On 11/20/2018 10:26 AM, Tiago Lam wrote:
> Instead of re-iterating through kvlist just to parse the
> ETH_AF_PACKET_IFACE_ARG argument in rte_pmd_init_internals(), we now use
> the already existing iteration in rte_eth_from_packet() to parse and
> validate the ETH_AF_PACKET_IFACE_ARG argument.
+1
>
> This will be useful for a later commit, which needs to access the
> interface name to get the underlying configured MTU.
>
> Signed-off-by: Tiago Lam <tiago....@intel.com>
> ---
>
> v2: Fix checkpatches.sh and check-git-log.sh warnings.
>
> ---
> drivers/net/af_packet/rte_eth_af_packet.c | 77
> ++++++++++++++++---------------
> 1 file changed, 41 insertions(+), 36 deletions(-)
>
> diff --git a/drivers/net/af_packet/rte_eth_af_packet.c
> b/drivers/net/af_packet/rte_eth_af_packet.c
> index 264cfc0..8d749a2 100644
> --- a/drivers/net/af_packet/rte_eth_af_packet.c
> +++ b/drivers/net/af_packet/rte_eth_af_packet.c
> @@ -540,15 +540,12 @@ rte_pmd_init_internals(struct rte_vdev_device *dev,
> unsigned int qdisc_bypass,
> struct pmd_internals **internals,
> struct rte_eth_dev **eth_dev,
> - struct rte_kvargs *kvlist)
> + const char *ifname)
Please align other parameter formatting.
<...>
> @@ -877,6 +856,32 @@ rte_eth_from_packet(struct rte_vdev_device *dev,
> }
> continue;
> }
> + if (strstr(pair->key, ETH_AF_PACKET_IFACE_ARG) != NULL) {
> + ifname = pair->value;
> + if (strlen(ifname) == 0) {
> + RTE_LOG(ERR, PMD,
> + "%s: invalid iface value\n",
> + name);
> + return -1;
> + }
> +
> + continue;
> + }
Indeed instead of accessing kvargs internal pair->values, this should call
rte_kvargs_process() but that is for another patch to fix all occurrences.
> + }
> +
> + if (ifname == NULL) {
> + RTE_LOG(ERR, PMD,
> + "%s: no interface specified for AF_PACKET ethdev\n",
> + name);
> + return -1;
> + }
> +
> + ifnamelen = strlen(ifname);
I am aware this is copy-paste but since update, perhaps we can use strnlen()
instead.
