* Desktop Integration Take 2
@ 2006-02-28 21:01 Ivan Gyurdiev
2006-03-08 19:20 ` Ivan Gyurdiev
0 siblings, 1 reply; 5+ messages in thread
From: Ivan Gyurdiev @ 2006-02-28 21:01 UTC (permalink / raw)
To: SELinux List; +Cc: Daniel J Walsh
Okay, after discussing various obstacles, I think it's time to come up
with a new plan for integrating selinux w/ nautilus.
How about the following:
- Have a file controlled by libsetrans which maps customizable types to
(unicode?) translated strings
- Have a file controlled by libsetrans which maps mls ranges to
(unicode?) translated strings
(setrans.conf)
- Require the mappings above to be 1:1.
- Have two functions (with better names):
get_customizable_types_by_user()
get_mls_labels_by_user()
The first takes a [Unix] user, and returns the translated labels of all
customizable types accessible { relabelto } from that user.
The second takes a [Unix] user, and returns the translated labels of all
mls ranges accessible from that user, [ only the labeled ones ].
===========
Make nautilus link to libsetrans.
Then on startup, nautilus calls both of those functions exactly once.
Then it has an enumeration to work with in set interfaces - have a
drop-down box for the type, and drop-down box for the mls range.
To label a file to a non-customizable context, add a "Restore Context"
button, which runs restorecon.
--
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.
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: Desktop Integration Take 2
2006-02-28 21:01 Desktop Integration Take 2 Ivan Gyurdiev
@ 2006-03-08 19:20 ` Ivan Gyurdiev
2006-03-08 19:59 ` Joe Nall
0 siblings, 1 reply; 5+ messages in thread
From: Ivan Gyurdiev @ 2006-03-08 19:20 UTC (permalink / raw)
Cc: SELinux List, Daniel J Walsh
Ivan Gyurdiev wrote:
> Okay, after discussing various obstacles, I think it's time to come up
> with a new plan for integrating selinux w/ nautilus.
> How about the following:
>
> - Have a file controlled by libsetrans which maps customizable types
> to (unicode?) translated strings
> - Have a file controlled by libsetrans which maps mls ranges to
> (unicode?) translated strings
> (setrans.conf)
>
> - Require the mappings above to be 1:1.
>
> - Have two functions (with better names):
> get_customizable_types_by_user()
> get_mls_labels_by_user()
>
> The first takes a [Unix] user, and returns the translated labels of
> all customizable types accessible { relabelto } from that user.
> The second takes a [Unix] user, and returns the translated labels of
> all mls ranges accessible from that user, [ only the labeled ones ].
>
> ===========
> Make nautilus link to libsetrans.
>
> Then on startup, nautilus calls both of those functions exactly once.
> Then it has an enumeration to work with in set interfaces - have a
> drop-down box for the type, and drop-down box for the mls range.
> To label a file to a non-customizable context, add a "Restore Context"
> button, which runs restorecon.
>
Any feedback on this plan? I can write patches against libsetrans if
there's agreement. Also, I think we should move libsetrans into this
planned future repository that I hear about for selinux components.
--
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.
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: Desktop Integration Take 2
2006-03-08 19:20 ` Ivan Gyurdiev
@ 2006-03-08 19:59 ` Joe Nall
2006-03-08 21:52 ` Ivan Gyurdiev
0 siblings, 1 reply; 5+ messages in thread
From: Joe Nall @ 2006-03-08 19:59 UTC (permalink / raw)
To: Ivan Gyurdiev; +Cc: SELinux List, Daniel J Walsh
On Mar 8, 2006, at 1:20 PM, Ivan Gyurdiev wrote:
>> - Have a file controlled by libsetrans which maps mls ranges to
>> (unicode?) translated strings
>> (setrans.conf)
>>
>> - Require the mappings above to be 1:1.
...
The work I'm doing on better MLS inverse bit handing (rel country1/
country2 ...) in libsetrans supports mapping label fragments
(individual categories and category ranges) in addition to the
current 1:1 mapping. With 256 bits, 1:1 MLS mappings are not
practical because the number of potential combinations is so great. I
hope to have a patch for review at by the end of next week.
In our CMW work, we have a GUI label builder that understands the MLS
combination constraints and a 'last 40 labels built' menu. Individual
users rarely go beyond the 40 labels. An initial manual label
constructor that saves state in a preferences file would be an
adequate start.
joe
--
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.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Desktop Integration Take 2
2006-03-08 19:59 ` Joe Nall
@ 2006-03-08 21:52 ` Ivan Gyurdiev
2006-03-08 22:50 ` Joe Nall
0 siblings, 1 reply; 5+ messages in thread
From: Ivan Gyurdiev @ 2006-03-08 21:52 UTC (permalink / raw)
To: Joe Nall; +Cc: SELinux List, Daniel J Walsh
>>> - Have a file controlled by libsetrans which maps mls ranges to
>>> (unicode?) translated strings
>>> (setrans.conf)
>>>
>>> - Require the mappings above to be 1:1.
> ...
>
> The work I'm doing on better MLS inverse bit handing (rel
> country1/country2 ...) in libsetrans supports mapping label fragments
> (individual categories and category ranges) in addition to the current
> 1:1 mapping.
It was my understanding that the range should be treated as a whole for
maximum generality, and we shouldn't unroll and translate individual
categories [ at least for the purposes of nautilus ]. That way you could
have a coded clearance label like: NSASecretProject, that unrolls to the
actual categories [ say NDA_1, NDA_3, Programmer, NSA, Security,
Confidential, ProjectX ], without listing them all individually.
> With 256 bits, 1:1 MLS mappings are not practical because the number
> of potential combinations is so great. I hope to have a patch for
> review at by the end of next week.
My comment said they should be 1:1, not onto.
(i.e. I don't expect the setrans file to define a translation for each
and every bit combination. In fact, I though I could ignore the
unspecified ones, and not provide access to them via the Nautilus GUI.
They'd still be available for use, but they wouldn't be exposed via
nautilus if they're not in the setrans file).
--
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.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Desktop Integration Take 2
2006-03-08 21:52 ` Ivan Gyurdiev
@ 2006-03-08 22:50 ` Joe Nall
0 siblings, 0 replies; 5+ messages in thread
From: Joe Nall @ 2006-03-08 22:50 UTC (permalink / raw)
To: Ivan Gyurdiev; +Cc: SELinux List, Daniel J Walsh
On Mar 8, 2006, at 3:52 PM, Ivan Gyurdiev wrote:
>
>>>> - Have a file controlled by libsetrans which maps mls ranges to
>>>> (unicode?) translated strings
>>>> (setrans.conf)
>>>>
>>>> - Require the mappings above to be 1:1.
>> ...
>>
>> The work I'm doing on better MLS inverse bit handing (rel country1/
>> country2 ...) in libsetrans supports mapping label fragments
>> (individual categories and category ranges) in addition to the
>> current 1:1 mapping.
> It was my understanding that the range should be treated as a whole
> for maximum generality, and we shouldn't unroll and translate
> individual categories [ at least for the purposes of nautilus ].
> That way you could have a coded clearance label like:
> NSASecretProject, that unrolls to the actual categories [ say
> NDA_1, NDA_3, Programmer, NSA, Security, Confidential, ProjectX ],
> without listing them all individually.
This doesn't scale in some (not all) environments. See below.
>> With 256 bits, 1:1 MLS mappings are not practical because the
>> number of potential combinations is so great. I hope to have a
>> patch for review at by the end of next week.
> My comment said they should be 1:1, not onto.
> (i.e. I don't expect the setrans file to define a translation for
> each and every bit combination. In fact, I though I could ignore
> the unspecified ones, and not provide access to them via the
> Nautilus GUI. They'd still be available for use, but they wouldn't
> be exposed via nautilus if they're not in the setrans file).
I agree that you can ignore the categories with unspecified mappings.
I want to define a translation for 238 countries and combinations
thereof (billions and billions to quote Carl Sagan) in the
setrans.conf file. In order to keep the size finite, I have to have
mappings that combine in addition to the current one name to many
categories mappings. I'll send you an image of what our current label
chooser looks like off list. If the selection mechanism supports a
plugin, we can replace a simple picker with something more complex
when warranted.
joe
--
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.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2006-03-08 22:50 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-02-28 21:01 Desktop Integration Take 2 Ivan Gyurdiev
2006-03-08 19:20 ` Ivan Gyurdiev
2006-03-08 19:59 ` Joe Nall
2006-03-08 21:52 ` Ivan Gyurdiev
2006-03-08 22:50 ` Joe Nall
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.