memblock: Fix alloc failure due to dumb underflow protection in membl…

…ock_find_in_range_node()

7bd0b0f ("memblock: Reimplement memblock allocation using
reverse free area iterator") implemented a simple top-down
allocator using a reverse memblock iterator.  To avoid underflow
in the allocator loop, it simply raised the lower boundary to
the requested size under the assumption that requested size
would be far smaller than available memblocks.

This causes early page table allocation failure under certain
configurations in Xen.  Fix it by checking for underflow directly
instead of bumping up lower bound.

Signed-off-by: Tejun Heo <[email protected]>
Reported-by: Konrad Rzeszutek Wilk <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: Benjamin Herrenschmidt <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Linus Torvalds <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>

mm/memblock.c

@@ -106,14 +106,17 @@ phys_addr_t __init_memblock memblock_find_in_range_node(phys_addr_t start,
 	if (end == MEMBLOCK_ALLOC_ACCESSIBLE)
 		end = memblock.current_limit;
 
-	/* adjust @start to avoid underflow and allocating the first page */
-	start = max3(start, size, (phys_addr_t)PAGE_SIZE);
+	/* avoid allocating the first page */
+	start = max_t(phys_addr_t, start, PAGE_SIZE);
 	end = max(start, end);
 
 	for_each_free_mem_range_reverse(i, nid, &this_start, &this_end, NULL) {
 		this_start = clamp(this_start, start, end);
 		this_end = clamp(this_end, start, end);
 
+		if (this_end < size)
+			continue;
+
 		cand = round_down(this_end - size, align);
 		if (cand >= this_start)
 			return cand;