https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102579
--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Well, the issue is simply that we don't remove the volatile load and the
call to 'new' is thought to possibly be a definition of 'd' via
stmt_may_clobber_ref_p_1 which is because there's a volatile_p check
before the may_be_aliased check:
/* A call that is not without side-effects might involve volatile
accesses and thus conflicts with all other volatile accesses. */
if (ref->volatile_p)
return true;
/* If the reference is based on a decl that is not aliased the call
cannot possibly clobber it. */
if (DECL_P (base)
&& !may_be_aliased (base)
/* But local non-readonly statics can be modified through recursion
or the call may implement a threading barrier which we must
treat as may-def. */
&& (TREE_READONLY (base)
|| !is_global_var (base)))
return false;
so it indeed seems we're overly cautionous about this here and we could
relax the volatile checks.
