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 Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id B933BC43334 for ; Fri, 3 Jun 2022 17:56:25 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1345807AbiFCR4Y (ORCPT ); Fri, 3 Jun 2022 13:56:24 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:57820 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1346325AbiFCRvE (ORCPT ); Fri, 3 Jun 2022 13:51:04 -0400 Received: from ams.source.kernel.org (ams.source.kernel.org [145.40.68.75]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C218D54020; Fri, 3 Jun 2022 10:47:32 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ams.source.kernel.org (Postfix) with ESMTPS id 7DFD3B823B0; Fri, 3 Jun 2022 17:47:31 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 9C78AC3411C; Fri, 3 Jun 2022 17:47:29 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1654278450; bh=V/yGTt7Ocp4ajWdl5FHnoH8c4RKdCJlTJGqL+no4af0=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=CI+ZXyKWUfXJaEliAU14xeyRbAkYLjq8qibbJ/E/j2NrdPSH3cLawwhnafVlhBeiJ RN5FYfyacyPY3Ho7KjBivf9IKihRNPCirA73GufBelTjS2+QVlrmYyoTSFrJ1gyLBH dZmfkP1Gb85cSenX5daB44m8oJ+BxHhBZRw+qR6s= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Sultan Alsawaf , Minchan Kim , Nitin Gupta , Sergey Senozhatsky , Andrew Morton Subject: [PATCH 5.10 37/53] zsmalloc: fix races between asynchronous zspage free and page migration Date: Fri, 3 Jun 2022 19:43:22 +0200 Message-Id: <20220603173819.801075996@linuxfoundation.org> X-Mailer: git-send-email 2.36.1 In-Reply-To: <20220603173818.716010877@linuxfoundation.org> References: <20220603173818.716010877@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Sultan Alsawaf commit 2505a981114dcb715f8977b8433f7540854851d8 upstream. The asynchronous zspage free worker tries to lock a zspage's entire page list without defending against page migration. Since pages which haven't yet been locked can concurrently migrate off the zspage page list while lock_zspage() churns away, lock_zspage() can suffer from a few different lethal races. It can lock a page which no longer belongs to the zspage and unsafely dereference page_private(), it can unsafely dereference a torn pointer to the next page (since there's a data race), and it can observe a spurious NULL pointer to the next page and thus not lock all of the zspage's pages (since a single page migration will reconstruct the entire page list, and create_page_chain() unconditionally zeroes out each list pointer in the process). Fix the races by using migrate_read_lock() in lock_zspage() to synchronize with page migration. Link: https://lkml.kernel.org/r/20220509024703.243847-1-sultan@kerneltoast.com Fixes: 77ff465799c602 ("zsmalloc: zs_page_migrate: skip unnecessary loops but not return -EBUSY if zspage is not inuse") Signed-off-by: Sultan Alsawaf Acked-by: Minchan Kim Cc: Nitin Gupta Cc: Sergey Senozhatsky Cc: Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman --- mm/zsmalloc.c | 37 +++++++++++++++++++++++++++++++++---- 1 file changed, 33 insertions(+), 4 deletions(-) --- a/mm/zsmalloc.c +++ b/mm/zsmalloc.c @@ -1748,11 +1748,40 @@ static enum fullness_group putback_zspag */ static void lock_zspage(struct zspage *zspage) { - struct page *page = get_first_page(zspage); + struct page *curr_page, *page; - do { - lock_page(page); - } while ((page = get_next_page(page)) != NULL); + /* + * Pages we haven't locked yet can be migrated off the list while we're + * trying to lock them, so we need to be careful and only attempt to + * lock each page under migrate_read_lock(). Otherwise, the page we lock + * may no longer belong to the zspage. This means that we may wait for + * the wrong page to unlock, so we must take a reference to the page + * prior to waiting for it to unlock outside migrate_read_lock(). + */ + while (1) { + migrate_read_lock(zspage); + page = get_first_page(zspage); + if (trylock_page(page)) + break; + get_page(page); + migrate_read_unlock(zspage); + wait_on_page_locked(page); + put_page(page); + } + + curr_page = page; + while ((page = get_next_page(curr_page))) { + if (trylock_page(page)) { + curr_page = page; + } else { + get_page(page); + migrate_read_unlock(zspage); + wait_on_page_locked(page); + put_page(page); + migrate_read_lock(zspage); + } + } + migrate_read_unlock(zspage); } static int zs_init_fs_context(struct fs_context *fc)