Make it as simple as possible, not simpler

The real "simple" solution is going back to single allocations
The individual allocation ensures better data locality
Which in turn means less cache miss events, and more performance

But we need a basically different approach...
The approach taken in the kernel is reversing the structure
Instead of including the payload in the list structure
                the list structure is included in the payload itself

#include <linux/list.h>
struct list_head {
        struct list_head *next, *prev;
};

static inline void list_add(struct list_head *new, struct list_head *head);
static inline void list_add_tail(struct list_head *new, struct list_head *head);
static inline void list_del(struct list_head *entry);
/* .... */