* kernel, by Linus Torvalds and others.
*/
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
-
#include <linux/dcache.h>
#include <linux/etherdevice.h>
#include <linux/if.h>
#include <linux/list.h>
#include <linux/mutex.h>
#include <linux/percpu.h>
+#include <linux/rcupdate.h>
#include <linux/rtnetlink.h>
#include <linux/compat.h>
#include <linux/version.h>
static const struct vport_ops **vport_ops_list;
static int n_vport_types;
+/* Protected by RCU read lock for reading, RTNL lock for writing. */
static struct hlist_head *dev_table;
#define VPORT_HASH_BUCKETS 1024
-/* Both RTNL lock and vport_mutex need to be held when updating dev_table.
- *
- * If you use vport_locate and then perform some operations, you need to hold
- * one of these locks if you don't want the vport to be deleted out from under
- * you.
- *
- * If you get a reference to a vport through a datapath, it is protected
- * by RCU and you need to hold rcu_read_lock instead when reading.
- *
- * If multiple locks are taken, the hierarchy is:
- * 1. RTNL
- * 2. DP
- * 3. vport
- */
-static DEFINE_MUTEX(vport_mutex);
-
-/**
- * vport_lock - acquire vport lock
- *
- * Acquire global vport lock. See above comment about locking requirements
- * and specific function definitions. May sleep.
- */
-void vport_lock(void)
-{
- mutex_lock(&vport_mutex);
-}
-
-/**
- * vport_unlock - release vport lock
- *
- * Release lock acquired with vport_lock.
- */
-void vport_unlock(void)
-{
- mutex_unlock(&vport_mutex);
-}
-
-#define ASSERT_VPORT() \
-do { \
- if (unlikely(!mutex_is_locked(&vport_mutex))) { \
- pr_err("vport lock not held at %s (%d)\n", \
- __FILE__, __LINE__); \
- dump_stack(); \
- } \
-} while (0)
-
/**
* vport_init - initialize vport subsystem
*
*
* @name: name of port to find
*
- * Either RTNL or vport lock must be acquired before calling this function
- * and held while using the found port. See the locking comments at the
- * top of the file.
+ * Must be called with RTNL or RCU read lock.
*/
struct vport *vport_locate(const char *name)
{
struct vport *vport;
struct hlist_node *node;
- if (unlikely(!mutex_is_locked(&vport_mutex) && !rtnl_is_locked())) {
- pr_err("neither RTNL nor vport lock held in vport_locate\n");
- dump_stack();
- }
-
- rcu_read_lock();
-
- hlist_for_each_entry(vport, node, bucket, hash_node)
+ hlist_for_each_entry_rcu(vport, node, bucket, hash_node)
if (!strcmp(name, vport_get_name(vport)))
- goto out;
-
- vport = NULL;
-
-out:
- rcu_read_unlock();
- return vport;
-}
-
-static void register_vport(struct vport *vport)
-{
- hlist_add_head(&vport->hash_node, hash_bucket(vport_get_name(vport)));
-}
+ return vport;
-static void unregister_vport(struct vport *vport)
-{
- hlist_del(&vport->hash_node);
+ return NULL;
}
static void release_vport(struct kobject *kobj)
vport->dp = parms->dp;
vport->port_no = parms->port_no;
- atomic_set(&vport->sflow_pool, 0);
+ vport->upcall_pid = parms->upcall_pid;
vport->ops = ops;
/* Initialize kobject for bridge. This will be added as
vport->kobj.kset = NULL;
kobject_init(&vport->kobj, &brport_ktype);
- if (vport->ops->flags & VPORT_F_GEN_STATS) {
- vport->percpu_stats = alloc_percpu(struct vport_percpu_stats);
- if (!vport->percpu_stats)
- return ERR_PTR(-ENOMEM);
+ vport->percpu_stats = alloc_percpu(struct vport_percpu_stats);
+ if (!vport->percpu_stats)
+ return ERR_PTR(-ENOMEM);
- spin_lock_init(&vport->stats_lock);
- }
+ spin_lock_init(&vport->stats_lock);
return vport;
}
* @vport: vport to free
*
* Frees a vport allocated with vport_alloc() when it is no longer needed.
+ *
+ * The caller must ensure that an RCU grace period has passed since the last
+ * time @vport was in a datapath.
*/
void vport_free(struct vport *vport)
{
- if (vport->ops->flags & VPORT_F_GEN_STATS)
- free_percpu(vport->percpu_stats);
+ free_percpu(vport->percpu_stats);
kobject_put(&vport->kobj);
}
* @parms: Information about new vport.
*
* Creates a new vport with the specified configuration (which is dependent on
- * device type) and attaches it to a datapath. Both RTNL and vport locks must
- * be held.
+ * device type). RTNL lock must be held.
*/
struct vport *vport_add(const struct vport_parms *parms)
{
int i;
ASSERT_RTNL();
- ASSERT_VPORT();
for (i = 0; i < n_vport_types; i++) {
if (vport_ops_list[i]->type == parms->type) {
goto out;
}
- register_vport(vport);
+ hlist_add_head_rcu(&vport->hash_node,
+ hash_bucket(vport_get_name(vport)));
return vport;
}
}
}
/**
- * vport_del - delete existing vport device (for kernel callers)
+ * vport_del - delete existing vport device
*
* @vport: vport to delete.
*
* Detaches @vport from its datapath and destroys it. It is possible to fail
- * for reasons such as lack of memory. Both RTNL and vport locks must be held.
- */
-int vport_del(struct vport *vport)
-{
- ASSERT_RTNL();
- ASSERT_VPORT();
-
- unregister_vport(vport);
-
- return vport->ops->destroy(vport);
-}
-
-/**
- * vport_set_mtu - set device MTU (for kernel callers)
- *
- * @vport: vport on which to set MTU.
- * @mtu: New MTU.
- *
- * Sets the MTU of the given device. Some devices may not support setting the
- * MTU, in which case the result will always be -EOPNOTSUPP. RTNL lock must
- * be held.
+ * for reasons such as lack of memory. RTNL lock must be held.
*/
-int vport_set_mtu(struct vport *vport, int mtu)
+void vport_del(struct vport *vport)
{
ASSERT_RTNL();
- if (mtu < 68)
- return -EINVAL;
-
- if (vport->ops->set_mtu) {
- int ret;
-
- ret = vport->ops->set_mtu(vport, mtu);
+ hlist_del_rcu(&vport->hash_node);
- if (!ret && !is_internal_vport(vport))
- set_internal_devs_mtu(vport->dp);
-
- return ret;
- } else
- return -EOPNOTSUPP;
+ vport->ops->destroy(vport);
}
/**
}
/**
- * vport_set_stats - sets offset device stats (for kernel callers)
+ * vport_set_stats - sets offset device stats
*
* @vport: vport on which to set stats
* @stats: stats to set
* Provides a set of transmit, receive, and error stats to be added as an
* offset to the collect data when stats are retreived. Some devices may not
* support setting the stats, in which case the result will always be
- * -EOPNOTSUPP. RTNL lock must be held.
+ * -EOPNOTSUPP.
+ *
+ * Must be called with RTNL lock.
*/
-int vport_set_stats(struct vport *vport, struct rtnl_link_stats64 *stats)
+void vport_set_stats(struct vport *vport, struct ovs_vport_stats *stats)
{
ASSERT_RTNL();
- if (vport->ops->flags & VPORT_F_GEN_STATS) {
- spin_lock_bh(&vport->stats_lock);
- vport->offset_stats = *stats;
- spin_unlock_bh(&vport->stats_lock);
-
- return 0;
- } else
- return -EOPNOTSUPP;
+ spin_lock_bh(&vport->stats_lock);
+ vport->offset_stats = *stats;
+ spin_unlock_bh(&vport->stats_lock);
}
/**
*
* @vport: vport from which to retrieve the type.
*
- * Retrieves the type of the given device. Either RTNL lock or rcu_read_lock
- * must be held.
+ * Retrieves the type of the given device.
*/
-enum odp_vport_type vport_get_type(const struct vport *vport)
+enum ovs_vport_type vport_get_type(const struct vport *vport)
{
return vport->ops->type;
}
return NULL;
}
-static int vport_call_get_stats(struct vport *vport, struct rtnl_link_stats64 *stats)
-{
- int err;
-
- rcu_read_lock();
- err = vport->ops->get_stats(vport, stats);
- rcu_read_unlock();
-
- return err;
-}
-
/**
- * vport_get_stats - retrieve device stats (for kernel callers)
+ * vport_get_stats - retrieve device stats
*
* @vport: vport from which to retrieve the stats
* @stats: location to store stats
*
* Retrieves transmit, receive, and error stats for the given device.
+ *
+ * Must be called with RTNL lock or rcu_read_lock.
*/
-int vport_get_stats(struct vport *vport, struct rtnl_link_stats64 *stats)
+void vport_get_stats(struct vport *vport, struct ovs_vport_stats *stats)
{
int i;
- if (!(vport->ops->flags & VPORT_F_GEN_STATS))
- return vport_call_get_stats(vport, stats);
-
/* We potentially have 3 sources of stats that need to be
* combined: those we have collected (split into err_stats and
* percpu_stats), offset_stats from set_stats(), and device
- * error stats from get_stats() (for errors that happen
+ * error stats from netdev->get_stats() (for errors that happen
* downstream and therefore aren't reported through our
- * vport_record_error() function). */
+ * vport_record_error() function).
+ * Stats from first two sources are merged and reported by ovs over
+ * OVS_VPORT_ATTR_STATS.
+ * netdev-stats can be directly read over netlink-ioctl.
+ */
spin_lock_bh(&vport->stats_lock);
spin_unlock_bh(&vport->stats_lock);
- if (vport->ops->get_stats) {
- struct rtnl_link_stats64 dev_stats;
- int err;
-
- err = vport_call_get_stats(vport, &dev_stats);
- if (err)
- return err;
-
- stats->rx_errors += dev_stats.rx_errors;
- stats->tx_errors += dev_stats.tx_errors;
- stats->rx_dropped += dev_stats.rx_dropped;
- stats->tx_dropped += dev_stats.tx_dropped;
- stats->multicast += dev_stats.multicast;
- stats->collisions += dev_stats.collisions;
- stats->rx_length_errors += dev_stats.rx_length_errors;
- stats->rx_over_errors += dev_stats.rx_over_errors;
- stats->rx_crc_errors += dev_stats.rx_crc_errors;
- stats->rx_frame_errors += dev_stats.rx_frame_errors;
- stats->rx_fifo_errors += dev_stats.rx_fifo_errors;
- stats->rx_missed_errors += dev_stats.rx_missed_errors;
- stats->tx_aborted_errors += dev_stats.tx_aborted_errors;
- stats->tx_carrier_errors += dev_stats.tx_carrier_errors;
- stats->tx_fifo_errors += dev_stats.tx_fifo_errors;
- stats->tx_heartbeat_errors += dev_stats.tx_heartbeat_errors;
- stats->tx_window_errors += dev_stats.tx_window_errors;
- stats->rx_compressed += dev_stats.rx_compressed;
- stats->tx_compressed += dev_stats.tx_compressed;
- }
-
for_each_possible_cpu(i) {
const struct vport_percpu_stats *percpu_stats;
struct vport_percpu_stats local_stats;
stats->tx_bytes += local_stats.tx_bytes;
stats->tx_packets += local_stats.tx_packets;
}
-
- return 0;
}
/**
*
* @vport: vport from which to retrieve the flags
*
- * Retrieves the flags of the given device. Either RTNL lock or rcu_read_lock
- * must be held.
+ * Retrieves the flags of the given device.
+ *
+ * Must be called with RTNL lock or rcu_read_lock.
*/
unsigned vport_get_flags(const struct vport *vport)
{
*
* @vport: vport on which to check status.
*
- * Checks whether the given device is running. Either RTNL lock or
- * rcu_read_lock must be held.
+ * Checks whether the given device is running.
+ *
+ * Must be called with RTNL lock or rcu_read_lock.
*/
int vport_is_running(const struct vport *vport)
{
*
* @vport: vport from which to check status
*
- * Retrieves the RFC2863 operstate of the given device. Either RTNL lock or
- * rcu_read_lock must be held.
+ * Retrieves the RFC2863 operstate of the given device.
+ *
+ * Must be called with RTNL lock or rcu_read_lock.
*/
unsigned char vport_get_operstate(const struct vport *vport)
{
*
* Retrieves the system interface index of the given device or 0 if
* the device does not have one (in the case of virtual ports).
- * Returns a negative index on error. Either RTNL lock or
- * rcu_read_lock must be held.
+ * Returns a negative index on error.
+ *
+ * Must be called with RTNL lock or rcu_read_lock.
*/
int vport_get_ifindex(const struct vport *vport)
{
}
/**
- * vport_get_iflink - retrieve device system link index
- *
- * @vport: vport from which to retrieve index
- *
- * Retrieves the system link index of the given device. The link is the index
- * of the interface on which the packet will actually be sent. In most cases
- * this is the same as the ifindex but may be different for tunnel devices.
- * Returns a negative index on error. Either RTNL lock or rcu_read_lock must
- * be held.
- */
-int vport_get_iflink(const struct vport *vport)
-{
- if (vport->ops->get_iflink)
- return vport->ops->get_iflink(vport);
-
- /* If we don't have an iflink, use the ifindex. In most cases they
- * are the same. */
- return vport_get_ifindex(vport);
-}
-
-/**
- * vport_get_mtu - retrieve device MTU (for kernel callers)
+ * vport_get_mtu - retrieve device MTU
*
* @vport: vport from which to retrieve MTU
*
- * Retrieves the MTU of the given device. Either RTNL lock or rcu_read_lock
- * must be held.
+ * Retrieves the MTU of the given device. Returns 0 if @vport does not have an
+ * MTU (as e.g. some tunnels do not). Either RTNL lock or rcu_read_lock must
+ * be held.
*/
int vport_get_mtu(const struct vport *vport)
{
+ if (!vport->ops->get_mtu)
+ return 0;
return vport->ops->get_mtu(vport);
}
* @skb: sk_buff where options should be appended.
*
* Retrieves the configuration of the given device, appending an
- * %ODP_VPORT_ATTR_OPTIONS attribute that in turn contains nested
- * vport-specific attributes to @skb. Either RTNL lock or rcu_read_lock must
- * be held.
+ * %OVS_VPORT_ATTR_OPTIONS attribute that in turn contains nested
+ * vport-specific attributes to @skb.
*
* Returns 0 if successful, -EMSGSIZE if @skb has insufficient room, or another
- * negative error code if a real error occurred.
+ * negative error code if a real error occurred. If an error occurs, @skb is
+ * left unmodified.
+ *
+ * Must be called with RTNL lock or rcu_read_lock.
*/
int vport_get_options(const struct vport *vport, struct sk_buff *skb)
{
struct nlattr *nla;
- nla = nla_nest_start(skb, ODP_VPORT_ATTR_OPTIONS);
+ nla = nla_nest_start(skb, OVS_VPORT_ATTR_OPTIONS);
if (!nla)
return -EMSGSIZE;
if (vport->ops->get_options) {
int err = vport->ops->get_options(vport, skb);
- if (err)
+ if (err) {
+ nla_nest_cancel(skb, nla);
return err;
+ }
}
nla_nest_end(skb, nla);
*/
void vport_receive(struct vport *vport, struct sk_buff *skb)
{
- if (vport->ops->flags & VPORT_F_GEN_STATS) {
- struct vport_percpu_stats *stats;
+ struct vport_percpu_stats *stats;
- local_bh_disable();
- stats = per_cpu_ptr(vport->percpu_stats, smp_processor_id());
+ local_bh_disable();
+ stats = per_cpu_ptr(vport->percpu_stats, smp_processor_id());
- write_seqcount_begin(&stats->seqlock);
- stats->rx_packets++;
- stats->rx_bytes += skb->len;
- write_seqcount_end(&stats->seqlock);
+ write_seqcount_begin(&stats->seqlock);
+ stats->rx_packets++;
+ stats->rx_bytes += skb->len;
+ write_seqcount_end(&stats->seqlock);
- local_bh_enable();
- }
+ local_bh_enable();
if (!(vport->ops->flags & VPORT_F_FLOW))
OVS_CB(skb)->flow = NULL;
dp_process_received_packet(vport, skb);
}
-static inline unsigned packet_length(const struct sk_buff *skb)
-{
- unsigned length = skb->len - ETH_HLEN;
-
- if (skb->protocol == htons(ETH_P_8021Q))
- length -= VLAN_HLEN;
-
- return length;
-}
-
/**
* vport_send - send a packet on a device
*
*/
int vport_send(struct vport *vport, struct sk_buff *skb)
{
- int mtu;
- int sent;
-
- mtu = vport_get_mtu(vport);
- if (unlikely(packet_length(skb) > mtu && !skb_is_gso(skb))) {
- if (net_ratelimit())
- pr_warn("%s: dropped over-mtu packet: %d > %d\n",
- dp_name(vport->dp), packet_length(skb), mtu);
- goto error;
- }
-
- sent = vport->ops->send(vport, skb);
+ struct vport_percpu_stats *stats;
+ int sent = vport->ops->send(vport, skb);
- if (vport->ops->flags & VPORT_F_GEN_STATS && sent > 0) {
- struct vport_percpu_stats *stats;
+ local_bh_disable();
+ stats = per_cpu_ptr(vport->percpu_stats, smp_processor_id());
- local_bh_disable();
- stats = per_cpu_ptr(vport->percpu_stats, smp_processor_id());
+ write_seqcount_begin(&stats->seqlock);
+ stats->tx_packets++;
+ stats->tx_bytes += sent;
+ write_seqcount_end(&stats->seqlock);
- write_seqcount_begin(&stats->seqlock);
- stats->tx_packets++;
- stats->tx_bytes += sent;
- write_seqcount_end(&stats->seqlock);
-
- local_bh_enable();
- }
+ local_bh_enable();
return sent;
-
-error:
- kfree_skb(skb);
- vport_record_error(vport, VPORT_E_TX_DROPPED);
- return 0;
}
/**
*/
void vport_record_error(struct vport *vport, enum vport_err_type err_type)
{
- if (vport->ops->flags & VPORT_F_GEN_STATS) {
-
- spin_lock_bh(&vport->stats_lock);
+ spin_lock_bh(&vport->stats_lock);
- switch (err_type) {
- case VPORT_E_RX_DROPPED:
- vport->err_stats.rx_dropped++;
- break;
+ switch (err_type) {
+ case VPORT_E_RX_DROPPED:
+ vport->err_stats.rx_dropped++;
+ break;
- case VPORT_E_RX_ERROR:
- vport->err_stats.rx_errors++;
- break;
+ case VPORT_E_RX_ERROR:
+ vport->err_stats.rx_errors++;
+ break;
- case VPORT_E_TX_DROPPED:
- vport->err_stats.tx_dropped++;
- break;
+ case VPORT_E_TX_DROPPED:
+ vport->err_stats.tx_dropped++;
+ break;
- case VPORT_E_TX_ERROR:
- vport->err_stats.tx_errors++;
- break;
- };
+ case VPORT_E_TX_ERROR:
+ vport->err_stats.tx_errors++;
+ break;
+ };
- spin_unlock_bh(&vport->stats_lock);
- }
+ spin_unlock_bh(&vport->stats_lock);
}