[Bug tree-optimization/126405] New: [15/16/17 Regression] Wrong code at -O2 on aarch64

xxs_chy at outlook dot com via Gcc-bugs <[email protected]>
Newsgroups gmane.comp.gcc.bugs
Message-ID <[email protected]/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126405

            Bug ID: 126405
           Summary: [15/16/17 Regression] Wrong code at -O2 on aarch64
           Product: gcc
           Version: 17.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: xxs_chy at outlook dot com
  Target Milestone: ---

Reproducer: https://godbolt.org/z/1jYenKv76
Testcase:
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
typedef int64_t v16i64 __attribute__((vector_size(128)));
int64_t g2, __chk, g12;
__attribute__((__vector_size__(16 * sizeof(int)))) int g3;
v16i64 g18;
void *g27;
void f31()
{
lbl_br1:
    g18 = ~g18;
    g3 = ~g3;
    if (g2) goto lbl_br1;
lbl_b5:
    switch (g12)
    case 4:
    case 0:
        goto lbl_sw8;
    abort();
lbl_sw8:
    if (g27) goto lbl_b5;
    g18 = ~g18;
}
int main()
{
    f31();
    __chk = g18[15];
    printf("checksum=0x%016llx\n", (unsigned long long)__chk);
}

This case outputs an indeterminate value at -O2, and outputs
checksum=0x0000000000000000 at -O0.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.