> I was surprised by Theo's answer because I recall BeagleBone Black was > open hardware, at least as a design.
The word "Open" means nothing in this instance. Most hardware + firmware combinations provide insufficient detail to know what pins are used for what, reserved for what, or wired to an auto-destruct. Even on PCs, we have encountered situations where critical clock controller chips are at placed at ISA addresses on one machine, but wired to a fan controller / temperature sensor on another. > But I won't bet my hand on the board's ARM cpu. It has nothing to do with the cpu. It is about what is around the cpu, and how these platforms lack a way to tell us what is what. And even if they dcreated > Anyway, I am not using it, I consider it for fun, but I checked it > again and here it is: > > Excerpt from the hardware support: > > "BeagleBone Black ships with two virtual capes already on it, one for > the on-board eMMC storage and one for the HDMI output. When configured > for use these virtual capes consume actual resources. > > If the eMMC is not placed in reset, the MMC1* signals may not be used > without potentially corrupting the contents of your on-board > eMMC---and possibly damaging the physical circuit as well." > > The pins used by MMC1* are shown in the picture: 3, 4, 5, 6, 21, 22, 23, 24, > 25. What it says in a book is not relevant. The kernel has to be told somehow. This "enumeration" problem is why PCIBIOS showed up, it is why ACPI showed up, it is why openfirware showed up, etc. Something has to tell you what is what, in a portable fashion for when the platforms change. Which THEY DO. Every day.
