The library also uses some absurdly long identifiers which are bound to clutter your code one way or another. To iterate a binary tree, use

while(lfds710_btree_au_get_by_absolute_position_and_then_by_relative_position(state, &element, LFDS710_BTREE_AU_SMALLEST_IN_TREE, LFDS710_BTREE_AU_INORDER_WALK_FROM_SMALLEST_TO_LARGEST))

It's not going to be possible to remember what names to use when you need to iterate. Or, take the initialization routine. You first call

lfds710_btree_au_init_valid_on_current_logical_core( &state, key_compare_function, LFDS710_BTREE_AU_EXISTING_KEY_FAIL, NULL );

on one core, which is perhaps reasonable, but then you call

LFDS710_MISC_MAKE_VALID_ON_CURRENT_LOGICAL_CORE_INITS_COMPLETED_BEFORE_NOW_ON_ANY_OTHER_LOGICAL_CORE()

on other cores. That is one hell of a #define, and it's not even clear from the name what it's doing. It seems like "LFDS710_MISC_SYNC_INITS" or something along those lines would be clearer and easier to remember.