https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121136
Bug ID: 121136
Summary: Missed optimization: (x <= 0xFFFFF) in '-Os' mode can
convert to ((x >> 20) == 0)
Product: gcc
Version: 15.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: other
Assignee: unassigned at gcc dot gnu.org
Reporter: Explorer09 at gmail dot com
Target Milestone: ---
Test code
```c
#include <stdbool.h>
#include <stdint.h>
bool func1(uint32_t x) {
return x < 0x100000U;
}
bool func2(uint32_t x) {
return ((x >> 20) == 0);
}
bool func3(uint32_t x) {
return ((x / 0x100000U) == 0);
}
```
The three functions are equivalent. Specifically the threshold constant to be
compared with is a power-of-two. If there is no shortage of temporary
registers, then the right shift version (that is, func2) will likely make the
smallest code.
The actual result I tested in gcc is:
gcc can recognize func3 and func1 are equivalent (it optimizes func3 to func1),
but missed func2. Also it doesn't realize func2 can make the smallest code (at
least in x86-64 and RISC-V targets; it likely misses for ARM too but I didn't
test it thoroughly and so can't tell for sure). I assume `-Os` optimization
mode here.
This is the assembly I expected for x86-64:
```
shrl $20, %edi
sete %al
ret
```
And for RV64:
```
srliw a0,a0,20
seqz a0,a0
ret
```
(This issue report is related to bug 115529 - maybe I can consider this a
sub-issue of it.)
