Comment by jcranmer

There's arguments to be made on both sides, but I think the problem with Linus's solution here is that it doesn't quite clearly establish the assumption being made, which gives it a bit too much of the 'cleverness' flavor that you allude to. A better implementation would be one that does establish why the use of pointers make sense:

   void remove_entry(node_t *entry) {
     // curr_ref is the address of the link pointing to curr
     node_t **curr_ref = &head;
     node_t *curr = *curr_ref;
     while (curr != NULL && curr != entry) {
       // Advance curr_ref and curr
       curr_ref = &curr->next;
       curr = *curr_ref;
     }
     if (curr) { *curr_ref = curr->next; }
   }

Choosing names somewhat more rationally, and making it clear that "curr" always points to the current node, and that "curr_ref" is the address of whatever pointer we followed to arrive at curr, makes it easier to establish the invariant that updating curr_ref is sufficient to insert or remove curr into a list, no matter if it's referring to the head link of a list or the next link of the prior node.