[net-next] tipc: add NULL pointer check

[net-next] tipc: add NULL pointer check

[Hoang Le]

skb somehow dequeued out of inputq before processing, it causes to
NULL pointer and kernel crashed.

Add checking skb valid before using.

Fixes: c55c8edafa9 ("tipc: smooth change between replicast and broadcast")
Reported-by: Tuong Lien Tong <tuong.t.lien@dektech.com.au>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
---
 net/tipc/bcast.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c
index 76e14dc08bb9..6c997d4a6218 100644
--- a/net/tipc/bcast.c
+++ b/net/tipc/bcast.c
@@ -769,6 +769,9 @@ void tipc_mcast_filter_msg(struct net *net, struct sk_buff_head *defq,
 	u32 node, port;
 
 	skb = skb_peek(inputq);
+	if (!skb)
+		return;
+
 	hdr = buf_msg(skb);
 
 	if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
-- 
2.17.1

RE: [net-next] tipc: add NULL pointer check

[Jon Maloy]

Acked-by: Jon Maloy <jon.maloy@ericsson.com>

Although "somehow" is not the good term here,- the reason is obvious when looking into tipc_sk_proto_rcv().


> -----Original Message-----
> From: netdev-owner@vger.kernel.org <netdev-owner@vger.kernel.org>
> On Behalf Of Hoang Le
> Sent: 3-Apr-19 02:05
> To: Jon Maloy <jon.maloy@ericsson.com>; maloy@donjonn.com;
> ying.xue@windriver.com; tipc-discussion@lists.sourceforge.net;
> netdev@vger.kernel.org
> Subject: [net-next] tipc: add NULL pointer check
> 
> skb somehow dequeued out of inputq before processing, it causes to NULL
> pointer and kernel crashed.
> 
> Add checking skb valid before using.
> 
> Fixes: c55c8edafa9 ("tipc: smooth change between replicast and broadcast")
> Reported-by: Tuong Lien Tong <tuong.t.lien@dektech.com.au>
> Acked-by: Ying Xue <ying.xue@windriver.com>
> Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
> ---
>  net/tipc/bcast.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index
> 76e14dc08bb9..6c997d4a6218 100644
> --- a/net/tipc/bcast.c
> +++ b/net/tipc/bcast.c
> @@ -769,6 +769,9 @@ void tipc_mcast_filter_msg(struct net *net, struct
> sk_buff_head *defq,
>  	u32 node, port;
> 
>  	skb = skb_peek(inputq);
> +	if (!skb)
> +		return;
> +
>  	hdr = buf_msg(skb);
> 
>  	if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
> --
> 2.17.1

Re: [net-next] tipc: add NULL pointer check

[Bo YU]

On Wed, Apr 3, 2019 at 8:55 PM Jon Maloy <jon.maloy@ericsson.com> wrote:
>
> Acked-by: Jon Maloy <jon.maloy@ericsson.com>
>
> Although "somehow" is not the good term here,- the reason is obvious when looking into tipc_sk_proto_rcv().
If so,skb_peek in tipc_mcast_xmit() maybe alert?
>
>
> > -----Original Message-----
> > From: netdev-owner@vger.kernel.org <netdev-owner@vger.kernel.org>
> > On Behalf Of Hoang Le
> > Sent: 3-Apr-19 02:05
> > To: Jon Maloy <jon.maloy@ericsson.com>; maloy@donjonn.com;
> > ying.xue@windriver.com; tipc-discussion@lists.sourceforge.net;
> > netdev@vger.kernel.org
> > Subject: [net-next] tipc: add NULL pointer check
> >
> > skb somehow dequeued out of inputq before processing, it causes to NULL
> > pointer and kernel crashed.
> >
> > Add checking skb valid before using.
> >
> > Fixes: c55c8edafa9 ("tipc: smooth change between replicast and broadcast")
> > Reported-by: Tuong Lien Tong <tuong.t.lien@dektech.com.au>
> > Acked-by: Ying Xue <ying.xue@windriver.com>
> > Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
> > ---
> >  net/tipc/bcast.c | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index
> > 76e14dc08bb9..6c997d4a6218 100644
> > --- a/net/tipc/bcast.c
> > +++ b/net/tipc/bcast.c
> > @@ -769,6 +769,9 @@ void tipc_mcast_filter_msg(struct net *net, struct
> > sk_buff_head *defq,
> >       u32 node, port;
> >
> >       skb = skb_peek(inputq);
> > +     if (!skb)
> > +             return;
> > +
> >       hdr = buf_msg(skb);
> >
> >       if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
> > --
> > 2.17.1
>

RE: [net-next] tipc: add NULL pointer check

[Jon Maloy]

> -----Original Message-----
> From: Bo YU <tsu.yubo@gmail.com>
> Sent: 3-Apr-19 09:10
> To: Jon Maloy <jon.maloy@ericsson.com>
> Cc: Hoang Huu Le <hoang.h.le@dektech.com.au>; maloy@donjonn.com;
> ying.xue@windriver.com; tipc-discussion@lists.sourceforge.net;
> netdev@vger.kernel.org
> Subject: Re: [net-next] tipc: add NULL pointer check
> 
> On Wed, Apr 3, 2019 at 8:55 PM Jon Maloy <jon.maloy@ericsson.com>
> wrote:
> >
> > Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> >
> > Although "somehow" is not the good term here,- the reason is obvious
> when looking into tipc_sk_proto_rcv().
> If so,skb_peek in tipc_mcast_xmit() maybe alert?

Don't understand your question. tipc_mcast_xmit() is on the send path,  we are on the receive path here.
Besides, this is a completely normal occurrence, so no need for any warning.

///jon
 
> >
> >
> > > -----Original Message-----
> > > From: netdev-owner@vger.kernel.org <netdev-
> owner@vger.kernel.org> On
> > > Behalf Of Hoang Le
> > > Sent: 3-Apr-19 02:05
> > > To: Jon Maloy <jon.maloy@ericsson.com>; maloy@donjonn.com;
> > > ying.xue@windriver.com; tipc-discussion@lists.sourceforge.net;
> > > netdev@vger.kernel.org
> > > Subject: [net-next] tipc: add NULL pointer check
> > >
> > > skb somehow dequeued out of inputq before processing, it causes to
> > > NULL pointer and kernel crashed.
> > >
> > > Add checking skb valid before using.
> > >
> > > Fixes: c55c8edafa9 ("tipc: smooth change between replicast and
> > > broadcast")
> > > Reported-by: Tuong Lien Tong <tuong.t.lien@dektech.com.au>
> > > Acked-by: Ying Xue <ying.xue@windriver.com>
> > > Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
> > > ---
> > >  net/tipc/bcast.c | 3 +++
> > >  1 file changed, 3 insertions(+)
> > >
> > > diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index
> > > 76e14dc08bb9..6c997d4a6218 100644
> > > --- a/net/tipc/bcast.c
> > > +++ b/net/tipc/bcast.c
> > > @@ -769,6 +769,9 @@ void tipc_mcast_filter_msg(struct net *net,
> > > struct sk_buff_head *defq,
> > >       u32 node, port;
> > >
> > >       skb = skb_peek(inputq);
> > > +     if (!skb)
> > > +             return;
> > > +
> > >       hdr = buf_msg(skb);
> > >
> > >       if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
> > > --
> > > 2.17.1
> >

Re: [net-next] tipc: add NULL pointer check

[Bo YU]

On Wed, Apr 3, 2019 at 10:27 PM Jon Maloy <jon.maloy@ericsson.com> wrote:
>
> > -----Original Message-----
> > From: Bo YU <tsu.yubo@gmail.com>
> > Sent: 3-Apr-19 09:10
> > To: Jon Maloy <jon.maloy@ericsson.com>
> > Cc: Hoang Huu Le <hoang.h.le@dektech.com.au>; maloy@donjonn.com;
> > ying.xue@windriver.com; tipc-discussion@lists.sourceforge.net;
> > netdev@vger.kernel.org
> > Subject: Re: [net-next] tipc: add NULL pointer check
> >
> > On Wed, Apr 3, 2019 at 8:55 PM Jon Maloy <jon.maloy@ericsson.com>
> > wrote:
> > >
> > > Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> > >
> > > Although "somehow" is not the good term here,- the reason is obvious
> > when looking into tipc_sk_proto_rcv().
> > If so,skb_peek in tipc_mcast_xmit() maybe alert?
>
> Don't understand your question. tipc_mcast_xmit() is on the send path,  we are on the receive path here.
> Besides, this is a completely normal occurrence, so no need for any warning.
Ok, thank you for explation.
>
> ///jon
>
> > >
> > >
> > > > -----Original Message-----
> > > > From: netdev-owner@vger.kernel.org <netdev-
> > owner@vger.kernel.org> On
> > > > Behalf Of Hoang Le
> > > > Sent: 3-Apr-19 02:05
> > > > To: Jon Maloy <jon.maloy@ericsson.com>; maloy@donjonn.com;
> > > > ying.xue@windriver.com; tipc-discussion@lists.sourceforge.net;
> > > > netdev@vger.kernel.org
> > > > Subject: [net-next] tipc: add NULL pointer check
> > > >
> > > > skb somehow dequeued out of inputq before processing, it causes to
> > > > NULL pointer and kernel crashed.
> > > >
> > > > Add checking skb valid before using.
> > > >
> > > > Fixes: c55c8edafa9 ("tipc: smooth change between replicast and
> > > > broadcast")
> > > > Reported-by: Tuong Lien Tong <tuong.t.lien@dektech.com.au>
> > > > Acked-by: Ying Xue <ying.xue@windriver.com>
> > > > Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
> > > > ---
> > > >  net/tipc/bcast.c | 3 +++
> > > >  1 file changed, 3 insertions(+)
> > > >
> > > > diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index
> > > > 76e14dc08bb9..6c997d4a6218 100644
> > > > --- a/net/tipc/bcast.c
> > > > +++ b/net/tipc/bcast.c
> > > > @@ -769,6 +769,9 @@ void tipc_mcast_filter_msg(struct net *net,
> > > > struct sk_buff_head *defq,
> > > >       u32 node, port;
> > > >
> > > >       skb = skb_peek(inputq);
> > > > +     if (!skb)
> > > > +             return;
> > > > +
> > > >       hdr = buf_msg(skb);
> > > >
> > > >       if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
> > > > --
> > > > 2.17.1
> > >

Re: [net-next] tipc: add NULL pointer check

[David Miller]

From: Hoang Le <hoang.h.le@dektech.com.au>
Date: Wed,  3 Apr 2019 13:05:04 +0700

> skb somehow dequeued out of inputq before processing, it causes to
> NULL pointer and kernel crashed.
> 
> Add checking skb valid before using.
> 
> Fixes: c55c8edafa9 ("tipc: smooth change between replicast and broadcast")
> Reported-by: Tuong Lien Tong <tuong.t.lien@dektech.com.au>
> Acked-by: Ying Xue <ying.xue@windriver.com>
> Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>

Applied.