Jiri Kosina wrote on Wed, Jan 16, 2019:

So if noone sees any principal problem there, I'll happily submit it with 
proper attribution etc.

I'm not convinced just the write permission check is enough for
mincore(), as Josh also seems to share the concern I raised (e.g. map a
git directory "hot" pages)
We probably need to add an inode_owner_or_capable() or similar, the open
question is do we still need the write access check after that - I don't
really know how expensive these calls are.

Thanks,
-- 
Dominique