On 8/24/2018 12:15 PM, Alejandro Lucero wrote:
> Current code assumes a MAC change can occur when the port has been
> started. In fact, there are some NICs which require this port state
> for being successful, but other NICs not always support MAC change
> in that case.
>
> This patch supports a new device flag for a device advertising this
> limitation, and if the flag is set, the MAC is changed before the
> port starts.
>
> Fixes: af75078fece3 ("first public release")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Alejandro Lucero <alejandro.luc...@netronome.com>
<...>
> @@ -2839,6 +2841,10 @@ int rte_eth_dev_mac_addr_add(uint16_t port_id, struct
> ether_addr *mac_addr,
> /**
> * Set the default MAC address.
> *
> + * A NIC not supporting MAC change after started should set
> + * RTE_ETH_DEV_NOLIVE_MAC_ADDR and this function should check such a flag
> + * and NIC state.
> + *
Only rte_eth_dev_start() API effected from this change, API behavior changes
based on if PMD provides this flag or not, I was thinking to document this in
rte_eth_dev_start(), something like:
"Driver RTE_ETH_DEV_NOLIVE_MAC_ADDR flag cause MAC address to be set before
start dev_ops"
As you mentioned in cover letter, rte_eth_dev_mac_addr_add() will return an
error if not supported, this is not changed with RTE_ETH_DEV_NOLIVE_MAC_ADDR
flag, so I think no need to add this comment.
