qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH 1/2] net: stellaris_enet: check packet length ag


From: Peter Maydell
Subject: Re: [Qemu-devel] [PATCH 1/2] net: stellaris_enet: check packet length against receive buffer
Date: Thu, 7 Apr 2016 12:02:13 +0100

On 7 April 2016 at 11:25, P J P <address@hidden> wrote:
> From: Prasad J Pandit <address@hidden>
>
> When receiving packets over Stellaris ethernet controller, it
> uses receive buffer of size 2048 bytes. In case the controller
> accepts large(MTU) packets, it could lead to memory corruption.
> Add check to avoid it.
>
> Reported by: Oleksandr Bazhaniuk <address@hidden>
>
> Signed-off-by: Prasad J Pandit <address@hidden>
> ---
>  hw/net/stellaris_enet.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/hw/net/stellaris_enet.c b/hw/net/stellaris_enet.c
> index 21a4773..47b869a 100644
> --- a/hw/net/stellaris_enet.c
> +++ b/hw/net/stellaris_enet.c
> @@ -237,6 +237,9 @@ static ssize_t stellaris_enet_receive(NetClientState *nc, 
> const uint8_t *buf, si
>          n -= 31;
>      s->np++;

We should do this check before we increase s->np, because
if we're going to bail out then we won't be putting this
packet into the RX FIFO.

> +    if (size >= sizeof(s->rx[n].data) - 6) {
> +        return -1;
> +    }

The datasheet for this chip says that we should report this
to the guest:

"For the Ethernet Controller, data sent/received can be larger
than 1500 bytes without causing a Frame Too Long error. Instead,
a FIFO overrun error is reported using the FOV bit in the
Ethernet MAC Raw Interrupt Status (MACRIS) register when the
frame received is too large to fit in the Ethernet Controller's
2K RAM."
So you want something like:

   /* If the packet won't fit into the emulated 2K RAM, this
    * is reported as a FIFO overrun error.
    */
   s->ris |= SE_INT_FOV;
   stellaris_enet_update(s);
   return -1;


thanks
-- PMM



reply via email to

[Prev in Thread] Current Thread [Next in Thread]