All of lore.kernel.org
 help / color / mirror / Atom feed
From: Artem Mygaiev <artem_mygaiev@epam.com>
To: Jarvis Roach <Jarvis.Roach@dornerworks.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Praveen Kumar <kpraveen.lkml@gmail.com>
Cc: "Davorin Mista" <davorin.mista@aggios.com>,
	"Lars Kurth" <lars.kurth@citrix.com>,
	"Edgar E. Iglesias" <edgar.iglesias@xilinx.com>,
	"paul_luperto@prqa.com" <paul_luperto@prqa.com>,
	"Volodymyr Babchuk" <volodymyr_babchuk@epam.com>,
	"Rich Persaud" <persaur@gmail.com>,
	"Mirela Simonović" <mirela.simonovic@aggios.com>,
	"Stewart Hildebrand" <Stewart.Hildebrand@dornerworks.com>,
	"julien.grall@arm.com" <julien.grall@arm.com>,
	"robin.randhawa@arm.com" <robin.randhawa@arm.com>,
	"committers@xenproject.org" <committers@xenproject.org>,
	"anastassios.nanos@onapp.com" <anastassios.nanos@onapp.com>,
	xen-devel <xen-devel@lists.xenproject.org>,
	"Jonathan Daugherty" <jtd@galois.com>,
	"Denys Balatsko" <denys.balatsko@globallogic.com>,
	"vfachin@de.adit-jv.com" <vfachin@de.adit-jv.com>
Subject: Re: Xen and safety certification, Minutes of the meeting on Apr 4th
Date: Tue, 22 May 2018 16:08:30 +0300	[thread overview]
Message-ID: <6af21f78-38fb-0f37-01dc-6a152c7f0cc0@epam.com> (raw)
In-Reply-To: <c137e3b91d234a2ca15c8e9c6ac2955f@dornerworks.com>

Hello Jarvis

On 22.05.18 15:08, Jarvis Roach wrote:
>> Hi Stefano
>>
>> On 10.05.18 22:51, Stefano Stabellini wrote:
>>> On Thu, 10 May 2018, Praveen Kumar wrote:
>>>>> Yeah, you are right. It looks like turning Dom0 into a DomU is not
>>>>> good enough. Maybe for this option to be viable we would actually
>>>>> have to terminate (or pause and never unpause?) dom0 after boot.
>>>>
>>>> Just a thought !
>>>> How about keeping Dom0 still be there, but DomUs given Dom0
>>>> privilege, with restricted permission on mission critical resources ?
>>>> And if anyhow Dom0 crashes, the best contended among the existing
>>>> DomUs take the ownership of Dom0 ?
>>>
>>> I don't think this is easily doable, also it wouldn't solve the issue
>>> of removing dom0 from the system. But see below.
>>>
>>>
>>>>>> However, you surely need an entity to handle domain crash. You
>>>>>> don't
>>>> want to
>>>>>> reboot your platform (and therefore you safety critical domain) for
>>>>>> a
>>>> crashed
>>>>>> UI, right? So how this is going to be handled in your option?
>>>>
>>>>> We need to understand the certification requirements better to know
>>>>> the answer to this. I am guessing that UI crashes are not handled
>>>>> from the certification point of view -- maybe we only need to
>>>>> demonstrate that the system is not affected by them?
>>>>
>>>> Where can we find the certification requirements details ?
>>>
>> ISO26262: https://www.iso.org/standard/51362.html
>> IEC61508: https://webstore.iec.ch/publication/5517
>>
>>> Yes, I think we need to understand the requirements better to figure
>>> out the right way forward for Dom0.
>>>
>>> For instance, here is another idea: we could have Xen boot multiple
>>> domains at boot time from device tree, as suggested in the dom0-less
>>> approach. All of the domains booted from Xen are "mission-critical".
>>> The first domain could still be dom0. Once booted, Dom0 can start
>>> other VMs, however, Xen would restrict Dom0 from doing any operations
>>> affecting the first set of mission-critical domains.
>>>
> 
> Does the first domain have to be dom0? Would it be possible to have domains boot in parallel (especially if allocated to separate CPU cores) such that a simple OS (like FreeRTOS) would complete booting before dom0/Linux? In other words, does the hypervisor have any dependencies on dom0 having performed certain functions (interrupt configuration, MMU table initialization, timers, etc.) before it can create and start additional VMs?
> 

We actually have one of the options to run FreeRTOS in dom0 (see earlier 
emails in this thread)

>>> This way, we would get the flexibility of being able to start/stop
>>> domains at run time, but at the same time we might still be able to
>>> avoid certifications for Dom0, because Dom0 cannot affect the mission
>>> critical applications.
> 
>> Such dom0 shall have no mission-critical domains memory access, no HW
>> access (SMMU, DVFS Power, etc.), and so on. EL3 software (optee or similar
>> on ARM) shall also be safety certified and not controlled from dom0
> 
>>>
>>> Is this approach actually feasible? We need to read the requirements
>>> to know. I am hoping Artem will chime in on this :-)
>>   >
>>
>> I think this approach is feasible indeed, if we can prove isolation and fault
>> tolerance for FuSa parts of the system.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

  reply	other threads:[~2018-05-22 13:08 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-04-06 14:13 Xen and safety certification, Minutes of the meeting on Apr 4th Lars Kurth
2018-04-06 17:01 ` Jarvis Roach
2018-04-06 17:23   ` Lars Kurth
2018-04-06 17:32   ` Artem Mygaiev
2018-04-06 20:47     ` Stefano Stabellini
2018-04-11 16:19       ` Artem Mygaiev
2018-04-12 18:38       ` Praveen Kumar
2018-05-08  0:11       ` Stefano Stabellini
2018-05-08 13:39         ` Julien Grall
2018-05-08 15:49           ` Stefano Stabellini
2018-05-10  4:55             ` Praveen Kumar
2018-05-10 19:51               ` Stefano Stabellini
2018-05-12 17:38                 ` Rich Persaud
2018-05-15  8:54                 ` Artem Mygaiev
2018-05-22 12:08                   ` Jarvis Roach
2018-05-22 13:08                     ` Artem Mygaiev [this message]
2018-05-22 17:50                     ` Stefano Stabellini
2018-04-06 17:18 ` Artem Mygaiev

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=6af21f78-38fb-0f37-01dc-6a152c7f0cc0@epam.com \
    --to=artem_mygaiev@epam.com \
    --cc=Jarvis.Roach@dornerworks.com \
    --cc=Stewart.Hildebrand@dornerworks.com \
    --cc=anastassios.nanos@onapp.com \
    --cc=committers@xenproject.org \
    --cc=davorin.mista@aggios.com \
    --cc=denys.balatsko@globallogic.com \
    --cc=edgar.iglesias@xilinx.com \
    --cc=jtd@galois.com \
    --cc=julien.grall@arm.com \
    --cc=kpraveen.lkml@gmail.com \
    --cc=lars.kurth@citrix.com \
    --cc=mirela.simonovic@aggios.com \
    --cc=paul_luperto@prqa.com \
    --cc=persaur@gmail.com \
    --cc=robin.randhawa@arm.com \
    --cc=sstabellini@kernel.org \
    --cc=vfachin@de.adit-jv.com \
    --cc=volodymyr_babchuk@epam.com \
    --cc=xen-devel@lists.xenproject.org \
    /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.