powerpc: Always trim numa memory to lmb_end_of_DRAM()

numa_enforce_memory_limit tried to be smart and only call lmb_end_of_DRAM
when a memory limit was set via mem= on the command line.  However,
the early boot code will also limit memory added to the lmb system
when iommu=off is specified.  When this happens, the page allocator
is given pages not in the linear mapping and this results in a fatal
data reference to the unmapped page.
Signed-off-by: NMilton Miller <miltonm@bga.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

arch/powerpc/mm/numa.c

@@ -528,12 +528,10 @@ static unsigned long __init numa_enforce_memory_limit(unsigned long start,
 	/*
 	 * We use lmb_end_of_DRAM() in here instead of memory_limit because
 	 * we've already adjusted it for the limit and it takes care of
-	 * having memory holes below the limit.
+	 * having memory holes below the limit.  Also, in the case of
+	 * iommu_is_off, memory_limit is not set but is implicitly enforced.
 	 */
 
-	if (! memory_limit)
-		return size;
-
 	if (start + size <= lmb_end_of_DRAM())
 		return size;