Assembly JLE jmp instruction example
The x86 assembly uses a system of bit-flags that represent the result of comparisons. The conditional jump instructions use these flags when deciding whether to perform the jump or not.
In your case you'd use the following two instructions:
cmp ebx, 10 ; compare EBX and 10
jle label ; jump if EBX is "less than or equal" to 10
…
label:
…
JB - work with unsigned numbers (Jump Below) <
JL - work with signed numbers
mov bx,0 // BX := 0
cmp bx,FF // 0 < -1 or 0 < 255 (Jump Flag and Sign Flag will change)
jl butter // if you use JL jump will not occurs, cus 0 > -1
jb butter // if you use JB jump will occurs, cus 0 < 255
JLE instruction conducts two tests:
- Signed Flag (
SF) != Overflow Flag (OF) - Zero flag (
ZF) == 1
If Zero flags is 1 and Signed Flag and Overflow Flag are not equal, then the short relative jump will be executed.
Maybe just a word how CMP instruction works. CMP instruction is like SUB (subtract), but the destination register will not be updated after exsecution. So the following code will perform the same result like CMP ebx, 10. CMP and SUB instruction affect to flags: Carry, Parity, Auxiliary, Zero, Sign and Overflow flags.
push ebx //store ebx value to stack
sub ebx, 10
pop ebx //restore ebx value from stack
The jump itself checks the flags in the EFL register. These are usually set with TEST or CMP(or as a side effect of many other instructions).
CMP ebx,10
JLE there
- CMP corresponds to calculating the difference of the operands, updating the flags and discarding the result. Typically used for greater/smaller checks
- TEST corresponds to calculating the binary AND of the operands, updating the flags and discarding the result. Typically used for equality checks.
See also: The art of assembly language on CMP
As a sidenote: You should get the Intel reference manuals. In particular the two part "Intel® 64 and IA-32 Architectures Software Developer's Manual Volume 2: Instruction Set Reference" which describes all x86 instructions.