Re: [dpdk-dev] [PATCH v2] net/bonding: fix socket id check

[Ferruh Yigit <ferruh.yigit@intel.com>]

On 4/27/2021 10:16 AM, Min Hu (Connor) wrote:
> From: Chengchang Tang <tangchengchang@huawei.com>
> 
> The socket ID entered by user is cast to an unsigned integer. However,
> the value may be an illegal negative value, which may cause some
> problems. In this case, an error should be returned.
> 
> In addition, the socket ID may be an invalid positive number, which is
> also processed in this patch.
> 
> Fixes: 2efb58cbab6e ("bond: new link bonding library")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> ---
> v2:
> * fixed socket id type.
> ---
>  drivers/net/bonding/rte_eth_bond_args.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/bonding/rte_eth_bond_args.c b/drivers/net/bonding/rte_eth_bond_args.c
> index 8c5f90d..977f3fe 100644
> --- a/drivers/net/bonding/rte_eth_bond_args.c
> +++ b/drivers/net/bonding/rte_eth_bond_args.c
> @@ -207,13 +207,13 @@ bond_ethdev_parse_socket_id_kvarg(const char *key __rte_unused,
>  		return -1;
>  
>  	errno = 0;
> -	socket_id = (uint8_t)strtol(value, &endptr, 10);
> +	socket_id = (int)strtol(value, &endptr, 10);

Already provided some comment to v1, why it is better to do checks first and
cast later.

>  	if (*endptr != 0 || errno != 0)
>  		return -1;
>  
>  	/* validate socket id value */
> -	if (socket_id >= 0) {
> -		*(uint8_t *)extra_args = (uint8_t)socket_id;
> +	if (socket_id >= 0 && socket_id < RTE_MAX_NUMA_NODES) {
> +		*(int *)extra_args = socket_id;

The provided 'extra_args' is 'uint8_t', we can't just cast it to "int *" and
assign to it.