From: Daniel J Walsh <dwalsh@redhat.com>
To: Russell Coker <rcoker@redhat.com>
Cc: Colin Walters <walters@redhat.com>,
Stephen Smalley <sds@epoch.ncsc.mil>,
SE Linux list <selinux@tycho.nsa.gov>,
Joshua Brindle <jbrindle@snu.edu>,
Jim Carter <jwcart2@epoch.ncsc.mil>,
Nalin Dahyabhai <nalin@redhat.com>
Subject: Re: Single home directory type for all roles.
Date: Thu, 09 Dec 2004 15:22:33 -0500 [thread overview]
Message-ID: <41B8B409.4070807@redhat.com> (raw)
In-Reply-To: <1102623195.4509.86.camel@aeon>
Russell Coker wrote:
>On Thu, 2004-12-09 at 14:45 -0500, Daniel J Walsh wrote:
>
>
>>>A bug IMHO. If we have two roles that become almost equivalent then the
>>>sensible thing to do is remove one. If we decide that for Fedora strict
>>>policy we don't want to have any regular users be denied the ability
>>>perform administrative tasks then the correct thing to do is to make
>>>staff_r the default user role.
>>>
>>>
>>>
>>I want to go back to the separation between user and staff without the
>>differences in file system.
>>
>>
>
>That's impossible. If you can write to someone's .bashrc file or
>similar then you can get their privs.
>
>
So we either go to full MAC which few people will use, or we have to
rely partially on DAC for protection.
I am looking for a way to raise the bar on policy without making a multi
use machine unmanageable.
BTW I like the idea of defaulting to staff_r. I think we should do that
and turn off user_canbe_sysadm.
>
>
>>>It's easy enough for anyone to add a new role if they need more roles
>>>than the default policy provides.
>>>
>>>
>>>
>>Not without relabing the file system.
>>
>>
>
>The expected practice should be to create the role before creating the
>user who will have it. This means that there should not be a need to
>relabel. There is only a need to relabel if you change the roles that
>are permitted after the machine has been running. But that also means
>you may have to have the user logout first to prevent processes becoming
>unlabeled.
>
>
>
Yes that is all well in good, but it does not work that way in the real
world. Admins
are going to want to be able to change peoples roles.
>> Currently if I want to add a new
>>role, say student that has less privs
>>then user, I need to massively rewrite the policy. If we came up with
>>a policy that shared homedir and tmpdir
>>file contexts between all types of users, I could begin to create
>>additional default roles for people.
>>
>>
>
>For what benefit? If they share file types and they share X access
>(with xdm logins) then what benefits can we gain from multiple roles?
>
>Multiple roles will still increase administrative overhead even without
>multiple file types. So multiple roles with the same types gives you
>some of the overhead with almost none of the benefit.
>
>
>
Roles can be used to govern which applications can be run. So I could
have a student role where only
student applications could be run.
--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.
next prev parent reply other threads:[~2004-12-09 20:22 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-12-07 0:08 patch: add can_create() macro, allow file_type_auto_trans(a,b,c, { file dir }) Thomas Bleher
2004-12-08 19:32 ` James Carter
2004-12-09 16:31 ` Some more fixes Daniel J Walsh
2004-12-09 18:35 ` Thomas Bleher
2004-12-10 20:14 ` James Carter
2004-12-09 16:50 ` Single home directory type for all roles Daniel J Walsh
2004-12-09 17:20 ` Stephen Smalley
2004-12-09 17:40 ` Stephen Smalley
2004-12-10 16:23 ` Manipulating user roles without policy-sources installed Daniel J Walsh
2004-12-10 16:37 ` Stephen Smalley
2004-12-10 18:09 ` Daniel J Walsh
2004-12-10 18:38 ` Stephen Smalley
2004-12-09 17:47 ` Single home directory type for all roles Russell Coker
2004-12-09 17:53 ` Stephen Smalley
2004-12-09 18:12 ` Russell Coker
2004-12-09 18:18 ` Stephen Smalley
2004-12-09 18:45 ` Stephen Smalley
2004-12-09 19:08 ` Russell Coker
2004-12-09 20:03 ` Casey Schaufler
2004-12-10 12:20 ` Russell Coker
2004-12-10 15:22 ` Valdis.Kletnieks
2004-12-10 16:19 ` Casey Schaufler
2004-12-10 17:00 ` Valdis.Kletnieks
2004-12-10 17:06 ` Stephen Smalley
2004-12-10 17:29 ` Casey Schaufler
2004-12-09 20:40 ` Valdis.Kletnieks
2004-12-10 3:03 ` Russell Coker
2004-12-10 14:09 ` Daniel J Walsh
2004-12-10 14:31 ` Stephen Smalley
2004-12-10 15:43 ` Colin Walters
2004-12-10 16:33 ` Casey Schaufler
2004-12-13 13:25 ` Russell Coker
2004-12-13 13:56 ` Daniel J Walsh
2004-12-13 14:19 ` Russell Coker
2004-12-09 19:07 ` Thomas Bleher
2004-12-09 19:19 ` Russell Coker
2004-12-09 17:28 ` Colin Walters
2004-12-09 18:02 ` Russell Coker
2004-12-09 19:45 ` Daniel J Walsh
2004-12-09 20:07 ` Stephen Smalley
2004-12-09 20:13 ` Russell Coker
2004-12-09 20:22 ` Daniel J Walsh [this message]
2004-12-09 20:30 ` Russell Coker
2004-12-09 21:38 ` Thomas Bleher
2004-12-10 2:56 ` Russell Coker
2004-12-09 22:29 ` Colin Walters
2004-12-10 13:11 ` Stephen Smalley
2004-12-10 16:28 ` Colin Walters
2004-12-09 21:16 ` Thomas Bleher
2004-12-10 2:58 ` Russell Coker
2004-12-09 22:43 ` Colin Walters
2004-12-10 2:23 ` Russell Coker
2004-12-10 15:48 ` Colin Walters
2004-12-10 21:58 ` Luke Kenneth Casson Leighton
2004-12-09 19:38 ` Daniel J Walsh
2004-12-09 19:58 ` Stephen Smalley
2004-12-09 20:09 ` Daniel J Walsh
2004-12-09 20:17 ` Russell Coker
2004-12-09 20:38 ` Daniel J Walsh
-- strict thread matches above, loose matches on Subject: below --
2004-12-09 18:50 Alex Ackerman
2004-12-09 19:29 ` Russell Coker
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=41B8B409.4070807@redhat.com \
--to=dwalsh@redhat.com \
--cc=jbrindle@snu.edu \
--cc=jwcart2@epoch.ncsc.mil \
--cc=nalin@redhat.com \
--cc=rcoker@redhat.com \
--cc=sds@epoch.ncsc.mil \
--cc=selinux@tycho.nsa.gov \
--cc=walters@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.