> If one wants to look up something in symbol table during late optimization
> one is supposed to check that ::get does not return NULL and be conservative
> otheriwse. So your change for PR120156 was OK with me, but I did not
> explicitly write (sorty for that). Similarly for vectorizer I would simply
> test for node to exist.
That's a bit ugly since we test decl_in_symtab_p just before, but of course
very safe for release branches (the crash for the Ada testcase is a regression
present in GCC 12 and later).
Tested on x86-64/Linux, OK for all branches (once they reopen)?
* tree-vect-data-refs.cc (vect_can_force_dr_alignment_p): Return
false if the variable has no symtab node.
--
Eric Botcazou
diff --git a/gcc/tree-vect-data-refs.cc b/gcc/tree-vect-data-refs.cc
index 9fd1ef29650..6e35f549a8c 100644
--- a/gcc/tree-vect-data-refs.cc
+++ b/gcc/tree-vect-data-refs.cc
@@ -7242,7 +7242,8 @@ vect_can_force_dr_alignment_p (const_tree decl, poly_uint64 alignment)
return false;
if (decl_in_symtab_p (decl)
- && !symtab_node::get (decl)->can_increase_alignment_p ())
+ && (!symtab_node::get (decl)
+ || !symtab_node::get (decl)->can_increase_alignment_p ()))
return false;
if (TREE_STATIC (decl))
