>> Sounds good? At least to me. If only swap/migration entries wouldn't be >> harder to handle than I'd wish, that's why it's taking a little and will >> take a little longer. > > Thanks for the quick response. I would have to see the logic to set/clear > PageAnonExclusive to fully understand how things are handled. > I'll let you know as soon as I have something. I'll most probably share a document explaining the design first. > BTW, I just saw this patch form PeterZ [1] that seems to be related, as > it deals with changing protection on pinned pages. Unfortunately the use of page_maybe_dma_pinned() is racy, as we can race with GUP fast. It's a problem for vmscan which currently relies on it for correctness. For migration, it might be good enough as we should fail later when trying to freeze the refcount. > > > [1] https://lore.kernel.org/linux-mm/20220120160822.666778608@xxxxxxxxxxxxx/ > -- Thanks, David / dhildenb
