I still would like to have the complete sam926x patch set implemented before we start to "play" with it
Yeah, but I still think some preparatory patches are a good thing in order to make things easier later on.
I only see two things as a result of those prepatory patches.
1) Bugs 2) Delays in availability
I am not questioning anything you want to do, just the timing. If we follow your recommendation you will have a bunch of untested board patches in the mainstream u-boot, for the sam926x.
I do not have the time to thoroughly test any changes you do (I dont even have an at91sam9263 board).
If we go my way, then we should be able to have *tested* sam926x support inside U-boot very soon, and while this results in duplication of a small part of the spi code on the source level (no addition to the binary) I believe that the benefit to the community of at91sam926x users of having native support in U-Boot outweighs this duplication a lot.
The number of users far outweisghs the number of implementers and I think that we need to look at it from their point of view.
We are not introducing any new interfaces here, just using the existing interfaces. This means that it will not be harder to do any modifications *after* the sam926x patches are applied. It will be easier since you have better overview.
The only alternative I can see is that you take over the responsibility for both AVR32 and AT91 U-boot and test all modifications on all boards before submission - Still with availability of working solution as the highest priority.
Haavard
Best Regards Ulf Samuelsson