[PATCH Lucid/Natty CVE-2012-3511] mm: Hold a file reference in madvise_remove

Tim Gardner tim.gardner at canonical.com
Thu Aug 30 15:22:48 UTC 2012


From: Andy Lutomirski <luto at amacapital.net>

CVE-2012-3511

BugLink: http://bugs.launchpad.net/bugs/1042447

Otherwise the code races with munmap (causing a use-after-free
of the vma) or with close (causing a use-after-free of the struct
file).

The bug was introduced by commit 90ed52ebe481 ("[PATCH] holepunch: fix
mmap_sem i_mutex deadlock")

Cc: Hugh Dickins <hugh at veritas.com>
Cc: Miklos Szeredi <mszeredi at suse.cz>
Cc: Badari Pulavarty <pbadari at us.ibm.com>
Cc: Nick Piggin <npiggin at suse.de>
Cc: stable at vger.kernel.org
Signed-off-by: Andy Lutomirski <luto at amacapital.net>
Signed-off-by: Linus Torvalds <torvalds at linux-foundation.org>
(back ported from commit 9ab4233dd08036fe34a89c7dc6f47a8bf2eb29eb)

Signed-off-by: Tim Gardner <tim.gardner at canonical.com>
---
 mm/madvise.c |   16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index 35b1479..e51291d 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -5,6 +5,7 @@
  * Copyright (C) 2002  Christoph Hellwig
  */
 
+#include <linux/file.h>
 #include <linux/mman.h>
 #include <linux/pagemap.h>
 #include <linux/syscalls.h>
@@ -190,14 +191,16 @@ static long madvise_remove(struct vm_area_struct *vma,
 	struct address_space *mapping;
 	loff_t offset, endoff;
 	int error;
+	struct file *f;
 
 	*prev = NULL;	/* tell sys_madvise we drop mmap_sem */
 
 	if (vma->vm_flags & (VM_LOCKED|VM_NONLINEAR|VM_HUGETLB))
 		return -EINVAL;
 
-	if (!vma->vm_file || !vma->vm_file->f_mapping
-		|| !vma->vm_file->f_mapping->host) {
+	f = vma->vm_file;
+
+	if (!f || !f->f_mapping || !f->f_mapping->host) {
 			return -EINVAL;
 	}
 
@@ -211,9 +214,16 @@ static long madvise_remove(struct vm_area_struct *vma,
 	endoff = (loff_t)(end - vma->vm_start - 1)
 			+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
 
-	/* vmtruncate_range needs to take i_mutex and i_alloc_sem */
+	/*
+	 * Filesystem's fallocate may need to take i_mutex.  We need to
+	 * explicitly grab a reference because the vma (and hence the
+	 * vma's reference to the file) can go away as soon as we drop
+	 * mmap_sem.
+	 */
+	get_file(f);
 	up_read(&current->mm->mmap_sem);
 	error = vmtruncate_range(mapping->host, offset, endoff);
+	fput(f);
 	down_read(&current->mm->mmap_sem);
 	return error;
 }
-- 
1.7.9.5





More information about the kernel-team mailing list