[kernel/f15] CVE-2011-2183: ksm: race between ksmd and exiting task
Chuck Ebbert
cebbert at fedoraproject.org
Fri Jun 24 10:12:22 UTC 2011
commit ced86e1611c925a35187083210c890f8bf7c83e8
Author: Chuck Ebbert <cebbert at redhat.com>
Date: Fri Jun 24 06:12:52 2011 -0400
CVE-2011-2183: ksm: race between ksmd and exiting task
kernel.spec | 8 ++
...er-dereference-in-scan-get-next-rmap-item.patch | 82 ++++++++++++++++++++
2 files changed, 90 insertions(+), 0 deletions(-)
---
diff --git a/kernel.spec b/kernel.spec
index e15ec24..cf07879 100644
--- a/kernel.spec
+++ b/kernel.spec
@@ -619,6 +619,9 @@ Patch30: linux-2.6-tracehook.patch
Patch31: linux-2.6-utrace.patch
Patch32: linux-2.6-utrace-ptrace.patch
+# CVE-2011-2183
+Patch80: ksm-fix-null-pointer-dereference-in-scan-get-next-rmap-item.patch
+
Patch150: linux-2.6.29-sparc-IOC_TYPECHECK.patch
Patch160: linux-2.6-32bit-mmap-exec-randomization.patch
@@ -1210,6 +1213,10 @@ ApplyPatch linux-2.6-tracehook.patch
ApplyPatch linux-2.6-utrace.patch
ApplyPatch linux-2.6-utrace-ptrace.patch
+# mm patches
+# CVE-2011-2183
+ApplyPatch ksm-fix-null-pointer-dereference-in-scan-get-next-rmap-item.patch
+
# Architecture patches
# x86(-64)
# Restore reliable stack backtraces, and hopefully
@@ -2037,6 +2044,7 @@ fi
- block-blkdev_get-should-access-bd_disk-only-after.patch:
fix potential oops introduced in 2.6.38.8
- ahci-add-another-pci-id-for-marvell.patch (#705960)
+- CVE-2011-2183: ksm: race between ksmd and exiting task
* Thu Jun 23 2011 Dave Airlie <airlied at redhat.com> 2.6.38.8-34
- drm-i915-snb-irq-stalls-fix.patch: fix Sandybridge IRQ stalls
diff --git a/ksm-fix-null-pointer-dereference-in-scan-get-next-rmap-item.patch b/ksm-fix-null-pointer-dereference-in-scan-get-next-rmap-item.patch
new file mode 100644
index 0000000..d67cdc7
--- /dev/null
+++ b/ksm-fix-null-pointer-dereference-in-scan-get-next-rmap-item.patch
@@ -0,0 +1,82 @@
+From: Hugh Dickins <hughd at google.com>
+Date: Wed, 15 Jun 2011 22:08:58 +0000 (-0700)
+Subject: ksm: fix NULL pointer dereference in scan_get_next_rmap_item()
+X-Git-Tag: v3.0-rc4~44
+X-Git-Url: http://git.kernel.org/?p=linux%2Fkernel%2Fgit%2Ftorvalds%2Flinux-2.6.git;a=commitdiff_plain;h=2b472611a32a72f4a118c069c2d62a1a3f087afd
+
+ksm: fix NULL pointer dereference in scan_get_next_rmap_item()
+
+Andrea Righi reported a case where an exiting task can race against
+ksmd::scan_get_next_rmap_item (http://lkml.org/lkml/2011/6/1/742) easily
+triggering a NULL pointer dereference in ksmd.
+
+ksm_scan.mm_slot == &ksm_mm_head with only one registered mm
+
+CPU 1 (__ksm_exit) CPU 2 (scan_get_next_rmap_item)
+ list_empty() is false
+lock slot == &ksm_mm_head
+list_del(slot->mm_list)
+(list now empty)
+unlock
+ lock
+ slot = list_entry(slot->mm_list.next)
+ (list is empty, so slot is still ksm_mm_head)
+ unlock
+ slot->mm == NULL ... Oops
+
+Close this race by revalidating that the new slot is not simply the list
+head again.
+
+Andrea's test case:
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/mman.h>
+
+#define BUFSIZE getpagesize()
+
+int main(int argc, char **argv)
+{
+ void *ptr;
+
+ if (posix_memalign(&ptr, getpagesize(), BUFSIZE) < 0) {
+ perror("posix_memalign");
+ exit(1);
+ }
+ if (madvise(ptr, BUFSIZE, MADV_MERGEABLE) < 0) {
+ perror("madvise");
+ exit(1);
+ }
+ *(char *)NULL = 0;
+
+ return 0;
+}
+
+Reported-by: Andrea Righi <andrea at betterlinux.com>
+Tested-by: Andrea Righi <andrea at betterlinux.com>
+Cc: Andrea Arcangeli <aarcange at redhat.com>
+Signed-off-by: Hugh Dickins <hughd at google.com>
+Signed-off-by: Chris Wright <chrisw at sous-sol.org>
+Cc: <stable at kernel.org>
+Signed-off-by: Andrew Morton <akpm at linux-foundation.org>
+Signed-off-by: Linus Torvalds <torvalds at linux-foundation.org>
+---
+
+diff --git a/mm/ksm.c b/mm/ksm.c
+index d708b3e..9a68b0c 100644
+--- a/mm/ksm.c
++++ b/mm/ksm.c
+@@ -1302,6 +1302,12 @@ static struct rmap_item *scan_get_next_rmap_item(struct page **page)
+ slot = list_entry(slot->mm_list.next, struct mm_slot, mm_list);
+ ksm_scan.mm_slot = slot;
+ spin_unlock(&ksm_mmlist_lock);
++ /*
++ * Although we tested list_empty() above, a racing __ksm_exit
++ * of the last mm on the list may have removed it since then.
++ */
++ if (slot == &ksm_mm_head)
++ return NULL;
+ next_mm:
+ ksm_scan.address = 0;
+ ksm_scan.rmap_list = &slot->rmap_list;
More information about the scm-commits
mailing list