On Thu, Nov 05, 2020 at 10:36:49AM -0500, Johannes Weiner wrote: > But the code is highly specific - synchronizing one struct page member > for one particular use case. Let's keep at least a reference to what > we are synchronizing against. There is a non-zero chance that if the > comment goes out of date, so does the code. How about this? > > /* > * page_idle does a lockless/optimistic rmap scan on page->mapping. > * Make sure the compiler doesn't split the stores of anon_vma and > * the PAGE_MAPPING_ANON type identifier, otherwise the rmap code > * could mistake the mapping for a struct address_space and crash. > */ Fine by me! There may be other cases where seeing a split store would be bad, so I didn't want to call out page_idle explicitly. But if you want to, I'm happy with this comment.
