We already have mutations, generally in women, for tetrachromaticism, who usually have male relatives with severe or moderate color blindness, in which the X chromosome encodes a different green cone. So they end up seeing red, strange-green, green, and blue, where strange-green is somewhere closer to red than green.

Only a few on record but they tend to have absolutely insane color matching and color perception. One of note worked in the fashion industry and could match fabrics perfectly even in varying lighting (e.g. working under fluorescent but able to match colors that would stay matched in halogen/stage lighting)

There's some evidence that tetrachromacy already exists in a few humans. If so we have the gene already. But why would it spread?

Most colorblind people aren't dichromats, they're so-called anomalous trichromats. Basically, the genes coding opsins in your eyes have a number of functional sites that tune the spectral sensitivity. Those sites are tuned as far apart as they can be in color-normal humans. Anomalous trichromats usually have a genetic error that causes their opsins' sensitivity curves to overlap more, which manifests as reduced color sensitivity.