[Bug rtl-optimization/127538] New: [16/17 Regression] Failure to merge SUB+CMP --> SUBS on aarch64

[ktkachov at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127538

            Bug ID: 127538
           Summary: [16/17 Regression] Failure to merge SUB+CMP --> SUBS
                    on aarch64
           Product: gcc
           Version: 17.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ktkachov at gcc dot gnu.org
  Target Milestone: ---
            Target: aarch64

#include <cstdint>

// subs x8, x0, x1
// csel x0, x8, x0, hs
uint64_t conditional_subtract(uint64_t x, uint64_t q) {
  return x - (x >= q ? q : 0);
}

On aarch64 this generates:
conditional_subtract(unsigned long, unsigned long):
        cmp     x0, x1
        csel    x1, x1, xzr, cs
        sub     x0, x0, x1
        ret

but before GCC 16 we generated the 2-insn form