I'm trying to understand how switch statement works in assembly and I have the following program :
int main (int argc, char **argv)
{
int x = argc > 1 ? atoi(argv[1]) : 2;
int y;
switch (x) {
case 0:
y = 0;
break;
case 1:
y = 1;
break;
case 2:
y = 2;
break;
case 3:
y = 3;
// this case "drops through"
case 4:
y = 4;
break;
default:
y = -1;
break;
}
return 0;
}
In assembly it looks like :
0x804842f <main+19> cmpl $0x1,(%eax)
0x8048432 <main+22> jle 0x804844a <main+46>
0x8048434 <main+24> mov 0x4(%eax),%eax
0x8048437 <main+27> add $0x4,%eax
0x804843a <main+30> mov (%eax),%eax
0x804843c <main+32> sub $0xc,%esp
0x804843f <main+35> push %eax
0x8048440 <main+36> call 0x8048310 <atoi@plt>
0x8048445 <main+41> add $0x10,%esp
0x8048448 <main+44> jmp 0x804844f <main+51>
0x804844a <main+46> mov $0x2,%eax
0x804844f <main+51> mov %eax,-0xc(%ebp)
0x8048452 <main+54> cmpl $0x4,-0xc(%ebp)
0x8048456 <main+58> ja 0x8048492 <main+118>
0x8048458 <main+60> mov -0xc(%ebp),%eax
0x804845b <main+63> shl $0x2,%eax
0x804845e <main+66> add $0x8048540,%eax
My first question is why there is ja instruction in <main+58>, instead of jg, as we are using signed integers. And second question is why there is a shift by 2 bits in <main+63>and not any other value.
The ja is a trick to fold the two signed checks into a single unsigned check. That effectively does if (x > 4 || x < 0), so the only things passed through are 0-4 as expected.
The shift by 2 bits is scaling for the pointer size. You have cut off the next part of the disassembly where I am pretty sure you would see an indirect jump through a table. That table holds pointers of size 4 (on your architecture) so the index should be scaled by 4, which is a shift of 2 bits.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With