[dpdk-dev] [PATCH] drivers/bonding: fix bond mac address reset

[dpdk-dev] [PATCH] drivers/bonding: fix bond mac address reset

[Kiran Kumar]

Currently when resetting bond mac address, we are getting the
persisted mac address from slave info considering primary port
as index. But we need to compare the port id from slave info
with the primary port to get the primary slave index and get
the persisted mac address. Without this fix, persisted mac addr
will be zero and rte_eth_dev_default_mac_addr_set will fail.

Fixes: a45b288ef21a ("bond: support link status polling")

Signed-off-by: Kiran Kumar <kiran.kokkilagadda@caviumnetworks.com>
---
 drivers/net/bonding/rte_eth_bond_api.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/net/bonding/rte_eth_bond_api.c b/drivers/net/bonding/rte_eth_bond_api.c
index d558df8b9..d1a110ec9 100644
--- a/drivers/net/bonding/rte_eth_bond_api.c
+++ b/drivers/net/bonding/rte_eth_bond_api.c
@@ -716,9 +716,21 @@ rte_eth_bond_mac_address_reset(uint16_t bonded_port_id)
 	internals->user_defined_mac = 0;
 
 	if (internals->slave_count > 0) {
+		int slave_port;
+		/* Get the primary slave location based on the primary port
+		 * number as, while slave_add(), we will keep the primary
+		 * slave based on slave_count,but not based on the primary port.
+		 */
+		for (slave_port = 0; slave_port < internals->slave_count;
+		     slave_port++) {
+			if (internals->slaves[slave_port].port_id ==
+			    internals->primary_port)
+				break;
+		}
+
 		/* Set MAC Address of Bonded Device */
 		if (mac_address_set(bonded_eth_dev,
-				&internals->slaves[internals->primary_port].persisted_mac_addr)
+			&internals->slaves[slave_port].persisted_mac_addr)
 				!= 0) {
 			RTE_BOND_LOG(ERR, "Failed to set MAC address on bonded device");
 			return -1;
-- 
2.17.0

Re: [dpdk-dev] [PATCH] drivers/bonding: fix bond mac address reset

[Ferruh Yigit]

On 5/23/2018 10:11 AM, Kiran Kumar wrote:
> Currently when resetting bond mac address, we are getting the
> persisted mac address from slave info considering primary port
> as index. But we need to compare the port id from slave info
> with the primary port to get the primary slave index and get
> the persisted mac address. Without this fix, persisted mac addr
> will be zero and rte_eth_dev_default_mac_addr_set will fail.
> 
> Fixes: a45b288ef21a ("bond: support link status polling")
> 
> Signed-off-by: Kiran Kumar <kiran.kokkilagadda@caviumnetworks.com>

Hi Declan, Radu, Chas,

If there is no objection, the patch will be merged as it is.
Can you please check?

Thanks,
ferruh

Re: [dpdk-dev] [PATCH] drivers/bonding: fix bond mac address reset

[Chas Williams]

On Thu, Jun 14, 2018 at 12:49 PM Ferruh Yigit <ferruh.yigit@intel.com>
wrote:

> On 5/23/2018 10:11 AM, Kiran Kumar wrote:
> > Currently when resetting bond mac address, we are getting the
> > persisted mac address from slave info considering primary port
> > as index. But we need to compare the port id from slave info
> > with the primary port to get the primary slave index and get
> > the persisted mac address. Without this fix, persisted mac addr
> > will be zero and rte_eth_dev_default_mac_addr_set will fail.
> >
> > Fixes: a45b288ef21a ("bond: support link status polling")
> >
> > Signed-off-by: Kiran Kumar <kiran.kokkilagadda@caviumnetworks.com>
>
> Hi Declan, Radu, Chas,
>
> If there is no objection, the patch will be merged as it is.
> Can you please check?
>

No objections. This patch looks correct.


>
> Thanks,
> ferruh
>

Re: [dpdk-dev] [PATCH] drivers/bonding: fix bond mac address reset

[Ferruh Yigit]

On 6/14/2018 9:38 PM, Chas Williams wrote:
> 
> 
> On Thu, Jun 14, 2018 at 12:49 PM Ferruh Yigit <ferruh.yigit@intel.com
> <mailto:ferruh.yigit@intel.com>> wrote:
> 
>     On 5/23/2018 10:11 AM, Kiran Kumar wrote:
>     > Currently when resetting bond mac address, we are getting the
>     > persisted mac address from slave info considering primary port
>     > as index. But we need to compare the port id from slave info
>     > with the primary port to get the primary slave index and get
>     > the persisted mac address. Without this fix, persisted mac addr
>     > will be zero and rte_eth_dev_default_mac_addr_set will fail.
>     >
>     > Fixes: a45b288ef21a ("bond: support link status polling")
>     >
>     > Signed-off-by: Kiran Kumar <kiran.kokkilagadda@caviumnetworks.com
>     <mailto:kiran.kokkilagadda@caviumnetworks.com>>
> 
>     Hi Declan, Radu, Chas,
> 
>     If there is no objection, the patch will be merged as it is.
>     Can you please check?
> 
> 
> No objections. This patch looks correct.

Applied to dpdk-next-net/master, thanks.