Next up in Linux's initialization sequence might be setting the number of CPU IDs to the index of the last-set-bit in the possible CPUs bitmask, via a couple accessors & a CPU-specific routine which might be built into the compiler. Following that it allocates memory for storing data to be used by a single CPU core. This might involve rounding up some properties, invoking a helper to memblock-allocate a computed amount of memory populating an array there, memblock-allocates some more, ... 1/2