[dpdk-stable] [PATCH] examples/vhost: fix potential overflow in args process

[Jiang, Cheng1]

Hi,

> -----Original Message-----
> From: David Marchand <david.marchand at redhat.com>
> Sent: Monday, April 26, 2021 4:25 PM
> To: Jiang, Cheng1 <cheng1.jiang at intel.com>
> Cc: Maxime Coquelin <maxime.coquelin at redhat.com>; Xia, Chenbo
> <chenbo.xia at intel.com>; dev <dev at dpdk.org>; dpdk stable
> <stable at dpdk.org>
> Subject: Re: [dpdk-stable] [PATCH] examples/vhost: fix potential overflow in
> args process
> 
> On Mon, Apr 19, 2021 at 7:59 AM Cheng Jiang <Cheng1.jiang at intel.com>
> wrote:
> >
> > Add args length check to fix potential overflow issue.
> >
> > Coverity issue: 363741
> > Fixes: 965b06f0358 ("examples/vhost: enhance getopt_long usage")
> > Cc: stable at dpdk.org
> 
> No need for Cc: stable since this issue only affects the current release.

Sure, sorry about that.

> 
> >
> > Signed-off-by: Cheng Jiang <Cheng1.jiang at intel.com>
> 
> It should indeed fix the coverity report but this limit does not make sense.
> Could you make dma_type point at either optarg (it should be fine as
> we only read this string)?
> 
> Something like (untested):
> 
> diff --git a/examples/vhost/main.c b/examples/vhost/main.c
> index 2ca7d98c58..158e5e9a8a 100644
> --- a/examples/vhost/main.c
> +++ b/examples/vhost/main.c
> @@ -55,9 +55,6 @@
> 
>  #define INVALID_PORT_ID 0xFF
> 
> -/* Maximum long option length for option parsing. */
> -#define MAX_LONG_OPT_SZ 64
> -
>  /* mask of enabled ports */
>  static uint32_t enabled_port_mask = 0;
> 
> @@ -97,7 +94,7 @@ static int builtin_net_driver;
> 
>  static int async_vhost_driver;
> 
> -static char dma_type[MAX_LONG_OPT_SZ];
> +static char *dma_type;
> 
>  /* Specify timeout (in useconds) between retries on RX. */
>  static uint32_t burst_rx_delay_time = BURST_RX_WAIT_US;
> @@ -201,7 +198,7 @@ struct vhost_bufftable
> *vhost_txbuff[RTE_MAX_LCORE
> * MAX_VHOST_DEVICE];
>  static inline int
>  open_dma(const char *value)
>  {
> -       if (strncmp(dma_type, "ioat", 4) == 0)
> +       if (dma_type != NULL && strncmp(dma_type, "ioat", 4) == 0)
>                 return open_ioat(value);
> 
>         return -1;
> @@ -669,7 +666,7 @@ us_vhost_parse_args(int argc, char **argv)
>                         break;
> 
>                 case OPT_DMA_TYPE_NUM:
> -                       strcpy(dma_type, optarg);
> +                       dma_type = optarg;
>                         break;
> 
>                 case OPT_DMAS_NUM:
> @@ -1472,7 +1469,7 @@ new_device(int vid)
>                 struct rte_vhost_async_features f;
>                 struct rte_vhost_async_channel_ops channel_ops;
> 
> -               if (strncmp(dma_type, "ioat", 4) == 0) {
> +               if (dma_type != NULL && strncmp(dma_type, "ioat", 4) == 0) {
>                         channel_ops.transfer_data = ioat_transfer_data_cb;
>                         channel_ops.check_completed_copies =
>                                 ioat_check_completed_copies_cb;
> 
> 
> 
> --
> David Marchand

I think that make sense. I'll send a new version according your comments.

Thanks.
Cheng