| Issue |
171676
|
| Summary |
constant addition not merged into subtract-with-borrow
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
purplesyringa
|
[Godbolt](https://godbolt.org/z/rWY51x7GE)
```cpp
#include <stdint.h>
uint64_t f(uint64_t a, uint64_t b) {
uint64_t x;
x += __builtin_add_overflow(a, b, &x);
return x + 10;
}
uint64_t g(uint64_t a, uint64_t b) {
uint64_t x;
x -= __builtin_sub_overflow(a, b, &x);
return x + 10;
}
```
On x86-64, I expected the lowering of `f` to use `adc ..., 10` and the lowering of `g` to use `sbb ..., -10`. `f` works, `g` doesn't:
```asm
f(unsigned long, unsigned long):
mov rax, rdi
add rax, rsi
adc rax, 10
ret
g(unsigned long, unsigned long):
sub rdi, rsi
sbb rdi, 0
lea rax, [rdi + 10]
ret
```
Same on aarch64. The addition of 10 could be integrated into `sbc`:
```asm
f(unsigned long, unsigned long):
mov w8, #10
adds x9, x0, x1
adc x0, x9, x8
ret
g(unsigned long, unsigned long):
subs x8, x0, x1
sbc x8, x8, xzr
add x0, x8, #10
ret
```
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
