问题
I am working through Richard Detmer's Assembly Language book.
The first chapter states:
A borrow occurs in the subtraction a - b when b is larger than a as unsigned numbers. Computer hardware can detect a borrow in subtraction by looking at whether a carry occurred in the corresponding addition. If there is no carry in the addition, then there is a borrow in the subtraction. If there is a carry in the addition, then there is no borrow in the subtraction.
The carry flag is the 0th bit of the EFL register.
Suppose we want to perform 195D - 618D = -423D as a subtraction operation. There is a borrow and hence the carry flag should not be set.
The following asm code compiles and runs, yet after sub rax, 618, the carry flag is indeed set.
The corresponding addition would be 00C3h + FD96h and this does not involve a carry since the final pairwise addition is 0 + F with no carry into it and hence, there is no carry out of the final pairwise addition.
.DATA
number QWORD 195
sum QWORD ?
.CODE
main PROC
mov rax, number ; 195 to RAX
sub rax, 618 ; subtract 618
;at this point, however, the carry flag is indeed set to 1. Why is this?
mov sum, rax ; sum to memory
mov rax, 0 ; return code
ret ; exit to operating system
main ENDP
END
I am unclear about how this could be.
Any help will be appreciated.
回答1:
First understand that there is unsigned integer arithmetic (where overflow is indicated by the carry flag) and signed integer arithmetic (where overflow is indicated by the overflow flag).
The same addition and subtraction instructions are used for both unsigned and signed integer arithmetic. The only difference is which flags you test afterwards and how you interpret the result (e.g. as -0x0000423D or as 0xFFFFBDC3).
A borrow is also indicated by the carry flag. This means that a borrow happens whenever an unsigned integer overflow happened. For 0x0000195D - 0x0000618D there is an unsigned integer overflow (unsigned integers can't be negative), so the carry flag will be set (but there wasn't a signed integer overflow, so the overflow flag won't be set). The result will be 0xFFFFBDC3 or -0x0000423D depending on whether the result was supposed to be signed or unsigned.
来源：https://stackoverflow.com/questions/47333458/assembly-x86-64-setting-carry-flag-for-sub-instruction