Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752555Ab2HLTcd (ORCPT ); Sun, 12 Aug 2012 15:32:33 -0400 Received: from mail-lpp01m010-f46.google.com ([209.85.215.46]:35000 "EHLO mail-lpp01m010-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750733Ab2HLTcb (ORCPT ); Sun, 12 Aug 2012 15:32:31 -0400 Date: Sun, 12 Aug 2012 23:32:10 +0400 From: Vasily Kulikov To: kernel-hardening@lists.openwall.com Cc: Al Viro , Andrew Morton , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, Eric Paris , Matthew Wilcox , Doug Ledford , Joe Korty , "Eric W. Biederman" , Ingo Molnar , David Howells , James Morris , linux-doc@vger.kernel.org, Dan Rosenberg Subject: Re: [kernel-hardening] [PATCH 1/2] fs: add link restrictions Message-ID: <20120812193210.GA16669@albatros> References: <1343262548-21743-1-git-send-email-keescook@chromium.org> <1343262548-21743-2-git-send-email-keescook@chromium.org> <20120808121938.GA9995@albatros> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2520 Lines: 62 On Sat, Aug 11, 2012 at 23:34 -0700, Kees Cook wrote: > On Wed, Aug 8, 2012 at 5:19 AM, Vasily Kulikov wrote: > > Hi Kees, > > > > On Wed, Jul 25, 2012 at 17:29 -0700, Kees Cook wrote: > >> +/** > >> + * safe_hardlink_source - Check for safe hardlink conditions > >> + * @inode: the source inode to hardlink from > >> + * > >> + * Return false if at least one of the following conditions: > >> + * - inode is not a regular file > >> + * - inode is setuid > >> + * - inode is setgid and group-exec > >> + * - access failure for read and write > >> + * > >> + * Otherwise returns true. > >> + */ > >> +static bool safe_hardlink_source(struct inode *inode) > >> +{ > >> + umode_t mode = inode->i_mode; > >> + > >> + /* Special files should not get pinned to the filesystem. */ > >> + if (!S_ISREG(mode)) > >> + return false; > >> + > >> + /* Setuid files should not get pinned to the filesystem. */ > >> + if (mode & S_ISUID) > >> + return false; > > > > We don't want to make hardlinks of SUID files, but we still allow to create > > hardlinks to SUID'ish cap'ed files. Probably check whether the inode is > > setcap'ed? > > Excellent idea. It doesn't look like there is anything "simple" to do > this already. It'd be close to get_file_caps() but without the bprm. > Maybe just get_vfs_caps_from_disk() and a walk of the caps? What would > you recommend? Yes, I think get_vfs_caps_from_disk() plus identifying whether any permitted or inheritable capability is set or effective bit is set. IOW, if there is _anything_ related to cababilities in the file, it is protected. > > Probably we can enhance this further and allow LSMs to define whether this > > particular file is special in LSM's point of view (IOW, it can be able to move > > a process to another security domain which is served by LSM). > > Yeah. Perhaps implementing the needed check above with a new security > check and have commoncaps do the vfs fetch with LSMs able to override? Yep. Then uid and gid checks from the above code should move to commoncaps function as default candidates for "privileged files" checks. Thanks, -- Vasily Kulikov http://www.openwall.com - bringing security into open computing environments -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/