Am Dienstag, den 12.03.2013, 16:03 -0700 schrieb ron minnich:
I think what is going on is just fine. I'm delighted to see the haswell stuff coming out so early. It's quite amazing to see such a short time between chipsets being announced and code appearing in coreboot.
I find that great too.
If we go back to the origin, that's how it always was with boards. Bringup happened in a very exposed way, with partial commits.
That is no argument to *not* trying to improve it.
This Is Not A Kernel. By doing release the way we do, you get visibility very early in the process.
Looking add the Haswell drm/i915 patches, these were contributed early on too. So I do not see the contradiction.
Let's not get too tied up in process details. Let's get on with the work.
Yes and part of that work is carefully reviewing the patches and getting them merged. This should be done in such a way that it is as efficient as possible to optimize time usage.
Theoretically, a careful review of the original Haswell patch would comment each point the next fix-up patches address. Or the commit message should point out each known TODO(?).
Adding new/correct PCI IDs should be squashed in my opinion.
Thanks,
Paul