On Tue, Dec 11, 2018 at 12:55:39PM +0000, Pankaj Bansal wrote: > I am not able to understand the reservation about this patch. The reservation is that from an outside perspective, this patch set gives the impression that somewhere deep down the line, a fundamental misunderstanding exists. This impression may be completely incorrect, but from the information presented here, we cannot understand what problem is being solved by this set. This is why we keep coming with endless follow-up questions. > At least in earlier version, the apprehension was that if dtb > supplied by kernel (using command line parameters) > Is supplied to firmware, it may break firmware, as firmware might > not understand the bindings in it. Then I am afraid there has been some miscommunication. I have re-read the thread on the original patch set, and my interpretation on the feedback is completely different. It is referring to creating multiple layers of stability requirements between kernel and firmware when dealing with command-line loaded device trees. The firmware could _never_ be a consumer of a device tree loaded by the kernel stub. Best Regards, Leif
