On Fri, Nov 05, 2010 at 02:52:08PM -0600, Alex Williamson wrote:
> When adding the length to the pseudo header, we're not properly
> accounting for overflow.
> 
> From: Mark Wu <d...@redhat.com>
> Signed-off-by: Alex Williamson <alex.william...@redhat.com>

Thanks, applied.

> ---
> 
>  hw/e1000.c |    5 ++++-
>  1 files changed, 4 insertions(+), 1 deletions(-)
> 
> diff --git a/hw/e1000.c b/hw/e1000.c
> index 532efdc..677165f 100644
> --- a/hw/e1000.c
> +++ b/hw/e1000.c
> @@ -384,9 +384,12 @@ xmit_seg(E1000State *s)
>          } else       // UDP
>              cpu_to_be16wu((uint16_t *)(tp->data+css+4), len);
>          if (tp->sum_needed & E1000_TXD_POPTS_TXSM) {
> +            unsigned int phsum;
>              // add pseudo-header length before checksum calculation
>              sp = (uint16_t *)(tp->data + tp->tucso);
> -            cpu_to_be16wu(sp, be16_to_cpup(sp) + len);
> +            phsum = be16_to_cpup(sp) + len;
> +            phsum = (phsum >> 16) + (phsum & 0xffff);
> +            cpu_to_be16wu(sp, phsum);
>          }
>          tp->tso_frames++;
>      }

Reply via email to