From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752609AbcEKLHm (ORCPT ); Wed, 11 May 2016 07:07:42 -0400 Received: from merlin.infradead.org ([205.233.59.134]:59594 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751565AbcEKLHk (ORCPT ); Wed, 11 May 2016 07:07:40 -0400 Subject: Re: [PATCH 1/2] mmap.2: clarify MAP_LOCKED semantic To: Michal Hocko , Michael Kerrisk References: <1431527892-2996-1-git-send-email-miso@dhcp22.suse.cz> <1431527892-2996-2-git-send-email-miso@dhcp22.suse.cz> Cc: Andrew Morton , Linus Torvalds , David Rientjes , LKML , Linux API , linux-mm@kvack.org, Michal Hocko From: Peter Zijlstra Message-ID: <57331275.9000805@infradead.org> Date: Wed, 11 May 2016 13:07:33 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.7.2 MIME-Version: 1.0 In-Reply-To: <1431527892-2996-2-git-send-email-miso@dhcp22.suse.cz> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 05/13/2015 04:38 PM, Michal Hocko wrote: > From: Michal Hocko > > MAP_LOCKED had a subtly different semantic from mmap(2)+mlock(2) since > it has been introduced. > mlock(2) fails if the memory range cannot get populated to guarantee > that no future major faults will happen on the range. mmap(MAP_LOCKED) on > the other hand silently succeeds even if the range was populated only > partially. > > Fixing this subtle difference in the kernel is rather awkward because > the memory population happens after mm locks have been dropped and so > the cleanup before returning failure (munlock) could operate on something > else than the originally mapped area. > > E.g. speculative userspace page fault handler catching SEGV and doing > mmap(fault_addr, MAP_FIXED|MAP_LOCKED) might discard portion of a racing > mmap and lead to lost data. Although it is not clear whether such a > usage would be valid, mmap page doesn't explicitly describe requirements > for threaded applications so we cannot exclude this possibility. > > This patch makes the semantic of MAP_LOCKED explicit and suggest using > mmap + mlock as the only way to guarantee no later major page faults. > URGH, this really blows chunks. It basically means MAP_LOCKED is pointless cruft and we might as well remove it. Why not fix it proper? From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Zijlstra Subject: Re: [PATCH 1/2] mmap.2: clarify MAP_LOCKED semantic Date: Wed, 11 May 2016 13:07:33 +0200 Message-ID: <57331275.9000805@infradead.org> References: <1431527892-2996-1-git-send-email-miso@dhcp22.suse.cz> <1431527892-2996-2-git-send-email-miso@dhcp22.suse.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1431527892-2996-2-git-send-email-miso-2MMpYkNvuYDjFM9bn6wA6Q@public.gmane.org> Sender: linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Michal Hocko , Michael Kerrisk Cc: Andrew Morton , Linus Torvalds , David Rientjes , LKML , Linux API , linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, Michal Hocko List-Id: linux-api@vger.kernel.org On 05/13/2015 04:38 PM, Michal Hocko wrote: > From: Michal Hocko > > MAP_LOCKED had a subtly different semantic from mmap(2)+mlock(2) since > it has been introduced. > mlock(2) fails if the memory range cannot get populated to guarantee > that no future major faults will happen on the range. mmap(MAP_LOCKED) on > the other hand silently succeeds even if the range was populated only > partially. > > Fixing this subtle difference in the kernel is rather awkward because > the memory population happens after mm locks have been dropped and so > the cleanup before returning failure (munlock) could operate on something > else than the originally mapped area. > > E.g. speculative userspace page fault handler catching SEGV and doing > mmap(fault_addr, MAP_FIXED|MAP_LOCKED) might discard portion of a racing > mmap and lead to lost data. Although it is not clear whether such a > usage would be valid, mmap page doesn't explicitly describe requirements > for threaded applications so we cannot exclude this possibility. > > This patch makes the semantic of MAP_LOCKED explicit and suggest using > mmap + mlock as the only way to guarantee no later major page faults. > URGH, this really blows chunks. It basically means MAP_LOCKED is pointless cruft and we might as well remove it. Why not fix it proper? From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-ob0-f198.google.com (mail-ob0-f198.google.com [209.85.214.198]) by kanga.kvack.org (Postfix) with ESMTP id BBA146B0005 for ; Wed, 11 May 2016 07:07:39 -0400 (EDT) Received: by mail-ob0-f198.google.com with SMTP id rd14so5272388obb.3 for ; Wed, 11 May 2016 04:07:39 -0700 (PDT) Received: from merlin.infradead.org (merlin.infradead.org. [2001:4978:20e::2]) by mx.google.com with ESMTPS id m20si4213618ita.72.2016.05.11.04.07.39 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Wed, 11 May 2016 04:07:39 -0700 (PDT) Subject: Re: [PATCH 1/2] mmap.2: clarify MAP_LOCKED semantic References: <1431527892-2996-1-git-send-email-miso@dhcp22.suse.cz> <1431527892-2996-2-git-send-email-miso@dhcp22.suse.cz> From: Peter Zijlstra Message-ID: <57331275.9000805@infradead.org> Date: Wed, 11 May 2016 13:07:33 +0200 MIME-Version: 1.0 In-Reply-To: <1431527892-2996-2-git-send-email-miso@dhcp22.suse.cz> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: Michal Hocko , Michael Kerrisk Cc: Andrew Morton , Linus Torvalds , David Rientjes , LKML , Linux API , linux-mm@kvack.org, Michal Hocko On 05/13/2015 04:38 PM, Michal Hocko wrote: > From: Michal Hocko > > MAP_LOCKED had a subtly different semantic from mmap(2)+mlock(2) since > it has been introduced. > mlock(2) fails if the memory range cannot get populated to guarantee > that no future major faults will happen on the range. mmap(MAP_LOCKED) on > the other hand silently succeeds even if the range was populated only > partially. > > Fixing this subtle difference in the kernel is rather awkward because > the memory population happens after mm locks have been dropped and so > the cleanup before returning failure (munlock) could operate on something > else than the originally mapped area. > > E.g. speculative userspace page fault handler catching SEGV and doing > mmap(fault_addr, MAP_FIXED|MAP_LOCKED) might discard portion of a racing > mmap and lead to lost data. Although it is not clear whether such a > usage would be valid, mmap page doesn't explicitly describe requirements > for threaded applications so we cannot exclude this possibility. > > This patch makes the semantic of MAP_LOCKED explicit and suggest using > mmap + mlock as the only way to guarantee no later major page faults. > URGH, this really blows chunks. It basically means MAP_LOCKED is pointless cruft and we might as well remove it. Why not fix it proper? -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org