Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932285AbcCHJQm (ORCPT ); Tue, 8 Mar 2016 04:16:42 -0500 Received: from mx1.redhat.com ([209.132.183.28]:46823 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753390AbcCHJQe (ORCPT ); Tue, 8 Mar 2016 04:16:34 -0500 Message-ID: <1457428591.27353.55.camel@redhat.com> Subject: Re: [PATCH] devpts: Add ptmx_uid and ptmx_gid options From: Alexander Larsson To: Andy Lutomirski , "Eric W. Biederman" Cc: James Bottomley , gnome-os-list@gnome.org, Linux Containers , "linux-kernel@vger.kernel.org" , mclasen@redhat.com, Linux FS Devel Date: Tue, 08 Mar 2016 10:16:31 +0100 In-Reply-To: References: <1427447013.2250.9.camel@HansenPartnership.com> <1427788642.4411.12.camel@redhat.com> <1427807248.2117.117.camel@HansenPartnership.com> <1427808184.2117.122.camel@HansenPartnership.com> <1427810118.2117.126.camel@HansenPartnership.com> <1427810886.2117.129.camel@HansenPartnership.com> <1427811444.4411.20.camel@redhat.com> <1427969525.3559.120.camel@HansenPartnership.com> <1427984969.13651.11.camel@redhat.com> <87zj6qs7v8.fsf@x220.int.ebiederm.org> <87oal4odne.fsf@x220.int.ebiederm.org> <1432832511.21304.6.camel@redhat.com> <87siagh4kh.fsf@x220.int.ebiederm.org> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1578 Lines: 35 On mån, 2016-03-07 at 20:59 -0800, Andy Lutomirski wrote: > On Thu, May 28, 2015 at 12:42 PM, Eric W. Biederman > wrote: > > Andy Lutomirski writes: > >  > Apparently alexl is encountering some annoyances related to the > current workaround, and the workaround is certainly ugly. It works, but it introduces an extra namespace that gets exposed to the world, which is pretty ugly. For instance, entering the namespace becomes hard. I can setns() into the intermediate user+mount namespace without problems, but if i try to setns into the final user+mount ns (it gets its own implicit mount ns) i get EPERM. I'm not sure exactly why though... > Your proposal seems like it could break some use cases involving > fscaps on a mount or mount-like binary. > > What if we change it to use the owner of the userns that owns the > current mount ns?  For anything that doesn't explicitly use > namespaces, this will be zero.  For namespace users, it should do the > right thing. Any of these is fine with me. One nice thing would if i could somehow detect whether this was supported or not so that i can fall back on the old workaround. -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Alexander Larsson Red Hat, Inc alexl@redhat.com alexander.larsson@gmail.com He's an all-American guitar-strumming househusband with no name. She's a scantily clad impetuous former first lady who don't take no shit from nobody. They fight crime!