From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751999AbdAZAMY (ORCPT ); Wed, 25 Jan 2017 19:12:24 -0500 Received: from mail-it0-f43.google.com ([209.85.214.43]:34828 "EHLO mail-it0-f43.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750853AbdAZAMW (ORCPT ); Wed, 25 Jan 2017 19:12:22 -0500 MIME-Version: 1.0 In-Reply-To: <1485380634.2998.161.camel@decadent.org.uk> References: <9318903980969a0e378dab2de4d803397adcd3cc.1485377903.git.luto@kernel.org> <1485380634.2998.161.camel@decadent.org.uk> From: Kees Cook Date: Wed, 25 Jan 2017 16:12:21 -0800 X-Google-Sender-Auth: m-h92YAboDa3hn7uVC1zMHbnPlA Message-ID: Subject: Re: [PATCH 1/2] fs: Check f_cred instead of current's creds in should_remove_suid() To: Ben Hutchings Cc: Andy Lutomirski , "security@kernel.org" , Konstantin Khlebnikov , Alexander Viro , Willy Tarreau , "linux-mm@kvack.org" , Andrew Morton , yalin wang , Linux Kernel Mailing List , Jan Kara , Linux FS Devel , "# 3.4.x" Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Jan 25, 2017 at 1:43 PM, Ben Hutchings wrote: > On Wed, 2017-01-25 at 13:06 -0800, Andy Lutomirski wrote: >> If an unprivileged program opens a setgid file for write and passes >> the fd to a privileged program and the privileged program writes to >> it, we currently fail to clear the setgid bit. Fix it by checking >> f_cred instead of current's creds whenever a struct file is >> involved. > [...] > > What if, instead, a privileged program passes the fd to an un > unprivileged program? It sounds like a bad idea to start with, but at > least currently the unprivileged program is going to clear the setgid > bit when it writes. This change would make that behaviour more > dangerous. > > Perhaps there should be a capability check on both the current > credentials and file credentials? (I realise that we've considered > file credential checks to be sufficient elsewhere, but those cases > involved virtual files with special semantics, where it's clearer that > a privileged process should not pass them to an unprivileged process.) We need a set of self-tests for this whole area. :( There are so many corner cases. We still have an unfixed corner case with mmap writes not clearing set*id bits that I tried to solve last year... -Kees -- Kees Cook Nexus Security