Skip to contentSkip to conversation

Bonding Overriding Configuration logic restored.

PATCH In review Anton Nayshtut 3 participants 3 messages Mar 29, 2015 – Mar 31, 2015

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Bonding Overriding Configuration logic restored.

Before commit 3900f29021f0bc7fe9815aa32f1a993b7dfdd402 ("bonding: slight optimizztion for bond_slave_override()") the override logic was to send packets with non-zero queue_id through the slave with corresponding queue_id, under two conditions only - if the slave can transmit and it's up.

The above mentioned commit changed this logic by introducing an additional condition - whether the bond is active (indirectly, using the slave_can_tx and later - bond_is_active_slave), that prevents the user from implementing more complex policies according to the Documentation/networking/bonding.txt.

---
 drivers/net/bonding/bond_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/bonding/bond_main.c +2 -1
index 0dceba1..68ad39a 100644
@@ -3797,7 +3797,8 @@ static inline int bond_slave_override(struct bonding *bond,
 	/* Find out if any slaves have the same mapping as this skb. */
 	bond_for_each_slave_rcu(bond, slave, iter) {
 		if (slave->queue_id == skb->queue_mapping) {
-			if (bond_slave_can_tx(slave)) {
+			if (bond_slave_is_up(slave) &&
+			    slave->link == BOND_LINK_UP) {
 				bond_dev_queue_xmit(bond, skb, slave->dev);
 				return 0;
 			}
Signature · 1 line
1.9.1
#1 lore raw
Quoted 9 lines · Anton Nayshtut

Before commit 3900f29021f0bc7fe9815aa32f1a993b7dfdd402 ("bonding: slight optimizztion for bond_slave_override()") the override logic was to send packets with non-zero queue_id through the slave with corresponding queue_id, under two conditions only - if the slave can transmit and it's up.

The above mentioned commit changed this logic by introducing an additional condition - whether the bond is active (indirectly, using the slave_can_tx and later - bond_is_active_slave), that prevents the user from implementing more complex policies according to the Documentation/networking/bonding.txt.

Yes, this does. Good catch.

Quoted 2 lines
Quoted 21 lines
---
 drivers/net/bonding/bond_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/bonding/bond_main.c +2 -1
index 0dceba1..68ad39a 100644
@@ -3797,7 +3797,8 @@ static inline int bond_slave_override(struct bonding *bond,
 	/* Find out if any slaves have the same mapping as this skb. */
 	bond_for_each_slave_rcu(bond, slave, iter) {
 		if (slave->queue_id == skb->queue_mapping) {
-			if (bond_slave_can_tx(slave)) {
+			if (bond_slave_is_up(slave) &&
+			    slave->link == BOND_LINK_UP) {
 				bond_dev_queue_xmit(bond, skb, slave->dev);
 				return 0;
 			}
Signature · 1 line
1.9.1
maintainer
#2 lore raw
In reply to Anton Nayshtut

From: Anton Nayshtut <[email protected]>
Date: Sun, 29 Mar 2015 14:20:25 +0300

Quoted 5 lines

Before commit 3900f29021f0bc7fe9815aa32f1a993b7dfdd402 ("bonding: slight optimizztion for bond_slave_override()") the override logic was to send packets with non-zero queue_id through the slave with corresponding queue_id, under two conditions only - if the slave can transmit and it's up.

Quoted 7 lines

The above mentioned commit changed this logic by introducing an additional condition - whether the bond is active (indirectly, using the slave_can_tx and later - bond_is_active_slave), that prevents the user from implementing more complex policies according to the Documentation/networking/bonding.txt.

Applied and queued up for -stable, thanks.