I would like a reference for the following result (you can assume more regularity and replace $C^2(\bar\Omega)$ with $C^2(\mathbb R^n)$ if needed):
Let $\Omega\subset\mathbb R^n$ be a bounded domain with a $C^2$ boundary. Let $f\in C^2(\bar\Omega)$ and $\gamma\in C^2(\bar\Omega)$ with $\gamma>0$. Then there is a unique strong solution $u\in C^2(\bar\Omega)$ to the PDE $\operatorname{div}(\gamma\nabla u)=0$ with the boundary condition $u=f$ at $\partial\Omega$. (If possible, I would like the reference to tell also that $u$ is the unique minimizer of $\int_\Omega\gamma|\nabla u|^2$ in $C^2(\bar\Omega)$ with the boundary condition.)
I am perfectly aware that a proof would make heavy use of modern PDE techniques with Sobolev spaces and whatnot. I would like to be able to offer PDE related topics for bachelor's theses. Proving the above statement or working with Sobolev functions would be too much and I would like to focus the theses on other issues. But for things to make any sense, I do want to give a theorem that states that solutions (in the strong sense) exist uniquely; the theorem will unfortunately be a black box for the students.
I would be very surprised if this would not be in Gilbarg/Trudinger "Elliptic Partial Differential Equations of Second Order" - I would start looking in Chapter 6 "Classical Solutions" (in fact they distinguish between "classical solutions" that are $C^2$ and "strong solutions" that are $W^{2,p}$).
An example added by another user: The following result is a special case of theorem 6.14 in section 6.3 of the book (the full theorem treats the Poisson equation):