The recent paper by V. G. Kogan and J. Schmalian Phys. Rev. B 83, 054515 (2011) argues that the widely used two-component Ginzburg-Landau (GL) models are not correct, and further concludes that in the regime which is described by a GL theory there could be no disparity in the coherence lengths of two superconducting components. This would in particular imply that (in contrast to $U(1)\times U(1)$ superconductors), there could be no "type-1.5" superconducting regime in U(1) multiband systems for any finite interband coupling strength. We point out that these claims are incorrect and based on an erroneous scheme of reduction of a two-component GL theory. We also attach a separate rejoinder on reply by Kogan and Schmalian. In their reply Phys. Rev. B 86, 016502 (2012) to our comment Phys. Rev. B 86, 016501 (2012) Kogan and Schmalian did not refute or, indeed, discuss the main points of criticism in the comment. Unfortunately they instead advance new incorrect claims regarding two-band and type-1.5 superconductivity. The main purpose of the attached rejoinder is to discuss these new incorrect claims.