[PATCH] net: Fix module refcount leak in kernel_accept()

[PATCH] net: Fix module refcount leak in kernel_accept()

[Wei Yongjun]

The kernel_accept() does not hold the module refcount of newsock->ops->owner,
so we need __module_get(newsock->ops->owner) code after call kernel_accept()
by hand.
In sunrpc, the module refcount is missing to hold. So this cause kernel panic.

Used following script to reproduct:

while [ 1 ];
do
    mount -t nfs4 192.168.0.19:/ /mnt
    touch /mnt/file
    umount /mnt
    lsmod | grep ipv6
done

This patch fixed the problem by add __module_get(newsock->ops->owner) to
kernel_accept(). So we do not need to used __module_get(newsock->ops->owner)
in every place when used kernel_accept().

Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
---
 net/bluetooth/rfcomm/core.c |    2 --
 net/socket.c                |    1 +
 2 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c
index ba537fa..ce68e04 100644
--- a/net/bluetooth/rfcomm/core.c
+++ b/net/bluetooth/rfcomm/core.c
@@ -1786,8 +1786,6 @@ static inline void rfcomm_accept_connection(struct rfcomm_session *s)
 	if (err < 0)
 		return;
 
-	__module_get(nsock->ops->owner);
-
 	/* Set our callbacks */
 	nsock->sk->sk_data_ready   = rfcomm_l2data_ready;
 	nsock->sk->sk_state_change = rfcomm_l2state_change;
diff --git a/net/socket.c b/net/socket.c
index 92764d8..76ba80a 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -2307,6 +2307,7 @@ int kernel_accept(struct socket *sock, struct socket **newsock, int flags)
 	}
 
 	(*newsock)->ops = sock->ops;
+	__module_get((*newsock)->ops->owner);
 
 done:
 	return err;
-- 
1.5.3.8

Re: [PATCH] net: Fix module refcount leak in kernel_accept()

[David Miller]

From: Wei Yongjun <yjwei@cn.fujitsu.com>
Date: Thu, 18 Dec 2008 18:20:46 +0800

> The kernel_accept() does not hold the module refcount of newsock->ops->owner,
> so we need __module_get(newsock->ops->owner) code after call kernel_accept()
> by hand.
> In sunrpc, the module refcount is missing to hold. So this cause kernel panic.
> 
> Used following script to reproduct:
 ...
> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>

Good fine, patch applied, thanks!

Re: [PATCH] net: Fix module refcount leak in kernel_accept()

[Marcel Holtmann]

Hi Wei,

> The kernel_accept() does not hold the module refcount of newsock->ops->owner,
> so we need __module_get(newsock->ops->owner) code after call kernel_accept()
> by hand.
> In sunrpc, the module refcount is missing to hold. So this cause kernel panic.
> 
> Used following script to reproduct:
> 
> while [ 1 ];
> do
>     mount -t nfs4 192.168.0.19:/ /mnt
>     touch /mnt/file
>     umount /mnt
>     lsmod | grep ipv6
> done
> 
> This patch fixed the problem by add __module_get(newsock->ops->owner) to
> kernel_accept(). So we do not need to used __module_get(newsock->ops->owner)
> in every place when used kernel_accept().
> 
> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
> ---
>  net/bluetooth/rfcomm/core.c |    2 --
>  net/socket.c                |    1 +
>  2 files changed, 1 insertions(+), 2 deletions(-)

I was just about to reply and ask you to double check the users since I
know that I am using that API. Hey, but you already did that. Thanks.

Dave, for what its worth, Acked-by me.

Regards

Marcel

Re: [PATCH] net: Fix module refcount leak in kernel_accept()

[David Miller]

From: Marcel Holtmann <marcel@holtmann.org>
Date: Fri, 19 Dec 2008 07:35:37 +0100

> Hi Wei,
> 
> > The kernel_accept() does not hold the module refcount of newsock->ops->owner,
> > so we need __module_get(newsock->ops->owner) code after call kernel_accept()
> > by hand.
> > In sunrpc, the module refcount is missing to hold. So this cause kernel panic.
> > 
> > Used following script to reproduct:
> > 
> > while [ 1 ];
> > do
> >     mount -t nfs4 192.168.0.19:/ /mnt
> >     touch /mnt/file
> >     umount /mnt
> >     lsmod | grep ipv6
> > done
> > 
> > This patch fixed the problem by add __module_get(newsock->ops->owner) to
> > kernel_accept(). So we do not need to used __module_get(newsock->ops->owner)
> > in every place when used kernel_accept().
> > 
> > Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
> > ---
> >  net/bluetooth/rfcomm/core.c |    2 --
> >  net/socket.c                |    1 +
> >  2 files changed, 1 insertions(+), 2 deletions(-)
> 
> I was just about to reply and ask you to double check the users since I
> know that I am using that API. Hey, but you already did that. Thanks.

That's the exact audit I did before applying his patch :-)