From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754105AbcJZJyS (ORCPT ); Wed, 26 Oct 2016 05:54:18 -0400 Received: from mail-wm0-f67.google.com ([74.125.82.67]:35242 "EHLO mail-wm0-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750720AbcJZJyQ (ORCPT ); Wed, 26 Oct 2016 05:54:16 -0400 Date: Wed, 26 Oct 2016 11:54:13 +0200 From: Michal Hocko To: Lorenzo Stoakes Cc: linux-mm@kvack.org, Linus Torvalds , Jan Kara , Hugh Dickins , Dave Hansen , Rik van Riel , Mel Gorman , Andrew Morton , Paolo Bonzini , Radim =?utf-8?B?S3LEjW3DocWZ?= , kvm@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH] mm: remove unnecessary __get_user_pages_unlocked() calls Message-ID: <20161026095413.GF18382@dhcp22.suse.cz> References: <20161025233609.5601-1-lstoakes@gmail.com> <20161025234631.GA5946@lucifer> <20161026091542.GD18382@dhcp22.suse.cz> <20161026093913.GA12814@lucifer> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20161026093913.GA12814@lucifer> User-Agent: Mutt/1.6.0 (2016-04-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed 26-10-16 10:39:13, Lorenzo Stoakes wrote: > On Wed, Oct 26, 2016 at 11:15:43AM +0200, Michal Hocko wrote: > > On Wed 26-10-16 00:46:31, Lorenzo Stoakes wrote: > > > The holdout for unexporting __get_user_pages_unlocked() is its invocation in > > > mm/process_vm_access.c: process_vm_rw_single_vec(), as this definitely _does_ > > > seem to invoke VM_FAULT_RETRY behaviour which get_user_pages_remote() will not > > > trigger if we were to replace it with the latter. > > > > I am not sure I understand. Prior to 1e9877902dc7e this used > > get_user_pages_unlocked. What prevents us from reintroducing it with > > FOLL_REMOVE which was meant to be added by the above commit? > > > > Or am I missing your point? > > The issue isn't the flags being passed, rather that in this case: > > a. Replacing __get_user_pages_unlocked() with get_user_pages_unlocked() won't > work as the latter assumes task = current and mm = current->mm but > process_vm_rw_single_vec() needs to pass different task, mm. Ohh, right. I should have checked more closely. > b. Moving to get_user_pages_remote() _will_ allow us to pass different task, mm > but won't however match existing behaviour precisely, since > __get_user_pages_unlocked() acquires mmap_sem then passes a pointer to a > local 'locked' variable to __get_user_pages_locked() which allows > VM_FAULT_RETRY to trigger. I do not see any reason why get_user_pages_remote should implicitely disallow VM_FAULT_RETRY. Releasing the mmap_sem on a remote task when we have to wait for IO is a good thing in general. So I would rather see a way to do allow that. Doing that implicitly sounds too dangerous and maybe we even have users which wouldn't cope with the mmap sem being dropped (get_arg_page sounds like a potential example) so I would rather add locked * parameter to get_user_pages_remote. -- Michal Hocko SUSE Labs