From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vlad Yasevich Subject: [PATCH v2 2/4] net: Allow for more then a single subclass for netif_addr_lock Date: Fri, 16 May 2014 13:18:15 -0400 Message-ID: <1400260697-32601-3-git-send-email-vyasevic@redhat.com> References: <1400260697-32601-1-git-send-email-vyasevic@redhat.com> Cc: dingtianhong@huawei.com, kaber@trash.net, vfalico@gmail.com, jiri@resnulli.us, Vlad Yasevich To: netdev@vger.kernel.org Return-path: Received: from mx1.redhat.com ([209.132.183.28]:35001 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756924AbaEPRSp (ORCPT ); Fri, 16 May 2014 13:18:45 -0400 In-Reply-To: <1400260697-32601-1-git-send-email-vyasevic@redhat.com> Sender: netdev-owner@vger.kernel.org List-ID: Currently netif_addr_lock_nested assumes that there can be only a single nesting level between 2 devices. However, if we have multiple devices of the same type stacked, this fails. For example: eth0 <-- vlan0.10 <-- vlan0.10.20 A more complicated configuration may stack more then one type of device in different order. Ex: eth0 <-- vlan0.10 <-- macvlan0 <-- vlan1.10.20 <-- macvlan1 This patch adds an ndo_* function that allows each stackable device to report its nesting level. If the device doesn't provide this function default subclass of 1 is used. Signed-off-by: Vlad Yasevich --- include/linux/netdevice.h | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e32d313..edafa7d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1144,6 +1144,7 @@ struct net_device_ops { netdev_tx_t (*ndo_dfwd_start_xmit) (struct sk_buff *skb, struct net_device *dev, void *priv); + int (*ndo_get_lock_subclass)(struct net_device *dev); }; /** @@ -2950,7 +2951,12 @@ static inline void netif_addr_lock(struct net_device *dev) static inline void netif_addr_lock_nested(struct net_device *dev) { - spin_lock_nested(&dev->addr_list_lock, SINGLE_DEPTH_NESTING); + int subclass = SINGLE_DEPTH_NESTING; + + if (dev->netdev_ops->ndo_get_lock_subclass) + subclass = dev->netdev_ops->ndo_get_lock_subclass(dev); + + spin_lock_nested(&dev->addr_list_lock, subclass); } static inline void netif_addr_lock_bh(struct net_device *dev) -- 1.9.0