Re: [Xen-devel] [PATCH v1 2/6] arm: add generic TEE mediator framework

[Julien Grall <julien.grall@xxxxxxx>]

On 27/08/18 20:09, Volodymyr Babchuk wrote:
> Hi Julien,
Hi,

> On 22.08.18 20:03, Julien Grall wrote:
>
> [...]
>
> > >       if ( is_hardware_domain(d) && (rc = domain_vuart_init(d)) )
> > >           goto fail;
> > > +    /* Notify TEE that new domain was created */
> > > +    tee_domain_create(d);
> > My concern about domain creation is still not addressed. I would 
> > expect the toolstack to decide whether TEE should be initialized for a 
> > given guest and potentially return an error on failure (e.g maximum 
> > client ID has been reached).
> > But very likely, you don't need to initialize TEE that early. This 
> > could be done in a separate DOMCTL as we did for VPL011.
> Yes, as we discussed in latter patches, I'll add DOMCTL support.But 
> what to do with dom0 construction?
> I think, it should be configurable. But how? With commandline option?
There are an high chance that OP-TEE will be required by Dom0 for some 
devices. So I think it is fine to have OP-TEE enabled by default for Dom0.
If you want to override the behavior, then a command-line option looks 
the best.
> > > +    printk(XENLOG_WARNING "No TEE mediator found\n");
> > Not having a TEE is a valid use case. So printing a warning seems a 
> > bit too much.
> I can change this to INFO. Or it is better to remove this print at all?
I would prefer the message to be removed. The message "Using TEE 
mediator..." is sufficient as if it is not printed it means there are no 
TEE.
Cheers,

--
Julien Grall

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