[PATCH] printk() should not be called under zone->lock

[PATCH] printk() should not be called under zone->lock

[Kirill Korotaev]

[-- Attachment #1: Type: text/plain, Size: 434 bytes --]

This patch fixes printk() under zone->lock in show_free_areas().
It can be unsafe to call printk() under this lock, since
caller can try to allocate/free some memory and selfdeadlock
on this lock.
I found allocations/freeing mem both in netconsole and serial console.

This issue was faced in reallity when meminfo was periodically
printed for debug purposes and netconsole was used.

Signed-Off-By: Kirill Korotaev <dev@openvz.org>


[-- Attachment #2: diff-show-free-areas --]
[-- Type: text/plain, Size: 929 bytes --]

--- linux-2.6.17-rc6-mm1s.orig/mm/page_alloc.c	2006-06-09 15:37:09.000000000 +0400
+++ linux-2.6.16-rc6-mm1/mm/page_alloc.c	2006-06-15 12:42:31.000000000 +0400
@@ -1580,7 +1580,7 @@ void show_free_areas(void)
 	}
 
 	for_each_zone(zone) {
- 		unsigned long nr, flags, order, total = 0;
+ 		unsigned long nr[MAX_ORDER], flags, order, total = 0;
 
 		show_node(zone);
 		printk("%s: ", zone->name);
@@ -1591,11 +1591,12 @@ void show_free_areas(void)
 
 		spin_lock_irqsave(&zone->lock, flags);
 		for (order = 0; order < MAX_ORDER; order++) {
-			nr = zone->free_area[order].nr_free;
-			total += nr << order;
-			printk("%lu*%lukB ", nr, K(1UL) << order);
+			nr[order] = zone->free_area[order].nr_free;
+			total += nr[order] << order;
 		}
 		spin_unlock_irqrestore(&zone->lock, flags);
+		for (order = 0; order < MAX_ORDER; order++)
+			printk("%lu*%lukB ", nr[order], K(1UL) << order);
 		printk("= %lukB\n", K(total));
 	}