It's funny because it reflects so well stuff from an essay I'm reading. The author presents "genre literature" as an artificial gimmick for the purpose of marketing, designed in publishing circles (mostly anglo-saxon originally) when they started to be faced with the multiplication of books in populat literature and the challenge to get their products the customers' attention, which they answered by artificially circumscribing the offer to orient readers toward certain books that can repeat the reading experience given by their previous purchases (it's much easier to determine what should be published and not if you manage to confine part of the readers to various niches than if you let them pick and choose freely what they'll read). Another similar trick, originally more common in French publishing, are publishers' Collections). Instead of putting a genre label on a novel, they publish the crime novels in the Série Noire, Masque, SF books in Ailleurs et Demain and so on, all with their distintive cover design. Putting a genre label on a novel also increases the chances the novel will be picked by bookstores (that each day offer a smaller percentage of all the books in print). That gradually replaced the literary trends/movements or forms (like the original novel), and originally if the genres tended to be well-defined, with certain rules/codes developping and writers (and readers) even forming circles around genres, but as the offering in each genre became larger and larger and the books again needed to get attention, the publishers repeated the recipe by inventing sub-genres, and sub-sub-genres, to the point of absurdity (we see much of those practical limitations of genre labelling on a forum like this one. Whenever someone asks "I like X, suggest more authors like this" there's never much consensus. To some, Jordan and Lynch or Hobb/Erikson/Martin/ are similar because they like both. To others, not at all because they like one and dislike the other - and in the end it's got barely anything to do with the so-called genre they have in common). The same marketing logic lead to a semantic shift concerning "classics", a label that has more to do nowadays with the fact a book sold well at one point than with long-term, enduring literary recognition, let alone with being written during classical times (you get LOTR called a classic, for instance, while opinion on its literary value or importance is very divided).

"Mainstream" he sees as merely another marketing ploy, and primarly the modern name of popular literature. It defines literature that has managed to get enough mediatic attention (for whatever reason, occasionally even for the literary quality) and became a best-seller, and to adress the fact genre literature sometime but too rarely became mainstream, the publishers have invented genre best-selling lists, and encouraged the emergence of "genre reviews" once reviews start to shift from being literary discussions/commentary usually written by and meant for elites to become purchasing advice for a mass audience). As "genres" came to dominance in marketing books, more and more writers have started, purposefully and not, to mix them up, further eroding the value of genre labels.

In the end, any mainstream novel can be fitted loosely into a genre (or several genres at once) if you try hard enough, but usually they're books the readers have chosen en masse regardless of genre (really few care to know if Dan Brown writes thrillers or if Harry Potter is or isn't Fantasy, let alone which sub-sub-genre of Fantasy it might belong to). The association to genres is downplayed once a book/writer has transcended niche readership and has wide exposure, and insisting too much on a label at this point might limit rather than widen your sales. Or you start cheating, saying stuff like Stephen King writes mainstream horror novels, or the Name of the Rose is a literary thriller/historical novel so the fact it de facto became popular literarure by its success (some of which is explained more by the mediatized fame of Eco as a scholar than the real mass appeal of his novels) doesn't cast a shadow on its intellectual/literary qualities.