In the previous post, we have seen the various Data Structures that the operating System uses in order to manage the files stored in the drives. Since, we are dealing with the system programming, it is important to know about Interrupts, Assembly Language, etc.
Anyways, if you don’t know that much about Interrupts, Assembly Language, the steps for developing the code are simpler to understand.
Let’s start with the practical implementation details.
Now, we want to read the file stored in Floppy Disk (FD). User will provide the name of the file through Command Line Interface (CLI).
Steps Involved:
1) Read the no. of sectors for:
i) FAT
ii) Root Directory (RD)
2) Find the RD entry for the given input file.
3) If RD entry found: Get the first cluster no. (Stored at 1AH offset in RD Entry) and read it. Also read FAT.
Else File Does not exist.
4) Now, read the remaining file by making use of FAT.
Now, we will elaborate these 4 steps in order to get into the rhythm.
Step 1:
Standard FD has predefined size of the Data Structures.
For instance,
Boot sector – 1 Sector
FAT – 9 Sectors
Copy of FAT – 9 Sectors
Root Directory – 14 Sectors
And then file area starts.
As these sizes are known previously, we are not going to retrieve them Boot Sector. Though, we can get the size of FAT from Number of sectors per FAT i.e. 16H offset & size of RD from Number of root-directory entries i.e. 11H offset of Boot Sector.
Note that: Size of RD = No. of RD Entries * 32
As size of each entry is 32 Bytes (See Fig. 3 of Previous Post).
Step 2:
User is going to provide the file name, which we have to search in the RD. For that purpose, we have to read the whole RD and then check first 11 Bytes (8 Bytes file name + 3 Bytes Extension) of each 32 Byte entries in it.
For reading RD, we are going to use INT 25H.
INT 25H (Absolute disk read)
AL = Drive Number (0 = A, 1 = B, etc.)
CX = Number of sectors to read
DX = Starting sector number
DS: BX = Segment: offset of buffer
Returns:
If function successful
Carry flag = clear
If function unsuccessful
Carry flag = set
AX = error code
MOV AL, 00
MOV BX, OFFSET BUFF
MOV CX, 14
MOV DX, 19
INT 25H
How does starting sector no. equal to 19?
RD comes after Boot Sector (1 Sector), FAT (9 Sectors), Copy of FAT (9 Sectors). So, 1+9+9 = 19 and from 19th sector onwards RD starts (as first sector is sector 0).
After the execution of INT 25H, we will get whole RD into buffer BUFF and we start checking first 11 Bytes of each RD entry.
A legal 32 Byte entry does not start with E5H or 00H. So, during file name check, we have to take care of this thing.
Step 3:
As soon as we find the RD entry for which we were looking, read the first cluster no. at an offset 1AH in that entry. Read and display the first cluster. Now FAT comes into the picture for displaying the remaining clusters of the file.
For reading 9 sectors of FAT, we make use of INT 25H as described in step 2. Only change will be in the values of CX, DX and use some other buffer.
Note:
MOV CX, 09
MOV DX, 01
MOV BX, BUFF1
One this should be noted here that we are going to make use of INT 21H and INT 25H only.
Let’s have a look about the 09H and 4CH functions of INT 21H
Function: 09H
It writes a $-terminated string to standard output.
DS must point to the string's segment, and DX must contain the string's offset:
For example:
.DATA
STRING BYTE “THIS IS A STRING$”
.CODE
MOV AH, 09H
MOV DX, OFFSET STRING
INT 21H
Function: 4CH
Ends the current process and returns an optional 8-bit code to the calling process.
A return code of 0 usually indicates successful completion. It is same as EXIT 0.
MOV AX, 4C00H
INT 21H
;Here AL = 00, After the execution register AL (AX = AH || AL) will contain Error Code if present.
Step 4:
This step is more interesting as compared to other steps. We got first cluster of the required file and now we have to find the next cluster no. in order to read the remaining file. If you have observed the nomenclature of the FAT at the time formatting the Pen-Drive, you may have seen FAT12, FAT16 or something like nowadays NTFS (New Technology File System).
Now, What does this FATX represents? What is the significance of X, here?
X represents X-bit wide cluster entries. Now let’s convert X into Bytes.
If FAT12 then X = 1.5 Bytes or If FAT16 then X = 2 Bytes, etc. One thing should be noted that Floppy Disk is FAT12 (i.e. 1.5 Byte).
Let F be the first cluster no. obtained from RD entry of the given input file. Now, we multiply F by 1.5 in order to get the next cluster no. Here, comes the tricky point:
Let after multiplication i.e. 1.5 * F, the four nibbles be wxyz (i.e. 1.5F = wxyz). If 1.5F is even, consider last three nibbles otherwise first three nibbles.
Suppose, 1.5F is even, consider xyz and add first nibble as 0. So, the next cluster no. becomes 0xyz. Now, read the location value from FAT present at 0xyz. From this location, we have to read the next sector of the file. In the similar fashion, we can read remaining file. At any instant of point, the next cluster no. equals to or greater than 0FFFH, stop i.e. file is completely read. Here, we are getting the location of the next sectors to be read from FAT in continuation. This continuation is referred as “Chaining”.
Note: 1 Cluster = 2^i Sector, i = 0,1,2……
Now, try to understand the below ASM code, which is written on the basis of the above discussion. I strongly suggest you people to read “The Microsoft(R) Guide for Assembly Language and C Programmers By Ray Duncan” for better understanding of System Programming.
print macro msg
mov ah,09h
lea dx,msg
int 21h
endm
.model tiny
.code
org 100h
begin:
jmp start
msg1 db 10d,13d,"Error in reading the root directory:$"
msg2 db 10d,13d,"Error in reading the sector:$"
msg3 db 10d,13d,"Found:$"
msg4 db 10d,13d,"Not Found:$"
msg5 db 10d,13d,"Error in reading FAT:$"
buff db 7168 dup('$')
buff1 db 4608 dup('$')
buff2 db 513 dup('$')
filename db 12 dup('$')
str_cls dw ?
temp dw ?
cnt db 224
count db ?
num dw ?
start:
mov count,08
mov si,80h
mov cl,[si]
dec cl
mov si,82h
lea di,filename
l1:
mov al,[si]
cmp al,'.'
je l2
mov [di],al
inc si
inc di
dec count
dec cl
jmp l1
l2:
cmp count,00
jz l4
inc si
dec cl
l3:
mov [di],20h
inc di
dec count
jnz l3
l4:
mov count,03h
l5:
mov al,[si]
mov [di],al
inc si
inc di
dec count
dec cl
jnz l5
cmp count,00
jnz l6
jmp l7 ;filename completed
l6:
mov [di],20h
inc di
dec count
jnz l6
l7:
mov al,00
mov bx,offset buff
mov cx,14
mov dx,19
int 25h
jc error1
add sp,02h
lea si,buff
l8:
mov al,[si]
cmp al,0E5h
je loop1
cmp al,00
je loop1
jmp loop2
loop1:
add si,0032
dec cnt
jnz l8
loop2:
lea di,filename
mov cx,11
l9:
mov al,[si]
cmp al,[di]
je loop3
add si,cx
add si,21
dec cnt
jnz loop2
jmp n_found
loop3:
inc si
inc di
dec cl
jnz l9
jmp found
n_found:
print msg4
jmp end1
found:
print msg3
sub si,11
add si,1Ah
mov bl,[si]
inc si
mov bh,[si]
mov str_cls,bx ;first cluster
mov al,00
mov bx,offset buff1 ;buff1 contains FAT
mov cx,09
mov dx,01
int 25h
add sp,02
jc error2
mov ax,str_cls
jmp en2
;FAT cluster reading
error1: jmp error11
en2:
add ax,31
mov dx,ax
mov al,00
mov cx,01
mov bx,offset buff2
int 25h
jc error3
add sp,02
mov ah,09h
lea dx,buff2 ;displaying cluster contents
int 21h
mov ax,str_cls
mov bx,03
mul bx
mov bx,02 ;1.5 multiplication
div bx
lea si,buff1
add si,ax
mov bl,[si]
inc si
mov bh,[si]
mov num,bx
mov ax,str_cls
and ax,01
cmp ax,00
je even1
jmp odd1
even1:
mov ax,num
and ax,0fffh
cmp ax,0fffh
je end1
mov str_cls,ax
jmp en2
odd1:
mov ax,num
and ax,0fff0h
mov cl,04
shr ax,cl
and ax,0fffh
cmp ax,0fffh
je end1
mov str_cls,ax
jmp en2
error11:
print msg1
jmp end1
error2:
print msg5
jmp end1
error3:
print msg2
end1:
mov ah,4ch
int 21h
end begin
I hope you people like this post.
Please do write comments !!!!