From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1161735AbdAEPvG (ORCPT ); Thu, 5 Jan 2017 10:51:06 -0500 Received: from mx2.suse.de ([195.135.220.15]:52463 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1161824AbdAEPu5 (ORCPT ); Thu, 5 Jan 2017 10:50:57 -0500 Date: Thu, 5 Jan 2017 16:50:54 +0100 From: Michal Hocko To: Vlastimil Babka Cc: Mel Gorman , David Rientjes , Andrew Morton , Jonathan Corbet , "Kirill A. Shutemov" , linux-kernel@vger.kernel.org, linux-mm@kvack.org Subject: Re: [patch] mm, thp: add new background defrag option Message-ID: <20170105155053.GW21618@dhcp22.suse.cz> References: <20170105101330.bvhuglbbeudubgqb@techsingularity.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.6.0 (2016-04-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu 05-01-17 14:58:47, Vlastimil Babka wrote: [...] > I'm not a fan of either name, so I've tried to implement my own > suggestion. Turns out it was easier than expected, as there's no kernel > boot option for "defer", just for "enabled", so that particular worry > was unfounded. > > And personally I think that it's less confusing when one can enable defer > and madvise together (and not any other combination), than having to dig > up the difference between "defer" and "background". > > I have only tested the sysfs manipulation, not actual THP, but seems to me > that alloc_hugepage_direct_gfpmask() already happens to process the flags > in a way that it works as expected. IMHO this looks indeed much simpler implementation wise, more consistent from the semantic point of view and less confusing from the usage POV. > diff --git a/mm/huge_memory.c b/mm/huge_memory.c > index 10eedbf14421..cc5ae86169a8 100644 > --- a/mm/huge_memory.c > +++ b/mm/huge_memory.c > @@ -150,7 +150,16 @@ static ssize_t triple_flag_store(struct kobject *kobj, > enum transparent_hugepage_flag deferred, > enum transparent_hugepage_flag req_madv) > { > - if (!memcmp("defer", buf, > + if (!memcmp("defer madvise", buf, > + min(sizeof("defer madvise")-1, count)) > + || !memcmp("madvise defer", buf, > + min(sizeof("madvise defer")-1, count))) { > + if (enabled == deferred) > + return -EINVAL; > + clear_bit(enabled, &transparent_hugepage_flags); > + set_bit(req_madv, &transparent_hugepage_flags); > + set_bit(deferred, &transparent_hugepage_flags); > + } else if (!memcmp("defer", buf, > min(sizeof("defer")-1, count))) { > if (enabled == deferred) > return -EINVAL; > @@ -251,9 +260,12 @@ static ssize_t defrag_show(struct kobject *kobj, > { > if (test_bit(TRANSPARENT_HUGEPAGE_DEFRAG_DIRECT_FLAG, &transparent_hugepage_flags)) > return sprintf(buf, "[always] defer madvise never\n"); > - if (test_bit(TRANSPARENT_HUGEPAGE_DEFRAG_KSWAPD_FLAG, &transparent_hugepage_flags)) > - return sprintf(buf, "always [defer] madvise never\n"); > - else if (test_bit(TRANSPARENT_HUGEPAGE_DEFRAG_REQ_MADV_FLAG, &transparent_hugepage_flags)) > + if (test_bit(TRANSPARENT_HUGEPAGE_DEFRAG_KSWAPD_FLAG, &transparent_hugepage_flags)) { > + if (test_bit(TRANSPARENT_HUGEPAGE_DEFRAG_REQ_MADV_FLAG, &transparent_hugepage_flags)) > + return sprintf(buf, "always [defer] [madvise] never\n"); > + else > + return sprintf(buf, "always [defer] madvise never\n"); > + } else if (test_bit(TRANSPARENT_HUGEPAGE_DEFRAG_REQ_MADV_FLAG, &transparent_hugepage_flags)) > return sprintf(buf, "always defer [madvise] never\n"); > else > return sprintf(buf, "always defer madvise [never]\n"); -- Michal Hocko SUSE Labs