For convenience of the expressions below I will redefine $k\mapsto k+1$, so that $k+1$ wins of a player are required to end the game.
In terms of generating functions the probability that the game will end in
$n+1$ rounds is:
$$
P_n(m,k)=\frac{n!C_{n}(m,k)}{m^{n}}\tag1
$$
where
$$
C_n(m,k)=[x^n]\frac{x^k}{k!}\left(\sum_{i=0}^k\frac{x^i}{i!}\right)^{m-1},\tag2
$$
with $[x^n]$ being the coefficient extractor.
The expected number of rounds can then be computed as:
$$
R(m,k)=\sum_{n\ge0}(n+1)P_n(m,k)=\sum_{n=k}^{km}\frac{(n+1)!C_{n}(m,k)}{m^{n}}.\tag3
$$
For the case $m=2$ it is not hard to find the closed form of the above expressions:
$$\begin{align}
P_n(2,k)&=\binom{n}{k}\frac1{2^n},\quad k\le n\le 2k,\\
R(2,k)&=(2k+1)\left[1-\binom{2k}k\frac1{4^k}\right]+1.
\end{align}$$