From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from out02.mta.xmission.com ([166.70.13.232]:47868 "EHLO out02.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751074AbeECVgv (ORCPT ); Thu, 3 May 2018 17:36:51 -0400 From: ebiederm@xmission.com (Eric W. Biederman) To: Mimi Zohar Cc: Casey Schaufler , David Howells , Matthew Garrett , linux-integrity@vger.kernel.org, linux-security-module@vger.kernel.org, kexec@lists.infradead.org, linux-kernel@vger.kernel.org References: <1523572911-16363-1-git-send-email-zohar@linux.vnet.ibm.com> <1523572911-16363-3-git-send-email-zohar@linux.vnet.ibm.com> <87h8nqglpx.fsf@xmission.com> <1525275904.5669.308.camel@linux.vnet.ibm.com> <87h8nospo5.fsf@xmission.com> <6203b1e4-70c3-6d0e-60e0-56c6e8f72ec9@schaufler-ca.com> <87y3h0pu72.fsf@xmission.com> <1525381619.3539.45.camel@linux.vnet.ibm.com> Date: Thu, 03 May 2018 16:36:40 -0500 In-Reply-To: <1525381619.3539.45.camel@linux.vnet.ibm.com> (Mimi Zohar's message of "Thu, 03 May 2018 17:06:59 -0400") Message-ID: <87lgd0o1zr.fsf@xmission.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Subject: Re: [PATCH 2/3] kexec: call LSM hook for kexec_load syscall Sender: linux-integrity-owner@vger.kernel.org List-ID: Mimi Zohar writes: > On Thu, 2018-05-03 at 11:42 -0500, Eric W. Biederman wrote: >> Casey Schaufler writes: >> >> > On 5/3/2018 8:51 AM, Eric W. Biederman wrote: >> >> Mimi Zohar writes: >> >> >> >>> On Wed, 2018-05-02 at 09:45 -0500, Eric W. Biederman wrote: >> >>>> Mimi Zohar writes: >> >>>> >> >>>>> Allow LSMs and IMA to differentiate between the kexec_load and >> >>>>> kexec_file_load syscalls by adding an "unnecessary" call to >> >>>>> security_kernel_read_file() in kexec_load. This would be similar to the >> >>>>> existing init_module syscall calling security_kernel_read_file(). >> >>>> Given the reasonable desire to load a policy that ensures everything >> >>>> has a signature I don't have fundamental objections. >> >>>> >> >>>> security_kernel_read_file as a hook seems an odd choice. At the very >> >>>> least it has a bad name because there is no file reading going on here. >> >>>> >> >>>> I am concerned that I don't see CONFIG_KEXEC_VERIFY_SIG being tested >> >>>> anywhere. Which means I could have a kernel compiled without that and I >> >>>> would be allowed to use kexec_file_load without signature checking. >> >>>> While kexec_load would be denied. >> >>>> >> >>>> Am I missing something here? >> >>> The kexec_file_load() calls kernel_read_file_from_fd(), which in turn >> >>> calls security_kernel_read_file(). So kexec_file_load and kexec_load >> >>> syscall would be using the same method for enforcing signature >> >>> verification. >> >> Having looked at your patches and the kernel a little more I think >> >> this should be a separate security hook that does not take a file >> >> parameter. >> >> >> >> Right now every other security module assumes !file is init_module. >> >> So I think this change has the potential to confuse other security >> >> modules, with the result of unintended policy being applied. >> >> >> >> So just for good security module hygeine I think this needs a dedicated >> >> kexec_load security hook. >> > >> > I would rather see the existing modules updated than a new >> > hook added. Too many hooks spoil the broth. Two hooks with >> > trivial differences just add to the clutter and make it harder >> > for non-lsm developers to figure out what to use in their >> > code. >> >> These are not non-trivial differences. There is absolutely nothing >> file related about kexec_load. Nor for init_module for that matter. >> >> If something is called security_kernel_read_file I think it is wholly >> appropriate for code that processes such a hook to assume file is >> non-NULL. >> >> When you have to dance a jig (which is what I see the security modules >> doing) to figure out who is calling a lsm hook for what purpose I think >> it is a maintenance problem waiting to happen and that the hook is badly >> designed. >> >> At this point I don't care what the lsm's do with the hooks but the >> hooks need to make sense for people outside of the lsm's and something >> about reading a file in a syscall that doesn't read files is complete >> and utter nonsense. > > Sure, we can define a wrapper around the security_kernel_read_file() > hook, calling it security_non-fd_syscall() or even > security_old_syscall(). I really don't see why you want to use the same hook. I just read through the code of all three users. None of them. Especially IMA shares any significant code between the !file case and the file case. Eric