On Thu, 22 Jan 2015 15:12:15 +0000 "Wilcox, Matthew R" <matthew.r.wilcox@xxxxxxxxx> wrote: > Looks like mips *declares* copy_user_page(), but never *defines* an implementation. > > It's documented in Documentation/cachetlb.txt, but it's not (currently) called if the architecture defines its own copy_user_highpage(), so some bitrot has occurred. ARM is currently fixing this, and MIPS will need to do the same. > > (We can't use copy_user_highpage() in DAX because we don't necessarily have a struct page for 'from'.) Has there been any progress on this? It would be unpleasant to merge DAX into 3.19 and break MIPS and ARM.
