sparc64: Fix array size reported by vmemmap_populate()

vmemmap_populate() attempts to report the used index and total size of vmemmap_table, but it wrongly shifts the total size so that it is always shown as 0.

---
 arch/sparc/mm/init_64.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

index aaebc48..b2831dc 100644
@@ -2117,7 +2117,7 @@ int __meminit vmemmap_populate(struct page *start, unsigned long nr, int node)
 			       "node=%d entry=%lu/%lu\n", start, block, nr,
 			       node,
 			       addr >> VMEMMAP_CHUNK_SHIFT,
-			       VMEMMAP_SIZE >> VMEMMAP_CHUNK_SHIFT);
+			       VMEMMAP_SIZE);
 		}
 	}
 	return 0;

1.7.0.3

From: Ben Hutchings <[email protected]>
Date: Sat, 03 Apr 2010 21:41:20 +0100

It's not the total, size, it's the total number of chunks that the area supports.

So it's trying to say:

	entry=${NUM_CHUNKS}/${MAX_CHUNKS}

So I think the printk statement is correct and if you check you'll find that every other vmemmap supporting platform has this same printk statement.

From: David Miller <[email protected]>
Date: Sat, 03 Apr 2010 13:50:29 -0700 (PDT)

Nevermind, Ben's patch is right, I'll apply it.

VMEMMAP_SIZE is number of entries, not a size in bytes.

Further confusing things is that before the VMEMMAP_SIZE bug fix the other month, the calculation was "correct" :-)