 1 and hence the userid field in the inode table is set to 1 for all data files loaded through the xfs interface. The permission is set to open(1). Note that any file in eXpFS file system is permitted to have a maximum of four data blocks.

3) An entry for this file will be made in the Root File also.

Before proceeding further you must be clear about eXpFS (eXperimental File System). In the following steps we will see the above mentioned updations.

Find out the block numbers of the Data Blocks corresponding to the loaded file. Use the copy command to copy the Inode Table(Inode Table is stored in disk blocks 3 and 4) to a UNIX file (say $HOME/myexpos/inode_table.txt).

# copy 3 4 $HOME/myexpos/inode_table.txt
# exit

Now check the Inode table entry for the file sample.dat in the UNIX file inode_table.txt and find the block numbers of its data blocks. The contents of the file inode_table.txt will be as follows:

1
root
512
0
0
-1
-1
-1
5
-1
-1
-1
-1
-1
-1
-1
2
sample.dat
19
1
1
-1
-1
-1
69
-1
-1
-1
-1
-1
-1
-1
-1
.
.
.

# dump --inodeusertable

This will write the contents of the inodetable into the file $HOME/myexpos/xfs-interface/inodeusertable.txt

Now check the contents of the disk free list and verify that the entry for the 69^th block is marked as used. This corresponds to the Data Block 1 of sample.dat.

Copy the data blocks from the XFS disk and display it as a UNIX file $HOME/myexpos/data.txt.

# copy 69 69 $HOME/myexpos/data.txt

You will get back the contents of the file $HOME/myexpos/sample.dat in $HOME/myexpos/data.txt. However in $HOME/myexpos/data.txt, each word is displayed in a line because a word in XFS is 16 characters long. Sample data.txt file is shown below.

There is a plac
e where the sid
ewalk ends

And before the
street begins,

And there the g
rass grows soft
and white,

And there the s
un burns crimso
n bright,

And there the m
oon-bird rests
from his flight


To cool in the
peppermint wind

xfs-interface provides the export command to export files from the XSM machine to the UNIX machine in a single step. Export the file sample.dat to the UNIX file $HOME/myexpos/data.txt using xfs-inteface as shown below and verify that the contents are same as sample.dat.

# export sample.dat $HOME/myexpos/data.txt