From: Jan Kara Subject: Re: [PATCH v3 3/5] dax: use common 4k zero page for dax mmap reads Date: Thu, 20 Jul 2017 12:27:23 +0200 Message-ID: <20170720102723.GB17689@quack2.suse.cz> References: <20170628220152.28161-1-ross.zwisler@linux.intel.com> <20170628220152.28161-4-ross.zwisler@linux.intel.com> <20170719153314.GC15908@quack2.suse.cz> <20170719162645.GA26445@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: linux-xfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Theodore Ts'o , Matthew Wilcox , "Darrick J. Wong" , Jonathan Corbet , linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Steven Rostedt , linux-doc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, Dave Hansen , Ingo Molnar , Andreas Dilger , Alexander Viro , linux-fsdevel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Jan Kara , Andrew Morton , linux-ext4-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Christoph Hellwig , linux-nvdimm-hn68Rpc1hR1g9hUCZPvPmw@public.gmane.org To: Ross Zwisler Return-path: Content-Disposition: inline In-Reply-To: <20170719162645.GA26445-VuQAYsv1563Yd54FQh9/CA@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: linux-nvdimm-bounces-hn68Rpc1hR1g9hUCZPvPmw@public.gmane.org Sender: "Linux-nvdimm" List-Id: linux-ext4.vger.kernel.org On Wed 19-07-17 10:26:45, Ross Zwisler wrote: > On Wed, Jul 19, 2017 at 05:33:14PM +0200, Jan Kara wrote: > > On Wed 28-06-17 16:01:50, Ross Zwisler wrote: > > > Another major change is that we remove dax_pfn_mkwrite() from our fault > > > flow, and instead rely on the page fault itself to make the PTE dirty and > > > writeable. The following description from the patch adding the > > > vm_insert_mixed_mkwrite() call explains this a little more: > > > > > > *** > > > To be able to use the common 4k zero page in DAX we need to have our PTE > > > fault path look more like our PMD fault path where a PTE entry can be > > > marked as dirty and writeable as it is first inserted, rather than > > > waiting for a follow-up dax_pfn_mkwrite() => finish_mkwrite_fault() call. > > > > > > Right now we can rely on having a dax_pfn_mkwrite() call because we can > > > distinguish between these two cases in do_wp_page(): > > > > > > case 1: 4k zero page => writable DAX storage > > > case 2: read-only DAX storage => writeable DAX storage > > > > > > This distinction is made by via vm_normal_page(). vm_normal_page() > > > returns false for the common 4k zero page, though, just as it does for > > > DAX ptes. Instead of special casing the DAX + 4k zero page case, we will > > > simplify our DAX PTE page fault sequence so that it matches our DAX PMD > > > sequence, and get rid of dax_pfn_mkwrite() completely. > > > > > > This means that insert_pfn() needs to follow the lead of insert_pfn_pmd() > > > and allow us to pass in a 'mkwrite' flag. If 'mkwrite' is set > > > insert_pfn() will do the work that was previously done by wp_page_reuse() > > > as part of the dax_pfn_mkwrite() call path. > > > *** > > > > Hum, thinking about this in context of this patch... So what if we have > > allocated storage, a process faults it read-only, we map it to page tables > > writeprotected. Then the process writes through mmap to the area - the code > > in handle_pte_fault() ends up in do_wp_page() if I'm reading it right. > > Yep. > > > Then, since we are missing ->pfn_mkwrite() handlers, the PTE will be marked > > writeable but radix tree entry stays clean - bug. Am I missing something? > > I don't think we ever end up with a writeable PTE but with a clean radix tree > entry. When we get the write fault we do a full fault through > dax_iomap_pte_fault() and dax_insert_mapping(). > > dax_insert_mapping() sets up the dirty radix tree entry via > dax_insert_mapping_entry() before it does anything with the page tables via > vm_insert_mixed_mkwrite(). > > So, this mkwrite fault path is exactly the path we would have taken if the > initial read to real storage hadn't happened, and we end up in the same end > state - with a dirty DAX radix tree entry and a writeable PTE. Ah sorry, I have missed that it is not that you would not have ->pfn_mkwrite() handler - you still have it but it is the same as standard fault handler now. So maybe can you rephrase the changelog a bit saying that: "We get rid of dax_pfn_mkwrite() helper and use dax_iomap_fault() to handle write-protection faults instead." Thanks! Honza -- Jan Kara SUSE Labs, CR