[PATCH] ppc64: Make NUMA code handle unexpected layouts

[Anton Blanchard]

From: Nathan Lynch <nathanl@xxxxxxxxxxxxxx>

Ran into this on a 4GB partition - all but about ~300MB was thrown away.

Does this look ok?  It works for me, but I've not tested on firmware
without the bug.

Fall back to non-numa setup upon discovering unexpected memory layout
as presented by firmware, instead of throwing away regions.

Signed-off-by: Nathan Lynch <nathanl@xxxxxxxxxxxxxx>
Acked-by: Anton Blanchard <anton@xxxxxxxxx>

diff -puN arch/ppc64/mm/numa.c~numa_broke_fix arch/ppc64/mm/numa.c
--- foobar2/arch/ppc64/mm/numa.c~numa_broke_fix	2005-01-10 23:36:39.922337007 +1100
+++ foobar2-anton/arch/ppc64/mm/numa.c	2005-01-10 23:46:46.069205815 +1100
@@ -345,8 +345,6 @@ new_range:
 			numa_domain = 0;
 		}
 
-		node_set_online(numa_domain);
-
 		if (max_domain < numa_domain)
 			max_domain = numa_domain;
 
@@ -361,14 +359,19 @@ new_range:
 				init_node_data[numa_domain].node_start_pfn +
 				init_node_data[numa_domain].node_spanned_pages;
 			if (shouldstart != (start / PAGE_SIZE)) {
-				printk(KERN_ERR "WARNING: Hole in node, "
-						"disabling region start %lx "
-						"length %lx\n", start, size);
-				continue;
+				/* Revert to non-numa for now */
+				printk(KERN_ERR
+				       "WARNING: Unexpected node layout: "
+				       "region start %lx length %lx\n",
+				       start, size);
+				printk(KERN_ERR "NUMA is disabled\n");
+				goto err;
 			}
 			init_node_data[numa_domain].node_spanned_pages +=
 				size / PAGE_SIZE;
 		} else {
+			node_set_online(numa_domain);
+
 			init_node_data[numa_domain].node_start_pfn =
 				start / PAGE_SIZE;
 			init_node_data[numa_domain].node_spanned_pages =
@@ -388,6 +391,14 @@ new_range:
 		node_set_online(i);
 
 	return 0;
+err:
+	/* Something has gone wrong; revert any setup we've done */
+	for_each_node(i) {
+		node_set_offline(i);
+		init_node_data[i].node_start_pfn = 0;
+		init_node_data[i].node_spanned_pages = 0;
+	}
+	return -1;
 }
 
 static void __init setup_nonnuma(void)
_

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/