From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Gunthorpe Subject: Re: [PATCH 0/4] get_user_pages*() and RDMA: first steps Date: Mon, 1 Oct 2018 09:31:01 -0600 Message-ID: <20181001153101.GA23286@ziepe.ca> References: <20180928053949.5381-1-jhubbard@nvidia.com> <20180928152958.GA3321@redhat.com> <4c884529-e2ff-3808-9763-eb0e71f5a616@nvidia.com> <20180928214934.GA3265@redhat.com> <20180929084608.GA3188@redhat.com> <20181001061127.GQ31060@dastard> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20181001061127.GQ31060@dastard> Sender: linux-kernel-owner@vger.kernel.org To: Dave Chinner Cc: Jerome Glisse , John Hubbard , john.hubbard@gmail.com, Matthew Wilcox , Michal Hocko , Christopher Lameter , Dan Williams , Jan Kara , Al Viro , linux-mm@kvack.org, LKML , linux-rdma , linux-fsdevel@vger.kernel.org, Christian Benvenuti , Dennis Dalessandro , Doug Ledford , Mike Marciniszyn List-Id: linux-rdma@vger.kernel.org On Mon, Oct 01, 2018 at 04:11:27PM +1000, Dave Chinner wrote: > This reminds me so much of Linux mmap() in the mid-2000s - mmap() > worked for ext3 without being aware of page faults, so most mm/ > developers at the time were of the opinion that all the other > filesystems should work just fine without being aware of page > faults. This is probably because RDMA was introduced around that time, before page_mkwrite was added.. It kind of sounds like page_mkwrite() broke all the writing get_user_pages users and it wasn't realized at the time. BTW, we are hosting a session on this at Plumbers during the RDMA track, hope everyone interested will be able to attend. Jason