From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-2.6 required=3.0 tests=DKIMWL_WL_HIGH,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, USER_AGENT_SANE_1 autolearn=no autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 662EAC41514 for ; Fri, 30 Aug 2019 06:23:45 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 2DE0C22CE3 for ; Fri, 30 Aug 2019 06:23:45 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1567146225; bh=eFs4UvYptjetSJttVh5WoztDFnBnEgLU+urvveUp3qk=; h=Date:From:To:Cc:Subject:References:In-Reply-To:List-ID:From; b=mBLUf/uYXsr6owZ4M+fmnDutVozwIJyz5PF89W3CuMa12dju9hPkKaXKszvjy4y5s FFI6v6ro4wOwOjN+ZjvtLwJAer1MaK6udRel+7TznHqXUPWKIqy7vF7P7JHLIc3Jrf FuIidlW17pbuB9m2QWwq/Y/awWYBInQlIkdjHnIU= Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727788AbfH3GXo (ORCPT ); Fri, 30 Aug 2019 02:23:44 -0400 Received: from mx2.suse.de ([195.135.220.15]:38678 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1726791AbfH3GXn (ORCPT ); Fri, 30 Aug 2019 02:23:43 -0400 X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx1.suse.de (Postfix) with ESMTP id B9049B643; Fri, 30 Aug 2019 06:23:41 +0000 (UTC) Date: Fri, 30 Aug 2019 08:23:40 +0200 From: Michal Hocko To: Yang Shi Cc: "Kirill A. Shutemov" , "Kirill A. Shutemov" , Yang Shi , Vlastimil Babka , Johannes Weiner , David Rientjes , Andrew Morton , Linux MM , Linux Kernel Mailing List Subject: Re: [v2 PATCH -mm] mm: account deferred split THPs into MemAvailable Message-ID: <20190830062340.GQ28313@dhcp22.suse.cz> References: <20190827120923.GB7538@dhcp22.suse.cz> <20190827121739.bzbxjloq7bhmroeq@box> <20190827125911.boya23eowxhqmopa@box> <20190828075708.GF7386@dhcp22.suse.cz> <20190828140329.qpcrfzg2hmkccnoq@box> <20190828141253.GM28313@dhcp22.suse.cz> <20190828144658.ar4fajfuffn6k2ki@black.fi.intel.com> <20190828160224.GP28313@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.10.1 (2018-07-13) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu 29-08-19 10:03:21, Yang Shi wrote: > On Wed, Aug 28, 2019 at 9:02 AM Michal Hocko wrote: > > > > On Wed 28-08-19 17:46:59, Kirill A. Shutemov wrote: > > > On Wed, Aug 28, 2019 at 02:12:53PM +0000, Michal Hocko wrote: > > > > On Wed 28-08-19 17:03:29, Kirill A. Shutemov wrote: > > > > > On Wed, Aug 28, 2019 at 09:57:08AM +0200, Michal Hocko wrote: > > > > > > On Tue 27-08-19 10:06:20, Yang Shi wrote: > > > > > > > > > > > > > > > > > > > > > On 8/27/19 5:59 AM, Kirill A. Shutemov wrote: > > > > > > > > On Tue, Aug 27, 2019 at 03:17:39PM +0300, Kirill A. Shutemov wrote: > > > > > > > > > On Tue, Aug 27, 2019 at 02:09:23PM +0200, Michal Hocko wrote: > > > > > > > > > > On Tue 27-08-19 14:01:56, Vlastimil Babka wrote: > > > > > > > > > > > On 8/27/19 1:02 PM, Kirill A. Shutemov wrote: > > > > > > > > > > > > On Tue, Aug 27, 2019 at 08:01:39AM +0200, Michal Hocko wrote: > > > > > > > > > > > > > On Mon 26-08-19 16:15:38, Kirill A. Shutemov wrote: > > > > > > > > > > > > > > Unmapped completely pages will be freed with current code. Deferred split > > > > > > > > > > > > > > only applies to partly mapped THPs: at least on 4k of the THP is still > > > > > > > > > > > > > > mapped somewhere. > > > > > > > > > > > > > Hmm, I am probably misreading the code but at least current Linus' tree > > > > > > > > > > > > > reads page_remove_rmap -> [page_remove_anon_compound_rmap ->\ deferred_split_huge_page even > > > > > > > > > > > > > for fully mapped THP. > > > > > > > > > > > > Well, you read correctly, but it was not intended. I screwed it up at some > > > > > > > > > > > > point. > > > > > > > > > > > > > > > > > > > > > > > > See the patch below. It should make it work as intened. > > > > > > > > > > > > > > > > > > > > > > > > It's not bug as such, but inefficientcy. We add page to the queue where > > > > > > > > > > > > it's not needed. > > > > > > > > > > > But that adding to queue doesn't affect whether the page will be freed > > > > > > > > > > > immediately if there are no more partial mappings, right? I don't see > > > > > > > > > > > deferred_split_huge_page() pinning the page. > > > > > > > > > > > So your patch wouldn't make THPs freed immediately in cases where they > > > > > > > > > > > haven't been freed before immediately, it just fixes a minor > > > > > > > > > > > inefficiency with queue manipulation? > > > > > > > > > > Ohh, right. I can see that in free_transhuge_page now. So fully mapped > > > > > > > > > > THPs really do not matter and what I have considered an odd case is > > > > > > > > > > really happening more often. > > > > > > > > > > > > > > > > > > > > That being said this will not help at all for what Yang Shi is seeing > > > > > > > > > > and we need a more proactive deferred splitting as I've mentioned > > > > > > > > > > earlier. > > > > > > > > > It was not intended to fix the issue. It's fix for current logic. I'm > > > > > > > > > playing with the work approach now. > > > > > > > > Below is what I've come up with. It appears to be functional. > > > > > > > > > > > > > > > > Any comments? > > > > > > > > > > > > > > Thanks, Kirill and Michal. Doing split more proactive is definitely a choice > > > > > > > to eliminate huge accumulated deferred split THPs, I did think about this > > > > > > > approach before I came up with memcg aware approach. But, I thought this > > > > > > > approach has some problems: > > > > > > > > > > > > > > First of all, we can't prove if this is a universal win for the most > > > > > > > workloads or not. For some workloads (as I mentioned about our usecase), we > > > > > > > do see a lot THPs accumulated for a while, but they are very short-lived for > > > > > > > other workloads, i.e. kernel build. > > > > > > > > > > > > > > Secondly, it may be not fair for some workloads which don't generate too > > > > > > > many deferred split THPs or those THPs are short-lived. Actually, the cpu > > > > > > > time is abused by the excessive deferred split THPs generators, isn't it? > > > > > > > > > > > > Yes this is indeed true. Do we have any idea on how much time that > > > > > > actually is? > > > > > > > > > > For uncontented case, splitting 1G worth of pages (2MiB x 512) takes a bit > > > > > more than 50 ms in my setup. But it's best-case scenario: pages not shared > > > > > across multiple processes, no contention on ptl, page lock, etc. > > > > > > > > Any idea about a bad case? > > > > > > Not really. > > > > > > How bad you want it to get? How many processes share the page? Access > > > pattern? Locking situation? > > > > Let's say how hard a regular user can make this? > > > > > Worst case scenarion: no progress on splitting due to pins or locking > > > conflicts (trylock failure). > > > > > > > > > > With memcg awareness, the deferred split THPs actually are isolated and > > > > > > > capped by memcg. The long-lived deferred split THPs can't be accumulated too > > > > > > > many due to the limit of memcg. And, cpu time spent in splitting them would > > > > > > > just account to the memcgs who generate that many deferred split THPs, who > > > > > > > generate them who pay for it. This sounds more fair and we could achieve > > > > > > > much better isolation. > > > > > > > > > > > > On the other hand, deferring the split and free up a non trivial amount > > > > > > of memory is a problem I consider quite serious because it affects not > > > > > > only the memcg workload which has to do the reclaim but also other > > > > > > consumers of memory beucase large memory blocks could be used for higher > > > > > > order allocations. > > > > > > > > > > Maybe instead of drive the split from number of pages on queue we can take > > > > > a hint from compaction that is struggles to get high order pages? > > > > > > > > This is still unbounded in time. > > > > > > I'm not sure we should focus on time. > > > > > > We need to make sure that we don't overal system health worse. Who cares > > > if we have pages on deferred split list as long as we don't have other > > > user for the memory? > > > > We do care for all those users which do not want to get stalled when > > requesting that memory. And you cannot really predict that, right? So > > the sooner the better. Modulo time wasted for the pointless splitting of > > course. I am afraid defining the best timing here is going to be hard > > but let's focus on workloads that are known to generate partial THPs and > > see how that behaves. > > I'm supposed we are just concerned by the global memory pressure > incurred by the excessive deferred split THPs. As long as no other > users for that memory we don't have to waste time to care about it. > So, I'm wondering why not we do harder in kswapd? kswapd is already late. There shouldn't be any need for the reclaim as long as there is a lot of memory that can be directly freed. > Currently, deferred split THPs get shrunk like slab. The number of > objects scanned is determined by some factors, i.e. scan priority, > shrinker->seeks, etc, to avoid over reclaim for filesystem caches to > avoid extra I/O. But, we don't have to worry about over reclaim for > deferred split THPs, right? We definitely could shrink them more > aggressively in kswapd context. This is certainly possible. I am just wondering why should we cram this into the reclaim when we have a reasonable trigger to do that. > For example, we could simply set shrinker->seeks to 0, now it is > DEFAULT_SEEKS. > > And, we also could consider boost water mark to wake up kswapd earlier > once we see excessive deferred split THPs accumulated. This has other side effect, right? -- Michal Hocko SUSE Labs