igb: reduce CPU0 latency when updating statistics

This change is based off of the work and suggestion of Jan Jablonsky
<jan.jablonsky@thalesgroup.com>.

The Watchdog workqueue in igb driver is scheduled every 2s for each
network interface. That includes updating a statistics protected by
spinlock. Function igb_update_stats in this case will be protected
against preemption. According to number of a statistics registers
(cca 60), processing this function might cause additional cpu load
 on CPU0.

In case of statistics spinlock may be replaced with mutex, which
reduce latency on CPU0.

CC: Bernhard Kaindl  <bernhard.kaindl@thalesgroup.com>
CC: Jan Jablonsky <jan.jablonsky@thalesgroup.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h
index ca54e26..fe1592a 100644 (file)
--- a/drivers/net/ethernet/intel/igb/igb.h
+++ b/drivers/net/ethernet/intel/igb/igb.h
@@ -515,7 +515,7 @@ struct igb_adapter {
        /* OS defined structs */
        struct pci_dev *pdev;
 
-       spinlock_t stats64_lock;
+       struct mutex stats64_lock;
        struct rtnl_link_stats64 stats64;
 
        /* structs defined in e1000_hw.h */

diff --git a/drivers/net/ethernet/intel/igb/igb_ethtool.c b/drivers/net/ethernet/intel/igb/igb_ethtool.c
index c576710..7426060 100644 (file)
--- a/drivers/net/ethernet/intel/igb/igb_ethtool.c
+++ b/drivers/net/ethernet/intel/igb/igb_ethtool.c
@@ -2295,7 +2295,7 @@ static void igb_get_ethtool_stats(struct net_device *netdev,
        int i, j;
        char *p;
 
-       spin_lock(&adapter->stats64_lock);
+       mutex_lock(&adapter->stats64_lock);
        igb_update_stats(adapter);
 
        for (i = 0; i < IGB_GLOBAL_STATS_LEN; i++) {
@@ -2338,7 +2338,7 @@ static void igb_get_ethtool_stats(struct net_device *netdev,
                } while (u64_stats_fetch_retry_irq(&ring->rx_syncp, start));
                i += IGB_RX_QUEUE_STATS_LEN;
        }
-       spin_unlock(&adapter->stats64_lock);
+       mutex_unlock(&adapter->stats64_lock);
 }
 
 static void igb_get_strings(struct net_device *netdev, u32 stringset, u8 *data)

diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 453ae1d..bb4f3f6 100644 (file)
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -2203,9 +2203,9 @@ void igb_down(struct igb_adapter *adapter)
        del_timer_sync(&adapter->phy_info_timer);
 
        /* record the stats before reset*/
-       spin_lock(&adapter->stats64_lock);
+       mutex_lock(&adapter->stats64_lock);
        igb_update_stats(adapter);
-       spin_unlock(&adapter->stats64_lock);
+       mutex_unlock(&adapter->stats64_lock);
 
        adapter->link_speed = 0;
        adapter->link_duplex = 0;
@@ -3840,7 +3840,7 @@ static int igb_sw_init(struct igb_adapter *adapter)
        adapter->min_frame_size = ETH_ZLEN + ETH_FCS_LEN;
 
        spin_lock_init(&adapter->nfc_lock);
-       spin_lock_init(&adapter->stats64_lock);
+       mutex_init(&adapter->stats64_lock);
 #ifdef CONFIG_PCI_IOV
        switch (hw->mac.type) {
        case e1000_82576:
@@ -5406,9 +5406,9 @@ no_wait:
                }
        }
 
-       spin_lock(&adapter->stats64_lock);
+       mutex_lock(&adapter->stats64_lock);
        igb_update_stats(adapter);
-       spin_unlock(&adapter->stats64_lock);
+       mutex_unlock(&adapter->stats64_lock);
 
        for (i = 0; i < adapter->num_tx_queues; i++) {
                struct igb_ring *tx_ring = adapter->tx_ring[i];
@@ -6235,10 +6235,10 @@ static void igb_get_stats64(struct net_device *netdev,
 {
        struct igb_adapter *adapter = netdev_priv(netdev);
 
-       spin_lock(&adapter->stats64_lock);
+       mutex_lock(&adapter->stats64_lock);
        igb_update_stats(adapter);
        memcpy(stats, &adapter->stats64, sizeof(*stats));
-       spin_unlock(&adapter->stats64_lock);
+       mutex_unlock(&adapter->stats64_lock);
 }
 
 /**