On Thu, May 14, 2026 at 01:48:36PM -0400, Gregory Price wrote: > On Thu, May 14, 2026 at 11:47:19AM -0400, Michael S. Tsirkin wrote: > > On Thu, May 14, 2026 at 11:41:45AM -0400, Gregory Price wrote: > > > On Thu, May 14, 2026 at 11:34:14AM -0400, Michael S. Tsirkin wrote: > > > > On Thu, May 14, 2026 at 11:32:26AM -0400, Gregory Price wrote: > > > > > > > > > > > > > > Maybe mildly out of scope but worth asking: Are there other > > > > > > > flags that > > > > > > > should be retained/propogated on a split? If so, rather than pass > > > > > > > was_reported, should we just take a temporary copy of the page > > > > > > > flags and > > > > > > > pass them all in? > > > > > > > > > > > > > > ~Gregory > > > > > > > > > > > > > > > > > > Not that I can see, no. > > > > > > > > > > > > > > > > You do this exact thing again later :] > > > > > > > > > > https://lore.kernel.org/linux-mm/9a22e0f9bbe1278913754db6df76e291a006181a.1778616612.git....@redhat.com/ > > > > > > > > > > ~Gregory > > > > > > > > I mean yes, reported and zeroed) Just zeroed does not exist at this > > > > stage in the series. > > > > > > > > > > More of a forward looking question: if we already know we're adding > > > was_reported and was_zeroed, maybe we should just pass the page flags > > > through entirely and let expand() check them instead of passing it > > > through individually. > > > > > > We end up doing something similar with post_alloc_hook later as well, so > > > I'm wondering if there's a general improvement that can be made here. > > > > > > ~Gregory > > > > sure, will do. do you want a bitwise thing for type safety? or just a > > long? > > > > Can probably just do like: > > flags = stable_page_flags(page); > > plumbing... > > expand(..., buddy_flags) { > if (test_bit(PG_reported, buddy_flags)) > ... > if (test_bit(PG_zeroed, buddy_flags)) > ... > } > > or whatever makes sense. > > ~Gregory
right, that's clear.
