[v8,3/3] hv_netvsc: Use vmbus_requestor to generate transaction IDs for VMBus hardening

From: Andres Beltran <lkmlabelt@gmail.com>

Currently, pointers to guest memory are passed to Hyper-V as
transaction IDs in netvsc. In the face of errors or malicious
behavior in Hyper-V, netvsc should not expose or trust the transaction
IDs returned by Hyper-V to be valid guest memory addresses. Instead,
use small integers generated by vmbus_requestor as requests
(transaction) IDs.

Signed-off-by: Andres Beltran <lkmlabelt@gmail.com>
Co-developed-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>
Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: netdev@vger.kernel.org
---
 drivers/net/hyperv/hyperv_net.h   | 13 +++++++++++++
 drivers/net/hyperv/netvsc.c       | 22 ++++++++++++++++------
 drivers/net/hyperv/rndis_filter.c |  1 +
 include/linux/hyperv.h            |  1 +
 4 files changed, 31 insertions(+), 6 deletions(-)

On Wed,  4 Nov 2020 16:40:27 +0100 Andrea Parri (Microsoft) wrote:
> From: Andres Beltran <lkmlabelt@gmail.com>

> 

> Currently, pointers to guest memory are passed to Hyper-V as

> transaction IDs in netvsc. In the face of errors or malicious

> behavior in Hyper-V, netvsc should not expose or trust the transaction

> IDs returned by Hyper-V to be valid guest memory addresses. Instead,

> use small integers generated by vmbus_requestor as requests

> (transaction) IDs.

> 

> Signed-off-by: Andres Beltran <lkmlabelt@gmail.com>

> Co-developed-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>

> Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>

> Reviewed-by: Michael Kelley <mikelley@microsoft.com>


I'm assuming this is targeting net-next? If so could you please tag it
as [PATCH net-next vN]?

> @@ -695,10 +695,19 @@ static void netvsc_send_tx_complete(struct net_device *ndev,

>  				    const struct vmpacket_descriptor *desc,

>  				    int budget)

>  {

> -	struct sk_buff *skb = (struct sk_buff *)(unsigned long)desc->trans_id;

> +	struct sk_buff *skb;

>  	struct net_device_context *ndev_ctx = netdev_priv(ndev);


Swap these two lines please to keep the variables declaration lines
longest to shortest.

On Wed, 4 Nov 2020 13:43:48 -0800 Jakub Kicinski wrote:
> On Wed,  4 Nov 2020 16:40:27 +0100 Andrea Parri (Microsoft) wrote:

> > From: Andres Beltran <lkmlabelt@gmail.com>

> > 

> > Currently, pointers to guest memory are passed to Hyper-V as

> > transaction IDs in netvsc. In the face of errors or malicious

> > behavior in Hyper-V, netvsc should not expose or trust the transaction

> > IDs returned by Hyper-V to be valid guest memory addresses. Instead,

> > use small integers generated by vmbus_requestor as requests

> > (transaction) IDs.

> > 

> > Signed-off-by: Andres Beltran <lkmlabelt@gmail.com>

> > Co-developed-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>

> > Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>

> > Reviewed-by: Michael Kelley <mikelley@microsoft.com>  

> 

> I'm assuming this is targeting net-next? If so could you please tag it

> as [PATCH net-next vN]?


Ah, you don't 'cause you only sent us the third patch. In that case with
the nit below addressed:

Acked-by: Jakub Kicinski <kuba@kernel.org>


> > @@ -695,10 +695,19 @@ static void netvsc_send_tx_complete(struct net_device *ndev,

> >  				    const struct vmpacket_descriptor *desc,

> >  				    int budget)

> >  {

> > -	struct sk_buff *skb = (struct sk_buff *)(unsigned long)desc->trans_id;

> > +	struct sk_buff *skb;

> >  	struct net_device_context *ndev_ctx = netdev_priv(ndev);  

> 

> Swap these two lines please to keep the variables declaration lines

> longest to shortest.

On Wed, Nov 04, 2020 at 01:45:05PM -0800, Jakub Kicinski wrote:
> On Wed, 4 Nov 2020 13:43:48 -0800 Jakub Kicinski wrote:
> > On Wed,  4 Nov 2020 16:40:27 +0100 Andrea Parri (Microsoft) wrote:
> > > From: Andres Beltran <lkmlabelt@gmail.com>
> > > 
> > > Currently, pointers to guest memory are passed to Hyper-V as
> > > transaction IDs in netvsc. In the face of errors or malicious
> > > behavior in Hyper-V, netvsc should not expose or trust the transaction
> > > IDs returned by Hyper-V to be valid guest memory addresses. Instead,
> > > use small integers generated by vmbus_requestor as requests
> > > (transaction) IDs.
> > > 
> > > Signed-off-by: Andres Beltran <lkmlabelt@gmail.com>
> > > Co-developed-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>
> > > Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>
> > > Reviewed-by: Michael Kelley <mikelley@microsoft.com>  
> > 
> > I'm assuming this is targeting net-next? If so could you please tag it
> > as [PATCH net-next vN]?
> 
> Ah, you don't 'cause you only sent us the third patch. In that case with
> the nit below addressed:
> 
> Acked-by: Jakub Kicinski <kuba@kernel.org>

I fixed the declarations locally.  Thank you for the review, Jakub.

(Yes, FWIW, I was imaging the series to go via hyperv-next...)

Thanks,
  Andrea


> 
> > > @@ -695,10 +695,19 @@ static void netvsc_send_tx_complete(struct net_device *ndev,
> > >  				    const struct vmpacket_descriptor *desc,
> > >  				    int budget)
> > >  {
> > > -	struct sk_buff *skb = (struct sk_buff *)(unsigned long)desc->trans_id;
> > > +	struct sk_buff *skb;
> > >  	struct net_device_context *ndev_ctx = netdev_priv(ndev);  
> > 
> > Swap these two lines please to keep the variables declaration lines
> > longest to shortest.
>