Re: [PATCH net-next v2] net: ti: icssg-prueth: Add multicast filtering support

[MD Danish Anwar]

Hi Wojciech,

On 04/06/24 8:19 pm, Wojciech Drewek wrote:
> 
> 
> On 04.06.2024 13:44, MD Danish Anwar wrote:
>> Add multicast filtering support for ICSSG Driver.
>>
>> The driver will keep a copy of multicast addresses in emac->mcast_list.
>> This list will be kept in sync with the netdev list and to add / del
>> multicast address icssg_prueth_mac_add_mcast / icssg_prueth_mac_del_mcast
>> APIs will be called.
>>
>> To add a mac_address for a port, driver need to call icssg_fdb_add_del()
>> and pass the mac_address and BIT(port_id) to the API. The ICSSG firmware
>> will then configure the rules and allow filtering.
>>
>> If a mac_address is added to port0 and the same mac_address needs to be
>> added for port1, driver needs to pass BIT(port0) | BIT(port1) to the
>> icssg_fdb_add_del() API. If driver just pass BIT(port1) then the entry for
>> port0 will be overwritten / lost. This is a design constraint on the
>> firmware side.
>>
>> To overcome this in the driver, to add any mac_address for let's say portX
>> driver first checks if the same mac_address is already added for any other
>> port. If yes driver calls icssg_fdb_add_del() with BIT(portX) |
>> BIT(other_existing_port). If not, driver calls icssg_fdb_add_del() with
>> BIT(portX).
>>
>> The same thing is applicable for deleting mac_addresses as well. This
>> logic is in icssg_prueth_mac_add_mcast / icssg_prueth_mac_del_mcast APIs.
>>
>> Signed-off-by: MD Danish Anwar <danishanwar@xxxxxx>
>> ---
>> v1 -> v2:
>> *) Rebased on latest net-next/main.
>>
>> NOTE: This series can be applied cleanly on the tip of net-next/main. This
>> series doesn't depend on any other ICSSG driver related series that is
>> floating around in netdev.
>>
>> v1 https://lore.kernel.org/all/20240516091752.2969092-1-danishanwar@xxxxxx/
>>
>>  drivers/net/ethernet/ti/icssg/icssg_prueth.c | 50 ++++++++++++++++++--
>>  drivers/net/ethernet/ti/icssg/icssg_prueth.h |  3 ++
>>  2 files changed, 49 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/ti/icssg/icssg_prueth.c b/drivers/net/ethernet/ti/icssg/icssg_prueth.c
>> index 6e65aa0977d4..03dd49f0afb7 100644
>> --- a/drivers/net/ethernet/ti/icssg/icssg_prueth.c
>> +++ b/drivers/net/ethernet/ti/icssg/icssg_prueth.c
>> @@ -439,6 +439,37 @@ const struct icss_iep_clockops prueth_iep_clockops = {
>>  	.perout_enable = prueth_perout_enable,
>>  };
>>  
>> +static int icssg_prueth_mac_add_mcast(struct net_device *ndev, const u8 *addr)
>> +{
>> +	struct prueth_emac *emac = netdev_priv(ndev);
>> +	int port_mask = BIT(emac->port_id);
>> +
>> +	port_mask |= icssg_fdb_lookup(emac, addr, 0);
>> +	icssg_fdb_add_del(emac, addr, 0, port_mask, true);
>> +	icssg_vtbl_modify(emac, 0, port_mask, port_mask, true);
>> +
>> +	return 0;
>> +}
>> +
>> +static int icssg_prueth_mac_del_mcast(struct net_device *ndev, const u8 *addr)
>> +{
>> +	struct prueth_emac *emac = netdev_priv(ndev);
>> +	int port_mask = BIT(emac->port_id);
>> +	int other_port_mask;
>> +
>> +	other_port_mask = port_mask ^ icssg_fdb_lookup(emac, addr, 0);
>> +
>> +	icssg_fdb_add_del(emac, addr, 0, port_mask, false);
>> +	icssg_vtbl_modify(emac, 0, port_mask, port_mask, false);
>> +
>> +	if (other_port_mask) {
>> +		icssg_fdb_add_del(emac, addr, 0, other_port_mask, true);
>> +		icssg_vtbl_modify(emac, 0, other_port_mask, other_port_mask, true);
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>>  /**
>>   * emac_ndo_open - EMAC device open
>>   * @ndev: network adapter device
>> @@ -547,6 +578,8 @@ static int emac_ndo_open(struct net_device *ndev)
>>  
>>  	prueth->emacs_initialized++;
>>  
>> +	__hw_addr_init(&emac->mcast_list);
>> +
>>  	queue_work(system_long_wq, &emac->stats_work.work);
>>  
>>  	return 0;
>> @@ -599,6 +632,9 @@ static int emac_ndo_stop(struct net_device *ndev)
>>  
>>  	icssg_class_disable(prueth->miig_rt, prueth_emac_slice(emac));
>>  
>> +	__dev_mc_unsync(ndev, icssg_prueth_mac_del_mcast);
>> +	__hw_addr_init(&emac->mcast_list);
>> +
>>  	atomic_set(&emac->tdown_cnt, emac->tx_ch_num);
>>  	/* ensure new tdown_cnt value is visible */
>>  	smp_mb__after_atomic();
>> @@ -675,10 +711,15 @@ static void emac_ndo_set_rx_mode_work(struct work_struct *work)
>>  		return;
>>  	}
>>  
>> -	if (!netdev_mc_empty(ndev)) {
>> -		emac_set_port_state(emac, ICSSG_EMAC_PORT_MC_FLOODING_ENABLE);
>> -		return;
>> -	}
>> +	/* make a mc list copy */
>> +
>> +	netif_addr_lock_bh(ndev);
>> +	__hw_addr_sync(&emac->mcast_list, &ndev->mc, ndev->addr_len);
>> +	netif_addr_unlock_bh(ndev);
>> +
>> +	__hw_addr_sync_dev(&emac->mcast_list, ndev,
>> +			   icssg_prueth_mac_add_mcast,
>> +			   icssg_prueth_mac_del_mcast);
> 
> Is there a reason __dev_mc_sync can't be used here?
> net_device has mc list already, no need to keep your own list
> in prueth_emac.

I checked the code and __dev_mc_sync() can be used here. No need for
keeping local list as you pointed out. I will drop this and re post the
patch.

Thanks for the review.

> 
>>  }
>>  
>>  /**
>> @@ -767,6 +808,7 @@ static int prueth_netdev_init(struct prueth *prueth,
>>  	SET_NETDEV_DEV(ndev, prueth->dev);
>>  	spin_lock_init(&emac->lock);
>>  	mutex_init(&emac->cmd_lock);
>> +	__hw_addr_init(&emac->mcast_list);
>>  
>>  	emac->phy_node = of_parse_phandle(eth_node, "phy-handle", 0);
>>  	if (!emac->phy_node && !of_phy_is_fixed_link(eth_node)) {
>> diff --git a/drivers/net/ethernet/ti/icssg/icssg_prueth.h b/drivers/net/ethernet/ti/icssg/icssg_prueth.h
>> index 5eeeccb73665..2bfda26b5901 100644
>> --- a/drivers/net/ethernet/ti/icssg/icssg_prueth.h
>> +++ b/drivers/net/ethernet/ti/icssg/icssg_prueth.h
>> @@ -155,6 +155,9 @@ struct prueth_emac {
>>  	unsigned int tx_ts_enabled : 1;
>>  	unsigned int half_duplex : 1;
>>  
>> +	/* List for storing multicast addresses */
>> +	struct netdev_hw_addr_list mcast_list;
>> +
>>  	/* DMA related */
>>  	struct prueth_tx_chn tx_chns[PRUETH_MAX_TX_QUEUES];
>>  	struct completion tdown_complete;
>>
>> base-commit: 2589d668e1a6ebe85329f1054cdad13647deac06

-- 
Thanks and Regards,
Danish