| Issue |
136816
|
| Summary |
why clang use Ltmp lable
|
| Labels |
clang
|
| Assignees |
|
| Reporter |
LukeSTM
|
example:
```
#include <stdio.h>
typedef struct {
volatile int counter;
} Counter;
void foo(Counter *v, int i) {
int result, val, tmp;
__asm__ __volatile__(
".arch_extension mp\n"
"9998: " "pldw\t%a0" "\n"
" .pushsection \".alt.smp.init\", \"a\"\n"
" .long 9998b - .\n"
" " "pld\t%a0" "\n"
" .popsection\n"
"1: ldrex %0, [%4]\n"
" add %1, %0, %5\n"
" strex %2, %1, [%4]\n"
" teq %2, #0\n"
" bne 1b"
: "=&r" (result), "=&r" (val), "=&r" (tmp), "+Qo" (v->counter)
: "r" (&v->counter), "Ir" (i)
: "cc"
);
printf("Updated counter: %d\n", v->counter);
}
int main() {
Counter c = {0};
printf("Initial counter: %d\n", c.counter);
foo(&c, 5);
return 0;
}```
https://godbolt.org/z/Px35ejvbn
Clang will split the `foo` function in the example into two parts, using `Ltmp0` to represent the embedded assembly logic in the source code. Why is this done?
![]()
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
