Patchwork [net-next] vmxnet3: fix ethtool ring buffer size setting

login
register
mail settings
Submitter Neil Horman
Date Feb. 22, 2013, 8:32 p.m.
Message ID <1361565144-18301-1-git-send-email-nhorman@tuxdriver.com>
Download mbox | patch
Permalink /patch/222639/
State Accepted
Delegated to: David Miller
Headers show

Comments

Neil Horman - Feb. 22, 2013, 8:32 p.m.
Noticed that vmxnet3's get_ringparam function was returning the summation of all
ring buffers on a NIC, rather than just the size of any one ring.  This causes
problems when a vmxnet3 instance has multiple queues, as ethtool, when setting
ring parameters, first gets the current ring parameters to set the existing
values in the set_ringparm commannd.  The result is, that unless both rx and tx
ring sizes are set in a single operation, which ever ring is not set will
silently have its ring count multiplied by the number of queues on the NIC until
it reaches a driver defined maxiumum value.

Fix it by not multiplying the rx and tx ring sizes by the number of queues in
the system, like every other driver.  Tested by myself successfully.

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Shreyas Bhatewara <sbhatewara@vmware.com>
CC: "VMware, Inc." <pv-drivers@vmware.com>
CC: "David S. Miller" <davem@davemloft.net>
---
 drivers/net/vmxnet3/vmxnet3_ethtool.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
David Miller - Feb. 25, 2013, 1:19 a.m.
From: Neil Horman <nhorman@tuxdriver.com>
Date: Fri, 22 Feb 2013 15:32:24 -0500

> Noticed that vmxnet3's get_ringparam function was returning the summation of all
> ring buffers on a NIC, rather than just the size of any one ring.  This causes
> problems when a vmxnet3 instance has multiple queues, as ethtool, when setting
> ring parameters, first gets the current ring parameters to set the existing
> values in the set_ringparm commannd.  The result is, that unless both rx and tx
> ring sizes are set in a single operation, which ever ring is not set will
> silently have its ring count multiplied by the number of queues on the NIC until
> it reaches a driver defined maxiumum value.
> 
> Fix it by not multiplying the rx and tx ring sizes by the number of queues in
> the system, like every other driver.  Tested by myself successfully.
> 
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>

Applied.
--
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

Patch

diff --git a/drivers/net/vmxnet3/vmxnet3_ethtool.c b/drivers/net/vmxnet3/vmxnet3_ethtool.c
index 9bc542b..a0feb17 100644
--- a/drivers/net/vmxnet3/vmxnet3_ethtool.c
+++ b/drivers/net/vmxnet3/vmxnet3_ethtool.c
@@ -448,10 +448,8 @@  vmxnet3_get_ringparam(struct net_device *netdev,
 	param->rx_mini_max_pending = 0;
 	param->rx_jumbo_max_pending = 0;
 
-	param->rx_pending = adapter->rx_queue[0].rx_ring[0].size *
-			    adapter->num_rx_queues;
-	param->tx_pending = adapter->tx_queue[0].tx_ring.size *
-			    adapter->num_tx_queues;
+	param->rx_pending = adapter->rx_queue[0].rx_ring[0].size;
+	param->tx_pending = adapter->tx_queue[0].tx_ring.size;
 	param->rx_mini_pending = 0;
 	param->rx_jumbo_pending = 0;
 }