On Wed, Jul 25, 2018 at 02:29:09PM +0200, Ard Biesheuvel wrote: > To prevent spending a disproportionate amount of time on inventing > ways to parameterize these 'models', which includes interfaces not > only between UEFI and QEMU but also between ARM-TF and QEMU (and > potentially between UEFI and ARM-TF as well), could we please consider
I guess a UEFI <=> ARM-TF interface will also need to work on real hardware, and therefore need a QEMU model. If whatever that interface needs can also satisfy the UEFI <=> QEMU and ARM-TF <=> interfaces, then that's all we need - no need for fw-cfg. > fixed configurations for the non-discoverable bits? If there is a new Well, at least one thing must to be fixed, which is a) address/register where the address of the DTB will be stored b) address of the QEMU interface device, e.g. fw-cfg. fw-cfg or whatever can then be used to get the DTB. And I guess everything the SBSA says should be fixed should be fixed. Otherwise, what's non-discoverable? > and improved sbsa model in the future, we could call it sbsa-2.0, no? Of course the model can be versioned, allowing for any change, but I don't think model versions were meant to be aliases for peripheral selections. That's what scripts wrapped around qemu are for. Thanks, drew