Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757977Ab1FGTBi (ORCPT ); Tue, 7 Jun 2011 15:01:38 -0400 Received: from mga02.intel.com ([134.134.136.20]:17219 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755868Ab1FGTBh (ORCPT ); Tue, 7 Jun 2011 15:01:37 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.65,333,1304319600"; d="scan'208";a="10318789" Message-ID: <4DEE758F.8060002@linux.intel.com> Date: Tue, 07 Jun 2011 12:01:35 -0700 From: Darren Hart User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.17) Gecko/20110424 Lightning/1.0b2 Thunderbird/3.1.10 MIME-Version: 1.0 To: Andrew Lutomirski CC: Eric Dumazet , Peter Zijlstra , David Oliver , linux-kernel@vger.kernel.org, Shawn Bohrer , Zachary Vonler , KOSAKI Motohiro , Hugh Dickins , Thomas Gleixner , Ingo Molnar Subject: Re: Change in functionality of futex() system call. References: <1307373819.3098.40.camel@edumazet-laptop> <1307376672.2322.167.camel@twins> <1307376989.2322.171.camel@twins> <1307377349.3098.65.camel@edumazet-laptop> <1307377782.2322.183.camel@twins> <1307378564.3098.67.camel@edumazet-laptop> <4DED1421.5000300@linux.intel.com> <1307383898.3098.90.camel@edumazet-laptop> <4DED976C.90009@linux.intel.com> <4DEE3944.5020005@mit.edu> <1307462300.3091.39.camel@edumazet-laptop> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3168 Lines: 85 On 06/07/2011 11:43 AM, Andrew Lutomirski wrote: > On Tue, Jun 7, 2011 at 11:58 AM, Eric Dumazet wrote: >> Le mardi 07 juin 2011 ? 10:44 -0400, Andy Lutomirski a ?crit : >>> On 06/06/2011 11:13 PM, Darren Hart wrote: >>>> >>>> >>>> On 06/06/2011 11:11 AM, Eric Dumazet wrote: >>>>> Le lundi 06 juin 2011 ? 10:53 -0700, Darren Hart a ?crit : >>>>>> >>>>> >>>>>> If I understand the problem correctly, RO private mapping really doesn't >>>>>> make any sense and we should probably explicitly not support it, while >>>>>> special casing the RO shared mapping in support of David's scenario. >>>>>> >>>>> >>>>> We supported them in 2.6.18 kernels, apparently. This might sounds >>>>> stupid but who knows ? >>>> >>>> >>>> I guess this is actually the key point we need to agree on to provide a >>>> solution. This particular case "worked" in 2.6.18 kernels, but that >>>> doesn't necessarily mean it was supported, or even intentional. >>>> >>>> It sounds to me that we agree that we should support RO shared mappings. >>>> The question remains about whether we should introduce deliberate >>>> support of RO private mappings, and if so, if the forced COW approach is >>>> appropriate or not. >>>> >>> >>> I disagree. >>> >>> FUTEX_WAIT has side-effects. Specifically, it eats one wakeup sent by >>> FUTEX_WAKE. So if something uses futexes on a file mapping, then a >>> process with only read access could (if the semantics were changed) DoS >>> the other processes by spawning a bunch of threads and FUTEX_WAITing >>> from each of them. >>> >>> If there were a FUTEX_WAIT_NOCONSUME that did not consume a wakeup and >>> worked on RO mappings, I would drop my objection. >> >> If a group of cooperating processes uses a memory segment to exchange >> critical information, do you really think this memory segment will be >> readable by other unrelated processes on the machine ? > > Depends on the design. > > I have some software I'm working on that uses shared files and could > easily use futexes. I don't want random read-only processes to > interfere with the futex protocol. So don't use world readable files. >> >> How is this related to futex code ? > > Because this usage is currently safe and would become unsafe with the > proposed change. > >> >> Same problem for legacy IPC (shm, msg, sem) : Appropriate protections >> are needed, obviously. >> >> BTW, kernel/futex.c uses a global hash table (futex_queues[256]) and a >> very predictable hash_futex(), so its easy to slow down futex users... > > There's a difference between slowing down users by abusing a kernel > hash and deadlocking users by eating a wakeup. (If you eat a wakeup > the wakeup won't magically come back later. It's gone.) That's the nature of SHARED, you have to protect the mapping independent of the futex mechanism. -- Darren Hart Intel Open Source Technology Center Yocto Project - Linux Kernel -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/