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=-8.6 required=3.0 tests=BAYES_00,DKIM_INVALID, DKIM_SIGNED,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham 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 91F4BC433E0 for ; Mon, 25 Jan 2021 16:26:03 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 0CA30221E3 for ; Mon, 25 Jan 2021 16:26:02 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0CA30221E3 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 751788D0010; Mon, 25 Jan 2021 11:26:02 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 701CE8D0001; Mon, 25 Jan 2021 11:26:02 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 63F188D0010; Mon, 25 Jan 2021 11:26:02 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0005.hostedemail.com [216.40.44.5]) by kanga.kvack.org (Postfix) with ESMTP id 4B8588D0001 for ; Mon, 25 Jan 2021 11:26:02 -0500 (EST) Received: from smtpin21.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay03.hostedemail.com (Postfix) with ESMTP id 041D6824999B for ; Mon, 25 Jan 2021 16:26:02 +0000 (UTC) X-FDA: 77744824164.21.bear21_3c0233227587 Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin21.hostedemail.com (Postfix) with ESMTP id AB3C1180442C7 for ; Mon, 25 Jan 2021 16:26:01 +0000 (UTC) X-HE-Tag: bear21_3c0233227587 X-Filterd-Recvd-Size: 7044 Received: from casper.infradead.org (casper.infradead.org [90.155.50.34]) by imf11.hostedemail.com (Postfix) with ESMTP for ; Mon, 25 Jan 2021 16:26:00 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=infradead.org; s=casper.20170209; h=In-Reply-To:Content-Type:MIME-Version: References:Message-ID:Subject:Cc:To:From:Date:Sender:Reply-To: Content-Transfer-Encoding:Content-ID:Content-Description; bh=ixdyS05p7ztMglop7rXtNcIDaMbcHqMuwN4v+O3qLuk=; b=tnzyKZlg/LeBjujzlgIYG7CdTJ pJ3QEFg6nagxlRxheDtmuGqZtDD6zVrtryCaGERZqm7fsvdPJ5qBk2Hx153fRH9Z3ZwIxkGmUgiCw dBpEcoH6QUDvq6hiloccNTRrBZtf2Qkv+vRrmJ+NJJ+P/7XFOeXf/vIqFtXKjE4+jTLI5Y/MPWynt xJhRIzo1dB7vbaORp8VzLmpZu6EV6PFTazSodEDWYpcZkV3MY5cWxXQJ9IAI1h2OVsa2RQodyS5QZ h0EqJ+UQ1sgANOxI/ewQLgo7G01amZrCPLFpQ/pCscl8fJbQVGucGy+8sTp69JwTN+BRmSTi1jWXp Z0gos8Fw==; Received: from willy by casper.infradead.org with local (Exim 4.94 #2 (Red Hat Linux)) id 1l44fq-004Oy8-6O; Mon, 25 Jan 2021 16:25:11 +0000 Date: Mon, 25 Jan 2021 16:25:06 +0000 From: Matthew Wilcox To: Michal Hocko Cc: Waiman Long , Andrew Morton , Johannes Weiner , Alex Shi , linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH] mm/filemap: Adding missing mem_cgroup_uncharge() to __add_to_page_cache_locked() Message-ID: <20210125162506.GF308988@casper.infradead.org> References: <20210125042441.20030-1-longman@redhat.com> <20210125092815.GB827@dhcp22.suse.cz> <20210125160328.GP827@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20210125160328.GP827@dhcp22.suse.cz> 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, Jan 25, 2021 at 05:03:28PM +0100, Michal Hocko wrote: > On Mon 25-01-21 10:57:54, Waiman Long wrote: > > On 1/25/21 4:28 AM, Michal Hocko wrote: > > > On Sun 24-01-21 23:24:41, Waiman Long wrote: > > > > The commit 3fea5a499d57 ("mm: memcontrol: convert page > > > > cache to a new mem_cgroup_charge() API") introduced a bug in > > > > __add_to_page_cache_locked() causing the following splat: > > > > > > > > [ 1570.068330] page dumped because: VM_BUG_ON_PAGE(page_memcg(page)) > > > > [ 1570.068333] pages's memcg:ffff8889a4116000 > > > > [ 1570.068343] ------------[ cut here ]------------ > > > > [ 1570.068346] kernel BUG at mm/memcontrol.c:2924! > > > > [ 1570.068355] invalid opcode: 0000 [#1] SMP KASAN PTI > > > > [ 1570.068359] CPU: 35 PID: 12345 Comm: cat Tainted: G S W I 5.11.0-rc4-debug+ #1 > > > > [ 1570.068363] Hardware name: HP HP Z8 G4 Workstation/81C7, BIOS P60 v01.25 12/06/2017 > > > > [ 1570.068365] RIP: 0010:commit_charge+0xf4/0x130 > > > > : > > > > [ 1570.068375] RSP: 0018:ffff8881b38d70e8 EFLAGS: 00010286 > > > > [ 1570.068379] RAX: 0000000000000000 RBX: ffffea00260ddd00 RCX: 0000000000000027 > > > > [ 1570.068382] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffff88907ebe05a8 > > > > [ 1570.068384] RBP: ffffea00260ddd00 R08: ffffed120fd7c0b6 R09: ffffed120fd7c0b6 > > > > [ 1570.068386] R10: ffff88907ebe05ab R11: ffffed120fd7c0b5 R12: ffffea00260ddd38 > > > > [ 1570.068389] R13: ffff8889a4116000 R14: ffff8889a4116000 R15: 0000000000000001 > > > > [ 1570.068391] FS: 00007ff039638680(0000) GS:ffff88907ea00000(0000) knlGS:0000000000000000 > > > > [ 1570.068394] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > > > > [ 1570.068396] CR2: 00007f36f354cc20 CR3: 00000008a0126006 CR4: 00000000007706e0 > > > > [ 1570.068398] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 > > > > [ 1570.068400] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 > > > > [ 1570.068402] PKRU: 55555554 > > > > [ 1570.068404] Call Trace: > > > > [ 1570.068407] mem_cgroup_charge+0x175/0x770 > > > > [ 1570.068413] __add_to_page_cache_locked+0x712/0xad0 > > > > [ 1570.068439] add_to_page_cache_lru+0xc5/0x1f0 > > > > [ 1570.068461] cachefiles_read_or_alloc_pages+0x895/0x2e10 [cachefiles] > > > > [ 1570.068524] __fscache_read_or_alloc_pages+0x6c0/0xa00 [fscache] > > > > [ 1570.068540] __nfs_readpages_from_fscache+0x16d/0x630 [nfs] > > > > [ 1570.068585] nfs_readpages+0x24e/0x540 [nfs] > > > > [ 1570.068693] read_pages+0x5b1/0xc40 > > > > [ 1570.068711] page_cache_ra_unbounded+0x460/0x750 > > > > [ 1570.068729] generic_file_buffered_read_get_pages+0x290/0x1710 > > > > [ 1570.068756] generic_file_buffered_read+0x2a9/0xc30 > > > > [ 1570.068832] nfs_file_read+0x13f/0x230 [nfs] > > > > [ 1570.068872] new_sync_read+0x3af/0x610 > > > > [ 1570.068901] vfs_read+0x339/0x4b0 > > > > [ 1570.068909] ksys_read+0xf1/0x1c0 > > > > [ 1570.068920] do_syscall_64+0x33/0x40 > > > > [ 1570.068926] entry_SYSCALL_64_after_hwframe+0x44/0xa9 > > > > [ 1570.068930] RIP: 0033:0x7ff039135595 > > > > > > > > Before that commit, there was a try_charge() and commit_charge() > > > > in __add_to_page_cache_locked(). These 2 separated charge functions > > > > were replaced by a single mem_cgroup_charge(). However, it forgot > > > > to add a matching mem_cgroup_uncharge() when the xarray insertion > > > > failed with the page released back to the pool. Fix this by adding a > > > > mem_cgroup_uncharge() call when insertion error happens. > > > > > > > > Fixes: 3fea5a499d57 ("mm: memcontrol: convert page cache to a new mem_cgroup_charge() API") > > > > Signed-off-by: Waiman Long > > > OK, this is indeed a subtle bug. The patch aimed at simplifying the > > > charge lifetime so that users do not really have to think about when to > > > uncharge as that happens when the page is freed. fscache somehow breaks > > > that assumption because it doesn't free up pages but it keeps some of > > > them in the cache. > > > > > > I have tried to wrap my head around the cached object life time in > > > fscache but failed and got lost in the maze. Is this the only instance > > > of the problem? Would it make more sense to explicitly handle charges in > > > the fscache code or there are other potential users to fall into this > > > trap? > > > > There may be other places that have similar problem. I focus on the > > filemap.c case as I have a test case that can reliably produce the bug > > splat. This patch does fix it for my test case. > > I believe this needs a more general fix than catching a random places > which you can trigger. Would it make more sense to address this at the > fscache level and always make sure that a page returned to the pool is > always uncharged instead? I believe you mean "page cache" -- there is a separate thing called 'fscache' which is used to cache network filesystems. I don't understand the memcg code at all, so I have no useful feedback on what you're saying other than this.