From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-lf0-f70.google.com (mail-lf0-f70.google.com [209.85.215.70]) by kanga.kvack.org (Postfix) with ESMTP id EA0786B0069 for ; Tue, 4 Oct 2016 09:14:21 -0400 (EDT) Received: by mail-lf0-f70.google.com with SMTP id x79so10408426lff.2 for ; Tue, 04 Oct 2016 06:14:21 -0700 (PDT) Received: from mail-lf0-x241.google.com (mail-lf0-x241.google.com. [2a00:1450:4010:c07::241]) by mx.google.com with ESMTPS id d193si1984645lfg.277.2016.10.04.06.14.19 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 04 Oct 2016 06:14:20 -0700 (PDT) Received: by mail-lf0-x241.google.com with SMTP id b75so2768706lfg.3 for ; Tue, 04 Oct 2016 06:14:19 -0700 (PDT) Date: Tue, 4 Oct 2016 16:14:17 +0300 From: Vladimir Davydov Subject: Re: [PATCH 1/2] mm: memcontrol: use special workqueue for creating per-memcg caches Message-ID: <20161004131417.GC1862@esperanza> References: <20161003120641.GC26768@dhcp22.suse.cz> <20161003123505.GA1862@esperanza> <20161003131930.GE26768@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20161003131930.GE26768@dhcp22.suse.cz> Sender: owner-linux-mm@kvack.org List-ID: To: Michal Hocko , Andrew Morton Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org On Mon, Oct 03, 2016 at 03:19:31PM +0200, Michal Hocko wrote: > On Mon 03-10-16 15:35:06, Vladimir Davydov wrote: > > On Mon, Oct 03, 2016 at 02:06:42PM +0200, Michal Hocko wrote: > > > On Sat 01-10-16 16:56:47, Vladimir Davydov wrote: > > > > Creating a lot of cgroups at the same time might stall all worker > > > > threads with kmem cache creation works, because kmem cache creation is > > > > done with the slab_mutex held. To prevent that from happening, let's use > > > > a special workqueue for kmem cache creation with max in-flight work > > > > items equal to 1. > > > > > > > > Link: https://bugzilla.kernel.org/show_bug.cgi?id=172981 > > > > > > This looks like a regression but I am not really sure I understand what > > > has caused it. We had the WQ based cache creation since kmem was > > > introduced more or less. So is it 801faf0db894 ("mm/slab: lockless > > > decision to grow cache") which was pointed by bisection that changed the > > > timing resp. relaxed the cache creation to the point that would allow > > > this runaway? > > > > It is in case of SLAB. For SLUB the issue was caused by commit > > 81ae6d03952c ("mm/slub.c: replace kick_all_cpus_sync() with > > synchronize_sched() in kmem_cache_shrink()"). > > OK, thanks for the confirmation. This would be useful in the changelog > imho. > > > > This would be really useful for the stable backport > > > consideration. > > > > > > Also, if I understand the fix correctly, now we do limit the number of > > > workers to 1 thread. Is this really what we want? Wouldn't it be > > > possible that few memcgs could starve others fromm having their cache > > > created? What would be the result, missed charges? > > > > Now kmem caches are created in FIFO order, i.e. if one memcg called > > kmem_cache_alloc on a non-existent cache before another, it will be > > served first. > > I do not see where this FIFO is guaranteed. > __memcg_schedule_kmem_cache_create doesn't seem to be using ordered WQ. Yeah, you're right - I thought max_active implies ordering, but it doesn't. Then we can use an ordered workqueue. Here's the updated patch: