As far as I have understood the problem, the subnormal, is equal to the ordinate of the point of tangency. If this is the case, then,
y(dy/dx) = y
i.e, dy/dx=1
this is not a linear differential equation; neither a second order, nor homogeneous in 1st degree. It is of course in the variable seperable form. So, I think c is correct.