Re: [PATCH 00/18] RFC: Merge IS_PRIV checks into XSM hooks

[Daniel De Graaf <dgdegra@tycho.nsa.gov>]

On 08/07/2012 02:07 PM, Shakeel Butt wrote:
>> I wasn't intending to exclude the other uses of XSM that this series will
>> benefit; dom0 disaggregation is just the most obvious case that requires
>> the larger changes like removing IS_PRIV checks.
> I was just saying that this patch series is more beneficial than claimed.
> 
>> Xenstore can already be split into its own stub domain (or domains, as in
>> the Xoar paper). The permissions model in Xenstore has a privileged bit
>> similar to IS_PRIV; extending XSM controls into Xenstore similar to how
>> SELinux controls were extended into DBus will address this.
> 
> My real concern here was the use of is_initial_domain() in the xenbus driver
> code. For example I am running all Linux PV and one of them is XenStore
> domain, the xenbus driver needs to do something different than
> is_initial_domain(),
> maybe something like is_xenstore_domain() [not saying this is right
> way to do it].
> Please correct me if I am wrong.
> 
> thanks,
> Shakeel
> 

The method in upstream Linux is more complete than this: if the domain
is started with xenstore information in the shared page, it will use it
(which happens when a domain builder is used to launch dom0 and xenstore
stub domains at the same time); otherwise, there is an ioctl that can
be used in dom0 to tell it about a newly launched xenstore stub domain.
The combination of these eliminates any need for an is_xenstore_domain()
function.

-- 
Daniel De Graaf
National Security Agency