The problem this poses is that the axes of the holes are not dimensioned to their intersections with datum A and they are not given a basic angle to datum A. As-is the slope of the holes eats into the allowable, but unknown angular tolerance on the control of the axis. For example, if the slope is 30 degrees, and the material is .1 inch thick, the axis displacement to A|B|C is sin(30)*.1 = .05 inches, for a positional diameter of 0.100, which is a bit larger than the tolerance.
What would be better is to establish a small taget area around each large hole as datum E and locate the countersunk holes to [E|D|B] INDIVIDUALLY. The primary thru holes should have the axes dimensioned to their intersections with A and given a basic angle to A. This duplicates how the drill will be positioned to create the holes.
This is necessary because the tolearance is on the entire axis of the hole, not the intersection of the hole axis with the 'obvious' side. This intersection will float depending on the angle of the bend.