From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wr0-f197.google.com (mail-wr0-f197.google.com [209.85.128.197]) by kanga.kvack.org (Postfix) with ESMTP id 4A9B86B0038 for ; Thu, 19 Oct 2017 15:35:45 -0400 (EDT) Received: by mail-wr0-f197.google.com with SMTP id g10so3367375wrg.6 for ; Thu, 19 Oct 2017 12:35:45 -0700 (PDT) Received: from mx2.suse.de (mx2.suse.de. [195.135.220.15]) by mx.google.com with ESMTPS id w63si1614949wmb.195.2017.10.19.12.35.43 for (version=TLS1 cipher=AES128-SHA bits=128/128); Thu, 19 Oct 2017 12:35:44 -0700 (PDT) Date: Thu, 19 Oct 2017 21:35:42 +0200 From: Michal Hocko Subject: Re: [PATCH] mm: mlock: remove lru_add_drain_all() Message-ID: <20171019193542.l5baqknxnfhljjkr@dhcp22.suse.cz> References: <20171018231730.42754-1-shakeelb@google.com> <20171019123206.3etacullgnarbnad@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: owner-linux-mm@kvack.org List-ID: To: Shakeel Butt Cc: Andrew Morton , "Kirill A. Shutemov" , Vlastimil Babka , Joonsoo Kim , Minchan Kim , Yisheng Xie , Ingo Molnar , Greg Thelen , Hugh Dickins , Linux MM , LKML On Thu 19-10-17 12:19:26, Shakeel Butt wrote: > On Thu, Oct 19, 2017 at 5:32 AM, Michal Hocko wrote: > > On Wed 18-10-17 16:17:30, Shakeel Butt wrote: > >> Recently we have observed high latency in mlock() in our generic > >> library and noticed that users have started using tmpfs files even > >> without swap and the latency was due to expensive remote LRU cache > >> draining. > > > > some numbers would be really nice > > > > On a production workload, customers complained that single mlock() > call took around 10 seconds on mapped tmpfs files and the perf profile > showed lru_add_drain_all as culprit. draining can take some time. I wouldn't expect orders of seconds so perf data would be definitely helpful in the changelog. [...] > > Is this really true? lru_add_drain_all will flush the previously cached > > LRU pages. We are not flushing after the pages have been faulted in so > > this might not do anything wrt. mlocked pages, right? > > > > Sorry for the confusion. I wanted to say that if the pages which are > being mlocked are on caches of remote cpus then lru_add_drain_all will > move them to their corresponding LRUs and then remaining functionality > of mlock will move them again from their evictable LRUs to unevictable > LRU. yes, but the point is that we are draining pages which might be not directly related to pages which _will_ be mlocked by the syscall. In fact those will stay on the cache. This is the primary reason why this draining doesn't make much sense. Or am I still misunderstanding what you are saying here? -- Michal Hocko SUSE Labs -- 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