On Tue, 12 Feb 2019, 19:21 Andrii Anisov, <andrii.anisov@gmail.com> wrote:

> Hello Julien,
>

Hi,

Sorry for the formatting.


> On 07.02.19 12:59, Julien Grall wrote:
> > In that case I would prefer if we don't keep the runstate mapped.
>
> Actually I'm going to see runstate update impact on the context switch
> time. For that I will extend TBM with runstate setup.
> I really do not like a bunch of `copy_to_guest()` done on each context
> switch because of runstate.


Please provide more meaningful arguments other than "I don't like it". I
provided potential drawbacks on my previous e-mails that you haven't yet
addressed.

FWIW, Volodymyr had the same argument on OP-TEE and I requested to avoid
the global mapping. I haven't seen any concern for performance devredation
afterwards. Please feel free to come with numbers here.

If you are against runstate mapping, I'd like to fit runstate into a page,
> than access it directly from the hypervisor.


This is not really a policy in Xen. If the guest cares about it, it can
ensure that the runstate does not cross a page boundary.

Cheers,


> --
> Sincerely,
> Andrii Anisov.
>