Channel load steering never triggered

[matrach]

It seems that the is_better_candidate function never returns UEV_SELECT_REASON_LOAD, because the condition is unsatisfiable. It is essentially p ∧ ¬p, unless some pointer aliasing happens in-between.

Probably the parameters were supposed to be flipped in the second call.

usteer/policy.c

Lines 110 to 112 in e218150

	if (has_better_load(current_node, new_node) &&
	!has_better_load(current_node, new_node))
	reasons |= (1 << UEV_SELECT_REASON_LOAD);

[AlD]

Just came across the same thing. Another issue seems to be that below_load_threshold actually implements above_load_threshold.