On Fri, Dec 09, 2022 at 11:24:55AM +0100, David Hildenbrand wrote: > For such cases, it would be good to have any evidence that it really helps. I don't know much on the s390 path, but if a process has a large hugetlb vma, even MADV_DONTNEED will be blocked for whatever long time if there's another process or thread scanning pagemap for this vma. Would this justify a bit? It's just that the vma lock is taken write far more than mmap_lock taken write I think, meanwhile what we need here is only release the lock and retake, nothing else. It didn't make things over complicated, IMO. No strong ipinion from my side, as I said to me it's really low hanging fruit. If you still think that doesn't justify and if Mike doesn't have a preference either I can just drop it for later. -- Peter Xu
