On Mon, Aug 11, 2014 at 5:26 AM, Andersson, Gunnar
<gunnar.x.anders...@volvocars.com> wrote:
> Everyone,
>
> Sorry for reacting to this so slowly.   (Copying genivi-dev for info.)
>
> On 5 Aug 2014, at 07:53, Dominig ar Foll (Intel OTC) 
> <dominig.arf...@fridu.net> wrote:
>
>> Le 05/08/2014 00:58, Rees, Kevron a écrit :
>>> On Mon, Aug 4, 2014 at 3:45 PM, Schaufler, Casey >>
>>> <casey.schauf...@intel.com> wrote:
>>>
>>>> We should start looking at using Cynara for application
>>>> (W3C) privilege enforcement.
>>>>
>>> I need to get with John to talk about whether Cynara would best be
>>> used inside AMB or at the crosswalk-extension level.  It's on my TODO
>>> list.
>>>
>>
>> It shall be done in AMB directly in order to allow non HTML5 Apps to benefit
>> of the security.  This is in particular critical when the socket transport
>> mode is used as there is in that case no other security enforcement point.
>
> I'd just like to request the AMB project and other component developers to
> consider not integrating too tightly with specific technologies.  Having the
> AMB tightly interact with Cynara might (depending on how you do it) 
> effectively
> make it a Tizen-only component.
>
> Will you be able to do it in a generic and replaceable way?
>

I completely agree.  The plan is to make cynara integration a plugin
that can be optionally built and optionally run.

-Kevron

> Dominig - discussing a pattern for adapting "generic" services to Tizen in a
> non-intrusive way would be good, IMHO.  Let me elaborate.
>
> Except from Tizen IVI, there might be no other GENIVI compliant platform that
> uses Cynara, and I doubt that it will be possible to require it as the only
> solution for service access policy enforcement.
>
> In this case I consider AMB to be a potentially generic component. AMB is a
> project that has been followed by GENIVI since its inception and was proposed
> by the Networking EG to be included into GENIVI compliance specification.  (It
> was withdrawn due to some technicalities, but that does not change the point
> for the future; it is still usable in GENIVI systems).
>
> It's probably OK for the "special daemons" (as we called them in the
> GENIVI/Tizen joint F2F; not sure if there is another name). They use a freely
> defined communication over sockets, and are in their code project strongly 
> tied
> to the corresponding Tizen Core API.  Tight interaction with other Tizen-only
> technologies should be fine there because those components are basically
> Tizen-unique anyway.
>
> But as Tizen moves into implementing its security mechanism also for D-Bus and
> integrating other generic services I would propose to also consider non-Tizen
> systems.  If we aim for a modular and generic approach for any generic 
> component
> there is the potential for broader reuse.   In particular the security 
> enforcement
> mechanisms is an area with much variation in different distros, so keeping its
> integration flexible would be good.
>
> Finally, this is probably a non-issue but just in case: for any component that
> you *do* make such changes to I would like to remind everyone to be diligent
> with popular versioning principles (likely best described by 
> http://semver.org/).
> Those are also used in GENIVI.  Specifically, if there is any incompatible API
> change (which I suppose an enforced security policy is), then the major number
> should be increased on the released version.  This will match how GENIVI uses
> version numbers in its specifications.  It will avoid the risk of an already
> released GENIVI spec (which frequently requires a component version N <=
> [used-version] < N+1), to be after-the-fact matching what has later
> become an incompatible component.
>
> Thanks for considering,
> - Gunnar
>
> --
> Gunnar Andersson
> Lead Architect, GENIVI Alliance
> Infotainment, Volvo Car Corporation
>
>>
>> Dominig
>> _______________________________________________
>> IVI mailing list
>> IVI@lists.tizen.org
>> https://lists.tizen.org/listinfo/ivi
>>
_______________________________________________
IVI mailing list
IVI@lists.tizen.org
https://lists.tizen.org/listinfo/ivi

Reply via email to