It seems that there had been no formal experiments that really prove/disprove the theory that contributions and pull requests on Github are judged according to sex/race that that kind of factors. It's a bit strange considering how easy it would be to do such an experiment. Decades ago formal experiments that proved the felt superiority of white ppl paved the way for civil rights, and it would be nice if we can have some formal, scientific support today as well.
Personally, I think that discrimination is possible in some corporate Github projects. However, it is incredibly unlikely in something as hard core as Linux. Generally speaking, the more hard-core and high level you get, the less politically opinionated ppl are, so it really is hardly imaginable that Linux contributors, being devoted developers, would really discriminate, especially when many contributors are simply known by an email address, and sex/race isn't even disclosed.