Hi Jens,
On Thu, Jun 11, 2015 at 01:21:31PM +0200, Jens Wiklander wrote:
It looks like Javier and I agrees here.
I'm also OK with this approach. I did however like the idea with having an opaque command. But I also see the point in using a more strictly defined interface.
One thing that bothers me which I don't think we can do much about is that we're having both ret and ret_origin. For OP-TEE which uses GP specified return values it's hard to do anything about it. But I wish they had set a side a few bits (2 or 3) for origin and the rest could be used for the real return value. However, as I said I don't think there is much to do about it right now. It's too late basically.
I will make a new patch set based on this and send it to this list first to give you a chance to give feedback before I send PATCH V4 to the kernel mailing lists .
-- Regards, Jens