Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753991AbeAKFbJ (ORCPT + 1 other); Thu, 11 Jan 2018 00:31:09 -0500 Received: from mail-pf0-f194.google.com ([209.85.192.194]:44949 "EHLO mail-pf0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751852AbeAKFbF (ORCPT ); Thu, 11 Jan 2018 00:31:05 -0500 X-Google-Smtp-Source: ACJfBovGMVmkYJdPCZSFGx5aXql7PgPHCyM8n1eEIw21oblYDBk0nJ2p1K9lJ7TbmRp64szOSVJKyw== From: Eric Biggers To: linux-fsdevel@vger.kernel.org Cc: Alexander Viro , Joe Lawrence , Michael Kerrisk , Willy Tarreau , Mikulas Patocka , "Luis R . Rodriguez" , Kees Cook , linux-kernel@vger.kernel.org, Eric Biggers Subject: [PATCH v2 7/7] pipe: read buffer limits atomically Date: Wed, 10 Jan 2018 21:29:02 -0800 Message-Id: <20180111052902.14409-8-ebiggers3@gmail.com> X-Mailer: git-send-email 2.15.1 In-Reply-To: <20180111052902.14409-1-ebiggers3@gmail.com> References: <20180111052902.14409-1-ebiggers3@gmail.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Return-Path: From: Eric Biggers The pipe buffer limits are accessed without any locking, and may be changed at any time by the sysctl handlers. In theory this could cause problems for expressions like the following: pipe_user_pages_hard && user_bufs > pipe_user_pages_hard ... since the assembly code might reference the 'pipe_user_pages_hard' memory location multiple times, and if the admin removes the limit by setting it to 0, there is a very brief window where processes could incorrectly observe the limit to be exceeded. Fix this by loading the limits with READ_ONCE() prior to use. Acked-by: Kees Cook Signed-off-by: Eric Biggers --- fs/pipe.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/fs/pipe.c b/fs/pipe.c index 458f866caf53..deeb303aa4ce 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -605,12 +605,16 @@ static unsigned long account_pipe_buffers(struct user_struct *user, static bool too_many_pipe_buffers_soft(unsigned long user_bufs) { - return pipe_user_pages_soft && user_bufs > pipe_user_pages_soft; + unsigned long soft_limit = READ_ONCE(pipe_user_pages_soft); + + return soft_limit && user_bufs > soft_limit; } static bool too_many_pipe_buffers_hard(unsigned long user_bufs) { - return pipe_user_pages_hard && user_bufs > pipe_user_pages_hard; + unsigned long hard_limit = READ_ONCE(pipe_user_pages_hard); + + return hard_limit && user_bufs > hard_limit; } static bool is_unprivileged_user(void) @@ -624,13 +628,14 @@ struct pipe_inode_info *alloc_pipe_info(void) unsigned long pipe_bufs = PIPE_DEF_BUFFERS; struct user_struct *user = get_current_user(); unsigned long user_bufs; + unsigned int max_size = READ_ONCE(pipe_max_size); pipe = kzalloc(sizeof(struct pipe_inode_info), GFP_KERNEL_ACCOUNT); if (pipe == NULL) goto out_free_uid; - if (pipe_bufs * PAGE_SIZE > pipe_max_size && !capable(CAP_SYS_RESOURCE)) - pipe_bufs = pipe_max_size >> PAGE_SHIFT; + if (pipe_bufs * PAGE_SIZE > max_size && !capable(CAP_SYS_RESOURCE)) + pipe_bufs = max_size >> PAGE_SHIFT; user_bufs = account_pipe_buffers(user, 0, pipe_bufs); -- 2.15.1