Let $\mathsf{A}$ be an abelian category. I understand that if the homotopy category $\mathsf{K}(\mathsf{A})$ is abelian, then $\mathsf{K}(\mathsf{A})$ is semi-simple. (There's a proof of this in another question of mine: 1.)
Now, I think that this should imply that $\mathsf{A}$ itself is semi-simple. (The analogous fact for $\mathsf{D}(\mathsf{A})$ is true. 2) But I can't seem to prove it.
Let $0\to A \to B\to C\to 0$ be an exact sequence in $\mathsf{A}$. If it is also exact when seen in $\mathsf{K}(\mathsf{A})$, then $A\to B$ is a split monomorphism in $\mathsf{K}(\mathsf{A})$ and, since $\mathsf{A}$ embeds fully faithfully in $\mathsf{K}(\mathsf{A})$, it is also split in $\mathsf{A}$, finishing the proof. However it is not clear to me why the sequence is also exact in $\mathsf{K}(\mathsf{A})$ or even why $A\to B$ is still a monomorphism in $\mathsf{K}(\mathsf{A})$.