On Tue 11-09-18 16:06:23, Zaslonko Mikhail wrote: [...] > > Well, I am afraid this is not the proper solution. We are relying on the > > full pageblock worth of initialized struct pages at many other place. We > > used to do that in the past because we have initialized the full > > section but this has been changed recently. Pavel, do you have any ideas > > how to deal with this partial mem sections now? > > I think this is not related to the recent changes of memory initialization. > If > you mean deferred init case, the problem exists even without > CONFIG_DEFERRED_STRUCT_PAGE_INIT kernel option. This is more about struct page initialization (which doesn't clear whole) memmap area and as such it stays unitialized. So you are right this is a much older issue we just happened to not notice without explicit memmap poisoning. -- Michal Hocko SUSE Labs
