Thread (4 messages) 4 messages, 3 authors, 2011-08-24

Re: [PATCH -v3] avoid null pointer access in vm_struct

From: Andrew Morton <akpm@linux-foundation.org>
Date: 2011-08-22 22:25:57
Also in: lkml

On Sun, 21 Aug 2011 17:21:32 +0900
Mitsuo Hayasaka [off-list ref] wrote:
The /proc/vmallocinfo shows information about vmalloc allocations in vmlist
that is a linklist of vm_struct. It, however, may access pages field of
vm_struct where a page was not allocated. This results in a null pointer
access and leads to a kernel panic.

Why this happen:
In __vmalloc_node_range() called from vmalloc(), newly allocated vm_struct
is added to vmlist at __get_vm_area_node() and then, some fields of
vm_struct such as nr_pages and pages are set at __vmalloc_area_node(). In
other words, it is added to vmlist before it is fully initialized. At the
same time, when the /proc/vmallocinfo is read, it accesses the pages field
of vm_struct according to the nr_pages field at show_numa_info(). Thus, a
null pointer access happens.

Patch:
This patch adds newly allocated vm_struct to the vmlist *after* it is fully
initialized. So, it can avoid accessing the pages field with unallocated
page when show_numa_info() is called.
Seems rather ugly, but I guess it's OK.  vmalloc() is "special" in that
it fills the area with allocated pages, whereas all the
get_vm_area()-type callers don't do that.
quoted hunk ↗ jump to hunk
...
@@ -1381,17 +1403,20 @@ struct vm_struct *remove_vm_area(const void *addr)
 	va = find_vmap_area((unsigned long)addr);
 	if (va && va->flags & VM_VM_AREA) {
 		struct vm_struct *vm = va->private;
-		struct vm_struct *tmp, **p;
-		/*
-		 * remove from list and disallow access to this vm_struct
-		 * before unmap. (address range confliction is maintained by
-		 * vmap.)
-		 */
-		write_lock(&vmlist_lock);
-		for (p = &vmlist; (tmp = *p) != vm; p = &tmp->next)
-			;
-		*p = tmp->next;
-		write_unlock(&vmlist_lock);
+
+		if (!(vm->flags & VM_UNLIST)) {
+			struct vm_struct *tmp, **p;
+			/*
+			 * remove from list and disallow access to
+			 * this vm_struct before unmap. (address range
+			 * confliction is maintained by vmap.)
+			 */
+			write_lock(&vmlist_lock);
+			for (p = &vmlist; (tmp = *p) != vm; p = &tmp->next)
+				;
+			*p = tmp->next;
+			write_unlock(&vmlist_lock);
+		}
Is this needed?  How can remove_vm_area() actually be called with a
VM_UNLIST area?


I think I'll let this patch cook in linux-next for a while and shall
tag it for backporting into 3.1.x later on.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help