Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: > Side note: I think we should just do the "FOLL_GET" doesn't touch the > refcount either, which would make this all become just Yeah... but there's a lot of places that would potentially need fixing. iov_iter_get_pages*(), for example, is used to grab pages and push them to all sorts of destinations, including pipes and sk_buffs. > but then we would need to fix try_grab_page() and gup_put_folio() and > friends to match. And any other cases I haven't thought of. And put_page(), folio_put(), ... I wonder if there would be a noticeable performance loss from adding an is_zero_page() check into those. It might also make sense to include the entire kernel post-init static image in that, perhaps a check: if ((unsigned long)(virt_address) & KERNEL_MASK == KERNEL_BASE) return; David
