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=-2.0 required=3.0 tests=DKIM_INVALID,DKIM_SIGNED, HEADER_FROM_DIFFERENT_DOMAINS,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 94078C432C0 for ; Mon, 18 Nov 2019 12:14:58 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 275E9206D4 for ; Mon, 18 Nov 2019 12:14:58 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (2048-bit key) header.d=infradead.org header.i=@infradead.org header.b="fR9iKXTX" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 275E9206D4 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=infradead.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 919A76B0006; Mon, 18 Nov 2019 07:14:57 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 8C9956B0007; Mon, 18 Nov 2019 07:14:57 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 793C06B000A; Mon, 18 Nov 2019 07:14:57 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0199.hostedemail.com [216.40.44.199]) by kanga.kvack.org (Postfix) with ESMTP id 63B016B0006 for ; Mon, 18 Nov 2019 07:14:57 -0500 (EST) Received: from smtpin11.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay05.hostedemail.com (Postfix) with SMTP id F0F3D181AEF00 for ; Mon, 18 Nov 2019 12:14:56 +0000 (UTC) X-FDA: 76169292192.11.push99_2b26a14737f3b X-HE-Tag: push99_2b26a14737f3b X-Filterd-Recvd-Size: 7419 Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) by imf28.hostedemail.com (Postfix) with ESMTP for ; Mon, 18 Nov 2019 12:14:56 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=infradead.org; s=bombadil.20170209; h=In-Reply-To:Content-Transfer-Encoding :Content-Type:MIME-Version:References:Message-ID:Subject:Cc:To:From:Date: Sender:Reply-To:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Id:List-Help: List-Unsubscribe:List-Subscribe:List-Post:List-Owner:List-Archive; bh=GhE7qMvIwwf/aYu9her0sBn1vDdPh7s1IE5BSkrLez4=; b=fR9iKXTXT4+XD+NhJdh5EwWQ5l ASFMW7HjICEpN4ugb9W379oArY6sdOuc5ijoipM0viY1tQs2o7BSudAaG2wbsNBla7gxYICZ1Jvdo TS3mbVdewoayXAVYyQ4/1tO3QuaU1VoUy4NV1nKzvLAUkcvQo3jtp0lJg5JsjfJmlPBiA2hArgz9k YKImF07gPYx+3ZOcONkLa1blhaXcisTK7+aRTTGkeUoZ4vPMrtvlKVgffH0SWuHKYlSm/S1SWtjli FFy6w7B8A1tbTW97fCdNHvYxj7PQuqrnTrzV9gmrXhwrTOdbySmOG40qHxW6dd5vJ+/y/xqIiwlTD /tVl7L9w==; Received: from willy by bombadil.infradead.org with local (Exim 4.92.3 #3 (Red Hat Linux)) id 1iWfvf-0002Rb-5Z; Mon, 18 Nov 2019 12:14:51 +0000 Date: Mon, 18 Nov 2019 04:14:51 -0800 From: Matthew Wilcox To: Alex Shi Cc: cgroups@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, akpm@linux-foundation.org, mgorman@techsingularity.net, tj@kernel.org, hughd@google.com, khlebnikov@yandex-team.ru, daniel.m.jordan@oracle.com, yang.shi@linux.alibaba.com, Johannes Weiner , Michal Hocko , Vladimir Davydov , Roman Gushchin , Shakeel Butt , Chris Down , Thomas Gleixner , Vlastimil Babka , Qian Cai , Andrey Ryabinin , "Kirill A. Shutemov" , =?iso-8859-1?B?Suly9G1l?= Glisse , Andrea Arcangeli , David Rientjes , "Aneesh Kumar K.V" , swkhack , "Potyra, Stefan" , Mike Rapoport , Stephen Rothwell , Colin Ian King , Jason Gunthorpe , Mauro Carvalho Chehab , Peng Fan , Nikolay Borisov , Ira Weiny , Kirill Tkhai , Yafang Shao Subject: Re: [PATCH v3 3/7] mm/lru: replace pgdat lru_lock with lruvec lock Message-ID: <20191118121451.GG20752@bombadil.infradead.org> References: <1573874106-23802-1-git-send-email-alex.shi@linux.alibaba.com> <1573874106-23802-4-git-send-email-alex.shi@linux.alibaba.com> <20191116043806.GD20752@bombadil.infradead.org> <0bfa9a03-b095-df83-9cfd-146da9aab89a@linux.alibaba.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <0bfa9a03-b095-df83-9cfd-146da9aab89a@linux.alibaba.com> User-Agent: Mutt/1.12.1 (2019-06-15) Content-Transfer-Encoding: quoted-printable X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: On Mon, Nov 18, 2019 at 07:55:43PM +0800, Alex Shi wrote: > =E5=9C=A8 2019/11/16 =E4=B8=8B=E5=8D=8812:38, Matthew Wilcox =E5=86=99=E9= =81=93: > > On Sat, Nov 16, 2019 at 11:15:02AM +0800, Alex Shi wrote: > >> This is the main patch to replace per node lru_lock with per memcg > >> lruvec lock. It also fold the irqsave flags into lruvec. > >=20 > > I have to say, I don't love the part where we fold the irqsave flags > > into the lruvec. I know it saves us an argument, but it opens up the > > possibility of mismatched expectations. eg we currently have: > >=20 > > static void __split_huge_page(struct page *page, struct list_head *li= st, > > struct lruvec *lruvec, pgoff_t end) > > { > > ... > > spin_unlock_irqrestore(&lruvec->lru_lock, lruvec->irqflags); > >=20 > > so if we introduce a new caller, we have to be certain that this call= er > > is also using lock_page_lruvec_irqsave() and not lock_page_lruvec_irq= (). > > I can't think of a way to make the compiler enforce that, and if we d= on't, > > then we can get some odd crashes with interrupts being unexpectedly > > enabled or disabled, depending on how ->irqflags was used last. > >=20 > > So it makes the code more subtle. And that's not a good thing. >=20 > Hi Matthew, >=20 > Thanks for comments! >=20 > Here, the irqflags is bound, and belong to lruvec, merging them into to= gether helps us to take them as whole, and thus reduce a unnecessary code= clues. It's not bound to the lruvec, though. Call chain A uses it and call chai= n B doesn't. If it was always used by every call chain, I'd see your point= , but we have call chains which don't use it, and so it adds complexity. > As your concern for a 'new' caller, since __split_huge_page is a static= helper here, no distub for anyothers. Even though it's static, there may be other callers within the same file. Or somebody may decide to make it non-static in the future. I think it's actually clearer to keep the irqflags as a separate parameter. > >> +static inline struct lruvec *lock_page_lruvec_irq(struct page *page= , > >> + struct pglist_data *pgdat) > >> +{ > >> + struct lruvec *lruvec =3D mem_cgroup_page_lruvec(page, pgdat); > >> + > >> + spin_lock_irq(&lruvec->lru_lock); > >> + > >> + return lruvec; > >> +} > >=20 > > ... > >=20 > >> +static struct lruvec *lock_page_lru(struct page *page, int *isolate= d) > >> { > >> pg_data_t *pgdat =3D page_pgdat(page); > >> + struct lruvec *lruvec =3D lock_page_lruvec_irq(page, pgdat); > >> =20 > >> - spin_lock_irq(&pgdat->lru_lock); > >> if (PageLRU(page)) { > >> - struct lruvec *lruvec; > >> =20 > >> - lruvec =3D mem_cgroup_page_lruvec(page, pgdat); > >> ClearPageLRU(page); > >> del_page_from_lru_list(page, lruvec, page_lru(page)); > >> *isolated =3D 1; > >> } else > >> *isolated =3D 0; > >> + > >> + return lruvec; > >> } > >=20 > > But what if the page is !PageLRU? What lruvec did we just lock? >=20 > like original pgdat->lru_lock, we need the lock from PageLRU racing. An= d it the lruvec which the page should be. >=20 >=20 > > According to the comments on mem_cgroup_page_lruvec(), > >=20 > > * This function is only safe when following the LRU page isolation > > * and putback protocol: the LRU lock must be held, and the page must > > * either be PageLRU() or the caller must have isolated/allocated it. > >=20 > > and now it's being called in order to find out which LRU lock to take= . > > So this comment needs to be updated, if it's wrong, or this patch has > > a race. >=20 >=20 > Yes, the function reminder is a bit misunderstanding with new patch, Ho= w about the following changes: >=20 > - * This function is only safe when following the LRU page isolation > - * and putback protocol: the LRU lock must be held, and the page must > - * either be PageLRU() or the caller must have isolated/allocated it. > + * The caller needs to grantee the page's mem_cgroup is undisturbed du= ring > + * using. That could be done by lock_page_memcg or lock_page_lruvec. I don't understand how lock_page_lruvec makes this guarantee. I'll look at the code again and see if I can understand that.