[PATCH 0/2]: LMB small bug fix and debugging tidy up.

[PATCH 0/2]: LMB small bug fix and debugging tidy up.

[David Miller]

While diagnosing Mikael's bug where half of his RAM disappeared
in 2.6.26-rc2 I found one bug (which I don't think fixes his
problem, but waiting for an updated set of debugging logs from
him) and an improvement to the debugging provided by the current
LMB code.

[rfc patch] memory_hotplug: Check for walk_memory_resource() failure in online_pages()

[Geoff Levand]

Add a check to online_pages() to test for failure of
walk_memory_resource().  This fixes a condition where a failure
of walk_memory_resource() can lead to online_pages() returning
success without the requested pages being onlined.

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
---

I'm not entirely sure this is the proper way to handle this
condition.  Comments welcome.

 mm/memory_hotplug.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -408,8 +408,15 @@ int online_pages(unsigned long pfn, unsi
 	if (!populated_zone(zone))
 		need_zonelists_rebuild = 1;
 
-	walk_memory_resource(pfn, nr_pages, &onlined_pages,
+	ret = walk_memory_resource(pfn, nr_pages, &onlined_pages,
 		online_pages_range);
+	if (ret) {
+		printk(KERN_DEBUG "online_pages %lx at %lx failed\n",
+			nr_pages, pfn);
+		memory_notify(MEM_CANCEL_ONLINE, &arg);
+		return ret;
+	}
+
 	zone->present_pages += onlined_pages;
 	zone->zone_pgdat->node_present_pages += onlined_pages;