I explore signatures of a possible dust formation in the late SN~2010jl that could be imprinted in the line blueshift and the radius evolution of the dusty infrared-emitting shell. I propose a simple model that permits one to reproduce emission lines of blueshifted hydrogen and helium emission lines. The model suggests that the hydrogen emission originates primarily from shocked fragmented circumstellar clumps partially obscured by the absorbing cool dense shell and by unshocked ejecta. In the He I 1.083 $mu$m line on day 178 this component is significantly weaker compared to broad component from unshocked ejecta that is obscured by the absorprion produced by ejecta itself. Simulations of late time ($t > 400$ d) H$alpha$ suggest that, apart from the dust in the cool dense shell, a significant amount of dust must form in the unshocked supernova ejecta. The supernova radius predicted by the interaction model coincides with the radius of the dusty shell recovered from late time (> 460 days) infrared data, which strongly support that infrared radiation indeed originates from supernova. The ejecta dust is presumably locked in opaque blobs.