mm: page_alloc: skip over regions of invalid pfns where possible

When using a sparse memory model memmap_init_zone() when invoked with
the MEMMAP_EARLY context will skip over pages which aren't valid - ie.
which aren't in a populated region of the sparse memory map.  However if
the memory map is extremely sparse then it can spend a long time
linearly checking each PFN in a large non-populated region of the memory
map & skipping it in turn.

When CONFIG_HAVE_MEMBLOCK_NODE_MAP is enabled, we have sufficient
information to quickly discover the next valid PFN given an invalid one
by searching through the list of memory regions & skipping forwards to
the first PFN covered by the memory region to the right of the
non-populated region.  Implement this in order to speed up
memmap_init_zone() for systems with extremely sparse memory maps.

James said "I have tested this patch on a virtual model of a Samurai CPU
with a sparse memory map.  The kernel boot time drops from 109 to
62 seconds. "

Link: http://lkml.kernel.org/r/20161125185518.29885-1-paul.burton@imgtec.com
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Tested-by: James Hartley <james.hartley@imgtec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

3 files changed, 36 insertions, 1 deletions

diff --git a/include/linux/memblock.h b/include/linux/memblock.h
index 5b759c9acf97..38bcf00cbed3 100644
--- a/include/linux/memblock.h
+++ b/include/linux/memblock.h
@@ -203,6 +203,7 @@ int memblock_search_pfn_nid(unsigned long pfn, unsigned long *start_pfn,
 			    unsigned long  *end_pfn);
 void __next_mem_pfn_range(int *idx, int nid, unsigned long *out_start_pfn,
 			  unsigned long *out_end_pfn, int *out_nid);
+unsigned long memblock_next_valid_pfn(unsigned long pfn, unsigned long max_pfn);
 
 /**
  * for_each_mem_pfn_range - early memory pfn range iterator
diff --git a/mm/memblock.c b/mm/memblock.c
index 7608bc305936..a476d28e0733 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -1105,6 +1105,31 @@ void __init_memblock __next_mem_pfn_range(int *idx, int nid,
 		*out_nid = r->nid;
 }
 
+unsigned long __init_memblock memblock_next_valid_pfn(unsigned long pfn,
+						      unsigned long max_pfn)
+{
+	struct memblock_type *type = &memblock.memory;
+	unsigned int right = type->cnt;
+	unsigned int mid, left = 0;
+	phys_addr_t addr = PFN_PHYS(pfn + 1);
+
+	do {
+		mid = (right + left) / 2;
+
+		if (addr < type->regions[mid].base)
+			right = mid;
+		else if (addr >= (type->regions[mid].base +
+				  type->regions[mid].size))
+			left = mid + 1;
+		else {
+			/* addr is within the region, so pfn + 1 is valid */
+			return min(pfn + 1, max_pfn);
+		}
+	} while (left < right);
+
+	return min(PHYS_PFN(type->regions[right].base), max_pfn);
+}
+
 /**
  * memblock_set_node - set node ID on memblock regions
  * @base: base of area to set node ID for
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 05c0a59323bd..6da3169d3750 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -5103,8 +5103,17 @@ void __meminit memmap_init_zone(unsigned long size, int nid, unsigned long zone,
 		if (context != MEMMAP_EARLY)
 			goto not_early;
 
-		if (!early_pfn_valid(pfn))
+		if (!early_pfn_valid(pfn)) {
+#ifdef CONFIG_HAVE_MEMBLOCK_NODE_MAP
+			/*
+			 * Skip to the pfn preceding the next valid one (or
+			 * end_pfn), such that we hit a valid pfn (or end_pfn)
+			 * on our next iteration of the loop.
+			 */
+			pfn = memblock_next_valid_pfn(pfn, end_pfn) - 1;
+#endif
 			continue;
+		}
 		if (!early_pfn_in_nid(pfn, nid))
 			continue;
 		if (!update_defer_init(pgdat, pfn, end_pfn, &nr_initialised))