Let G be a global variable, and let H be a variable allocated on the heap. Both G and H are structs with many fields, and a programmer wishes to protect each with a sequence lock. Why is it necessary...

Let G be a global variable, and let H be a variable allocated on the heap. Both G and H are structs with many fields, and a programmer wishes to protect each with a sequence lock. Why is it necessary to use a safe memory reclamation strategy for H, but not for G?