Re: [dpdk-dev] [PATCH v3 5/5] net/tap: fix netlink socket file descriptor check before close

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

On 4/7/2020 5:23 AM, wangyunjian wrote:
> From: Yunjian Wang <wangyunjian@huawei.com>
> 
> The "internals->nlsk_fd" is possibly a zero value. Error comparison
> should check for negative fd.
> 
> Fixes: de96fe68ae95 ("net/tap: add basic flow API patterns and actions")
> CC: stable@dpdk.org
> 
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> ---
>  drivers/net/tap/rte_eth_tap.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
> index 88cd615e0..ff149342b 100644
> --- a/drivers/net/tap/rte_eth_tap.c
> +++ b/drivers/net/tap/rte_eth_tap.c
> @@ -2442,7 +2442,7 @@ rte_pmd_tap_remove(struct rte_vdev_device *dev)
>  	TAP_LOG(DEBUG, "Closing %s Ethernet device on numa %u",
>  		tuntap_types[internals->type], rte_socket_id());
>  
> -	if (internals->nlsk_fd) {
> +	if (internals->nlsk_fd >= 0) {

I guess it is unlikely socket() will return 0 as fd, but technically it seems
possible.

'tap_nl_init()' either returns '-1' for error or a valid fd, and this has been
checked as "pmd->nlsk_fd == -1" in other locations, what about doing similar here:

"if (internals->nlsk_fd != -1)"

>  		tap_flow_flush(eth_dev, NULL);
>  		tap_flow_implicit_flush(internals, NULL);
>  		tap_nl_final(internals->nlsk_fd);
>