I understand that you are seeking a solution with generating functions, but I find the approach to these Fibonacci-type problems to be much more direct with a generalized Binet formula.
There have been many extensions of the sequence with adjustable (integer) coefficients and different (integer) initial conditions, e.g., $f_n=af_{n-1}+bf_{n-2}$. (You can look up Pell, Jacobsthal, Lucas, Pell-Lucas, and Jacobsthal-Lucas sequences.) Maynard has extended the analysis to $a,b\in\mathbb{R}$, (Ref: Maynard, P. (2008), “Generalised Binet Formulae,” $Applied \ Probability \ Trust$.
We have extended Maynard's analysis to include arbitrary $f_0,f_1\in\mathbb{R}$. It is relatively straightforward to show, by induction, that
$$f_n=\left(f_1-\frac{af_0}{2}\right) \frac{\alpha^n-\beta^n}{\alpha-\beta}+\frac{af_0}{2}\frac{\alpha^n+\beta^n}{\alpha+\beta} $$
where $\alpha,\beta=(a\pm\sqrt{a^2+4b})/2$ are the characteristic roots.
With the roots that you have already correctly identified, you can now create a meaningful solution in terms of the known initial conditions. Normally, when people us the charactristic equation they would express the solution as $f_n=A\alpha^n+B\beta^n$ and calculate $A,B$ from the initial conditions.
By the way, $\delta=1+\sqrt{2}$ is known as the silver ratio and $\delta^*=1-\sqrt{2}$ as its conjugate.