Re: [PATCH v3 0/1] mm: introduce put_user_page*(), placeholder versions

From: Jason Gunthorpe
Date: Tue Mar 12 2019 - 11:53:14 EST


On Tue, Mar 12, 2019 at 11:35:29AM -0400, Jerome Glisse wrote:
> > > > Yes you now have the filesystem as well as the GUP pinner claiming
> > > > authority over the contents of a single memory segment. Maybe better not
> > > > allow that?
> > >
> > > This goes back to regressing existing driver with existing users.
> >
> > There is no regression if that behavior never really worked.
>
> Well RDMA driver maintainer seems to report that this has been a valid
> and working workload for their users.

I think it is more O_DIRECT that is the history here..

In RDMA land long term GUPs of file backed pages tend to crash the
kernel (what John is trying to fix here) so I'm not sure there are
actual real & tested users, only people that wish they could do this..

Jason