I found a strange issue when using multiplication (*=) and one of the operands is 1.

[zzdmfk]

C code:
`int printf(const char* str, ...);

int main()
{
int x = 1;
x *= 100;
printf("%d\r\n", x);
}`

asm code:

section .data section .text extern _printf global main ; main function main: push ebp mov ebp, esp sub esp, 16 push dword 1 pop eax mov dword [ebp-4], eax push dword 100 pop eax mov ecx, eax mov eax, [ebp-4] imul eax mov dword [ebp-4], eax lea ebx, [_printf] push ebx pop ebx mov dword [function_call_1], ebx push dword [ebp-4] mov eax, str_2 push eax call [function_call_1] add esp, 8 push eax pop eax push eax add esp, 4 add esp, 16 pop ebp ret section .data function_call_1: dd 0 section .rodata str_2: db '%', 'd', 0

The function printf always outputs 1, but I change int x = 1 to int x = 2, it outputs correctly. Can you tell me why? the multiplication operand can not be 1?

[nibblebits]

*= is handled are you basing the issue from the master repository or an earlier commit?

Thanks

[nibblebits]

Can you format your assembly so its not on one line please

[zzdmfk]

section .data
section .text
extern printf
global main
; main function
main:
push ebp
mov ebp, esp
sub esp, 16
push dword 1
pop eax
mov dword [ebp-4], eax
push dword 100
pop eax
mov ecx, eax
mov eax, [ebp-4]
imul eax
mov dword [ebp-4], eax
lea ebx, [printf]
push ebx
pop ebx
mov ecx, ebx
push dword [ebp-4]
mov eax, str_1
push eax
call ecx
add esp, 8
push eax
pop eax
push eax
add esp, 4
add esp, 16
pop ebp
ret
section .rodata
str_1: db '%', 'd', 13, 10, 0

Above asm code was formated. I wrote code follow your lecture video, even I got the C code from https://github.com/nibblebits/PeachCompiler.git, but the problem still exists.

George Chow

[nibblebits]

Okay, please can you clone the repository from the master and if the same problem persists its a bug. If it doesnt persist then you made a mistake, please let me know

Thanks

[zzdmfk]

Here are detailed steps I did again today:

step 1, git clone https://github.com/nibblebits/PeachCompiler.git

step 2, replace content of test.c with:
int printf(const char* str, ...);

int main()
{
int x = 1;
x *= 100;
printf("%d\n", x);
}

step 3, make clean && make

step4:
george@george-ubuntu:~/Desktop/gitcompiler/PeachCompiler$ ./main
section .data
section .text
extern printf
global main
; main function
main:
push ebp
mov ebp, esp
sub esp, 16
push dword 1
pop eax
mov dword [ebp-4], eax
push dword 100
pop eax
mov ecx, eax
mov eax, [ebp-4]
imul eax
mov dword [ebp-4], eax
lea ebx, [printf]
push ebx
pop ebx
mov dword [function_call_1], ebx
push dword [ebp-4]
mov eax, str_2
push eax
call [function_call_1]
add esp, 8
push eax
pop eax
push eax
add esp, 4
add esp, 16
pop ebp
ret
section .data
function_call_1: dd 0
section .rodata
str_2: db '%', 'd', 10, 0
everything compiled file
/usr/bin/ld: ./test.o: warning: relocation in read-only section `.text'
/usr/bin/ld: warning: creating DT_TEXTREL in a PIE
nasm -f elf32 ./test -o ./test.o && gcc -m32 ./test.o -o ./test

george@george-ubuntu:~/Desktop/gitcompiler/PeachCompiler$ ./test
1

And this is my operate system info:
george@george-ubuntu:~/Desktop/gitcompiler/PeachCompiler$ uname -a
Linux george-ubuntu 5.15.0-57-generic #63-Ubuntu SMP Thu Nov 24 13:43:17 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux

[nibblebits]

Hi,
Yes its a valid bug thank you for reporting it after some debugging ive found the problem: See the debugging window: https://ibb.co/3mTxd0R

The two operands are calculated correctly, the issue lies in that the IMUL instruction should of mulitplied on the ECX register but multiplies on the EAX register which leads to 1 multiplied by 1 causing this problem.

Thanks for reporting this I will look into fixing it.

[nibblebits]

Hello,
Thanks for reaching out, once theres enough issue reports I will go through them all and create a lecture addressing any issues.

Thanks
Dan