On 2/27/10 2:57 PM, Peter Stuge wrote:
Um..? Oh, ok, they aren't all bootsplash types. But, I think there should be just one meaningful parameter for each file. My point is that a bootsplash type is redundant, since a specific filename must already be used for bootsplashes. The same is true for VSA.
Since we only do name based matching in coreboot anyways, do you suggest we drop the type field?
I figured, the names might be changed or varied in the future, while the type is an additional consistency check. Also it makes sense to specify this so cbfstool might be able to use reasonable defaults for compression or placement of the files in the future. (Like, bootsplash jpg files should never be lzma compressed)
What is MBI?
Some Intel add-on for SMI/oprom communication.
How could this be handled better? Maybe we should change things around so that the filename is insignificant, and make the type more important? That makes user error much less obvious however, the filename concept is already well known by everyone.
Also, I think it should be possible to add files to the cbfs that we don't have to know how to parse. Like encryption public keys, or license files for oproms, or, ...
Do we need so specific types at all?
Well we have them, and I think it's better to use them unless/until we decide to drop them. Payloads may or may not make use of them.
I even forgot a type: cpu microcode.