[SRU][T/X/B][PATCH 0/1] Fix for CVE-12232

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[SRU][T/X/B][PATCH 0/1] Fix for CVE-12232

Khalid Elmously
Cleanly cherry-picked for Bionic. Simple backport for Trusty and Xenial

Cong Wang (1):
  socket: close race condition between sock_close() and sockfs_setattr()

 net/socket.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

--
2.17.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[SRU][B]PATCH 1/1] socket: close race condition between sock_close() and sockfs_setattr()

Khalid Elmously
From: Cong Wang <[hidden email]>

fchownat() doesn't even hold refcnt of fd until it figures out
fd is really needed (otherwise is ignored) and releases it after
it resolves the path. This means sock_close() could race with
sockfs_setattr(), which leads to a NULL pointer dereference
since typically we set sock->sk to NULL in ->release().

As pointed out by Al, this is unique to sockfs. So we can fix this
in socket layer by acquiring inode_lock in sock_close() and
checking against NULL in sockfs_setattr().

sock_release() is called in many places, only the sock_close()
path matters here. And fortunately, this should not affect normal
sock_close() as it is only called when the last fd refcnt is gone.
It only affects sock_close() with a parallel sockfs_setattr() in
progress, which is not common.

Fixes: 86741ec25462 ("net: core: Add a UID field to struct sock.")
Reported-by: shankarapailoor <[hidden email]>
Cc: Tetsuo Handa <[hidden email]>
Cc: Lorenzo Colitti <[hidden email]>
Cc: Al Viro <[hidden email]>
Signed-off-by: Cong Wang <[hidden email]>
Signed-off-by: David S. Miller <[hidden email]>
(cherry-picked from 6d8c50dcb029872b298eea68cc6209c866fd3e14)
Signed-off-by: Khalid Elmously <[hidden email]>
---
 net/socket.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

diff --git a/net/socket.c b/net/socket.c
index 6f05d5c4bf30..f2957aa205e6 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -544,7 +544,10 @@ static int sockfs_setattr(struct dentry *dentry, struct iattr *iattr)
  if (!err && (iattr->ia_valid & ATTR_UID)) {
  struct socket *sock = SOCKET_I(d_inode(dentry));
 
- sock->sk->sk_uid = iattr->ia_uid;
+ if (sock->sk)
+ sock->sk->sk_uid = iattr->ia_uid;
+ else
+ err = -ENOENT;
  }
 
  return err;
@@ -594,12 +597,16 @@ EXPORT_SYMBOL(sock_alloc);
  * an inode not a file.
  */
 
-void sock_release(struct socket *sock)
+static void __sock_release(struct socket *sock, struct inode *inode)
 {
  if (sock->ops) {
  struct module *owner = sock->ops->owner;
 
+ if (inode)
+ inode_lock(inode);
  sock->ops->release(sock);
+ if (inode)
+ inode_unlock(inode);
  sock->ops = NULL;
  module_put(owner);
  }
@@ -614,6 +621,11 @@ void sock_release(struct socket *sock)
  }
  sock->file = NULL;
 }
+
+void sock_release(struct socket *sock)
+{
+ __sock_release(sock, NULL);
+}
 EXPORT_SYMBOL(sock_release);
 
 void __sock_tx_timestamp(__u16 tsflags, __u8 *tx_flags)
@@ -1128,7 +1140,7 @@ static int sock_mmap(struct file *file, struct vm_area_struct *vma)
 
 static int sock_close(struct inode *inode, struct file *filp)
 {
- sock_release(SOCKET_I(inode));
+ __sock_release(SOCKET_I(inode), inode);
  return 0;
 }
 
--
2.17.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[SRU][[T/X][PATCH 1/1] socket: close race condition between sock_close() and sockfs_setattr()

Khalid Elmously
In reply to this post by Khalid Elmously
From: Cong Wang <[hidden email]>

fchownat() doesn't even hold refcnt of fd until it figures out
fd is really needed (otherwise is ignored) and releases it after
it resolves the path. This means sock_close() could race with
sockfs_setattr(), which leads to a NULL pointer dereference
since typically we set sock->sk to NULL in ->release().

As pointed out by Al, this is unique to sockfs. So we can fix this
in socket layer by acquiring inode_lock in sock_close() and
checking against NULL in sockfs_setattr().

sock_release() is called in many places, only the sock_close()
path matters here. And fortunately, this should not affect normal
sock_close() as it is only called when the last fd refcnt is gone.
It only affects sock_close() with a parallel sockfs_setattr() in
progress, which is not common.

Fixes: 86741ec25462 ("net: core: Add a UID field to struct sock.")
Reported-by: shankarapailoor <[hidden email]>
Cc: Tetsuo Handa <[hidden email]>
Cc: Lorenzo Colitti <[hidden email]>
Cc: Al Viro <[hidden email]>
Signed-off-by: Cong Wang <[hidden email]>
Signed-off-by: David S. Miller <[hidden email]>
(backported from 6d8c50dcb029872b298eea68cc6209c866fd3e14)
Signed-off-by: Khalid Elmously <[hidden email]>
---
 net/socket.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/net/socket.c b/net/socket.c
index 1895b9eff43b..338d64756cb2 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -564,12 +564,16 @@ static struct socket *sock_alloc(void)
  * an inode not a file.
  */
 
-void sock_release(struct socket *sock)
+static void __sock_release(struct socket *sock, struct inode *inode)
 {
  if (sock->ops) {
  struct module *owner = sock->ops->owner;
 
+ if (inode)
+ inode_lock(inode);
  sock->ops->release(sock);
+ if (inode)
+ inode_unlock(inode);
  sock->ops = NULL;
  module_put(owner);
  }
@@ -584,6 +588,11 @@ void sock_release(struct socket *sock)
  }
  sock->file = NULL;
 }
+
+void sock_release(struct socket *sock)
+{
+ __sock_release(sock, NULL);
+}
 EXPORT_SYMBOL(sock_release);
 
 void __sock_tx_timestamp(const struct sock *sk, __u8 *tx_flags)
@@ -1019,7 +1028,7 @@ static int sock_mmap(struct file *file, struct vm_area_struct *vma)
 
 static int sock_close(struct inode *inode, struct file *filp)
 {
- sock_release(SOCKET_I(inode));
+ __sock_release(SOCKET_I(inode), inode);
  return 0;
 }
 
--
2.17.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

NACK: [SRU][T/X/B][PATCH 0/1] Fix for CVE-12232

Khalid Elmously
In reply to this post by Khalid Elmously
Sending V2 instead

On 2018-07-04 23:52:24 , Khalid Elmously wrote:

> Cleanly cherry-picked for Bionic. Simple backport for Trusty and Xenial
>
> Cong Wang (1):
>   socket: close race condition between sock_close() and sockfs_setattr()
>
>  net/socket.c | 18 +++++++++++++++---
>  1 file changed, 15 insertions(+), 3 deletions(-)
>
> --
> 2.17.1
>

--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team