[Bug tree-optimization/128008] wrong code at -O2: store through a restrict member pointing to its own struct type is deleted

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128008

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot 
gnu.org
   Last reconfirmed|                            |2026-10-20

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.   We compute q->next to point to nothing.

  # PT = null 
  _1 = MEM[(struct T *)p_3(D) clique 1 base 1].next;
  MEM[(struct T *)_1 clique 1 base 0].v = 7;

constraints are

p = &PARM_NOALIAS.0+32
PARM_NOALIAS.0+32 = &NONLOCAL
_1 = *p + 64
derefaddrtmp(13) = &NONLOCAL
*_1 = derefaddrtmp(13)

the reason is we avoid creating restrict vars "recursively" but then
fail to create any initializing constraints at all, added by
r6-6208-g0b4b6ef2272f75.

I'm testing a patch.