https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123583

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |ASSIGNED
      Known to work|                            |6.1.0, 6.5.0
            Summary|Named address space         |[13/14/15/16 Regression]
                   |qualifier causes undeclared |Named address space
                   |identifier error            |qualifier causes undeclared
                   |                            |identifier error
   Last reconfirmed|                            |2026-01-14
      Known to fail|                            |15.2.0, 16.0, 7.1.0, 7.4.0
   Target Milestone|---                         |13.5
           Keywords|                            |rejects-valid

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
here is a x86 testcase which shows the same issue (and should be included in
the patch too):
```
void h() {
    if(1)
      ;
  __seg_gs const int *x = 0;
}
```

And it used to work in GCC 6.4.0. Started to fail in GCC 7.1.0. So this is a
regression.

(In reply to Senthil Kumar Selvaraj from comment #2)
> This appears to be a regression introduced by PR 67784.

Yes most likely by r7-272-g2448a956be32d4 (not the first try at the fix).

Reply via email to