On 25.08.23 06:52, Williams, Hannah wrote:
I think your suggestion to use similar interface to libgfxinit (gma_gfxinit(), gma_gfxstop()) is a good idea. Please add your comments to the CLs and we will work on it.
Thanks. I'll wait though, until the decisions about open-sourcing and integrating uGOP are made.
Regarding the MP PPI and the Kconfig cleanup that you mentioned, we will take this up. I am going to open a internal ticket to follow up on this.
Thanks again! However the MP PPI was just one example. It would be good to have up-to-date information and clean integration for all features of Intel's blobs. Information, I think, is the more impor- tant part. Staying with the MP PPI example: IIRC, when it was intro- duced there was a list of advanced features added to our documen- tation, arguing that the features would only be made possible by the MP PPI integration. But I'm not sure if any of these features ever surfaced in coreboot. So maybe this is something to add to your ticket: Do we actually need the MP PPI in coreboot?
Also, on your comment on implementing libgfxinit for MTL, this will not speed up the process for us because we do not have resources to work on it
Last time I checked Intel was still a multi-billion dollar company. I think this translates to: SOL is not a high enough priority for Intel.
and this is why we went the faster route to re-use what we already have (GOP) and modified it as uGOP for SOL use case.
If you only go one route, we'll never have a definite answer which is the faster one.
Regards, Nico