diff mbox

[net-next-2.6] bonding: set primary param via sysfs

Message ID 20090918121321.GB2801@psychotron.redhat.com
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Jiri Pirko Sept. 18, 2009, 12:13 p.m. UTC
Primary module parameter passed to bonding is pernament. That means if you
release the primary slave and enslave it again, it becomes the primary slave
again. But if you set primary slave via sysfs, the primary slave is only set
once and it's not remembered in bond->params structure. Therefore the setting is
lost after releasing the primary slave. This simple one-liner fixes this.

Signed-off-by: Jiri Pirko <jpirko@redhat.com>

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Nicolas de Peslo√ľan Sept. 18, 2009, 8:21 p.m. UTC | #1
Jiri Pirko wrote:
> Primary module parameter passed to bonding is pernament. That means if you

typo in "permanent".

> release the primary slave and enslave it again, it becomes the primary slave
> again. But if you set primary slave via sysfs, the primary slave is only set
> once and it's not remembered in bond->params structure. Therefore the setting is
> lost after releasing the primary slave. This simple one-liner fixes this.

You patch also has the side effect of fixing this strange behavior:

If you move the primary slave from one bond device to another one, it becomes 
the primary for this other bond device, ignoring what you might have set as the 
primary for this other bond device.

#modprobe bonding mode=active-backup primary=eth0
#echo +eth0 > /sys/class/net/bond0/bonding/slaves
#cat /sys/class/net/bond0/bonding/primary
eth0
#echo -eth0 > /sys/class/net/bond0/bonding/slaves

#echo +eth1 > /sys/class/net/bond1/bonding/slaves
#echo eth1 > /sys/class/net/bond1/bonding/primary
#cat /sys/class/net/bond1/bonding/primary
eth1
#echo +eth0 > /sys/class/net/bond1/bonding/slaves
#cat /sys/class/net/bond1/bonding/primary
eth0

=> Primary just changed, for no good reason.

Can someone imagine that some current configurations rely on this incredible 
side effect ?

	Nicolas.

> Signed-off-by: Jiri Pirko <jpirko@redhat.com>
> 
> diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
> index 6044e12..ff449de 100644
> --- a/drivers/net/bonding/bond_sysfs.c
> +++ b/drivers/net/bonding/bond_sysfs.c
> @@ -1182,6 +1182,7 @@ static ssize_t bonding_store_primary(struct device *d,
>  				       ": %s: Setting %s as primary slave.\n",
>  				       bond->dev->name, slave->dev->name);
>  				bond->primary_slave = slave;
> +				strcpy(bond->params.primary, slave->dev->name);
>  				bond_select_active_slave(bond);
>  				goto out;
>  			}

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jay Vosburgh Sept. 18, 2009, 11:02 p.m. UTC | #2
Jiri Pirko <jpirko@redhat.com> wrote:

>Primary module parameter passed to bonding is pernament. That means if you
>release the primary slave and enslave it again, it becomes the primary slave
>again. But if you set primary slave via sysfs, the primary slave is only set
>once and it's not remembered in bond->params structure. Therefore the setting is
>lost after releasing the primary slave. This simple one-liner fixes this.
>
>Signed-off-by: Jiri Pirko <jpirko@redhat.com>
>
>diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
>index 6044e12..ff449de 100644
>--- a/drivers/net/bonding/bond_sysfs.c
>+++ b/drivers/net/bonding/bond_sysfs.c
>@@ -1182,6 +1182,7 @@ static ssize_t bonding_store_primary(struct device *d,
> 				       ": %s: Setting %s as primary slave.\n",
> 				       bond->dev->name, slave->dev->name);
> 				bond->primary_slave = slave;
>+				strcpy(bond->params.primary, slave->dev->name);
> 				bond_select_active_slave(bond);
> 				goto out;
> 			}

	Looks good, fixes another bug I'd not heard of previously.

Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller Oct. 1, 2009, 9:39 p.m. UTC | #3
From: Jiri Pirko <jpirko@redhat.com>
Date: Fri, 18 Sep 2009 14:13:22 +0200

> Primary module parameter passed to bonding is pernament. That means if you
> release the primary slave and enslave it again, it becomes the primary slave
> again. But if you set primary slave via sysfs, the primary slave is only set
> once and it's not remembered in bond->params structure. Therefore the setting is
> lost after releasing the primary slave. This simple one-liner fixes this.
> 
> Signed-off-by: Jiri Pirko <jpirko@redhat.com>

Applied, thanks Jiri.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
index 6044e12..ff449de 100644
--- a/drivers/net/bonding/bond_sysfs.c
+++ b/drivers/net/bonding/bond_sysfs.c
@@ -1182,6 +1182,7 @@  static ssize_t bonding_store_primary(struct device *d,
 				       ": %s: Setting %s as primary slave.\n",
 				       bond->dev->name, slave->dev->name);
 				bond->primary_slave = slave;
+				strcpy(bond->params.primary, slave->dev->name);
 				bond_select_active_slave(bond);
 				goto out;
 			}