linux-sg2042/include/linux/if_team.h

254 lines
6.4 KiB
C
Raw Normal View History

net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
/*
* include/linux/if_team.h - Network team device driver header
* Copyright (c) 2011 Jiri Pirko <jpirko@redhat.com>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*/
#ifndef _LINUX_IF_TEAM_H_
#define _LINUX_IF_TEAM_H_
#include <linux/netpoll.h>
#include <net/sch_generic.h>
#include <uapi/linux/if_team.h>
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
struct team_pcpu_stats {
u64 rx_packets;
u64 rx_bytes;
u64 rx_multicast;
u64 tx_packets;
u64 tx_bytes;
struct u64_stats_sync syncp;
u32 rx_dropped;
u32 tx_dropped;
};
struct team;
struct team_port {
struct net_device *dev;
struct hlist_node hlist; /* node in enabled ports hash list */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
struct list_head list; /* node in ordinary list */
struct team *team;
int index; /* index of enabled port. If disabled, it's set to -1 */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
bool linkup; /* either state.linkup or user.linkup */
struct {
bool linkup;
u32 speed;
u8 duplex;
} state;
/* Values set by userspace */
struct {
bool linkup;
bool linkup_enabled;
} user;
/* Custom gennetlink interface related flags */
bool changed;
bool removed;
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
/*
* A place for storing original values of the device before it
* become a port.
*/
struct {
unsigned char dev_addr[MAX_ADDR_LEN];
unsigned int mtu;
} orig;
#ifdef CONFIG_NET_POLL_CONTROLLER
struct netpoll *np;
#endif
s32 priority; /* lower number ~ higher priority */
u16 queue_id;
struct list_head qom_list; /* node in queue override mapping list */
long mode_priv[0];
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
};
static inline bool team_port_enabled(struct team_port *port)
{
return port->index != -1;
}
static inline bool team_port_txable(struct team_port *port)
{
return port->linkup && team_port_enabled(port);
}
#ifdef CONFIG_NET_POLL_CONTROLLER
static inline void team_netpoll_send_skb(struct team_port *port,
struct sk_buff *skb)
{
struct netpoll *np = port->np;
if (np)
netpoll_send_skb(np, skb);
}
#else
static inline void team_netpoll_send_skb(struct team_port *port,
struct sk_buff *skb)
{
}
#endif
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
struct team_mode_ops {
int (*init)(struct team *team);
void (*exit)(struct team *team);
rx_handler_result_t (*receive)(struct team *team,
struct team_port *port,
struct sk_buff *skb);
bool (*transmit)(struct team *team, struct sk_buff *skb);
int (*port_enter)(struct team *team, struct team_port *port);
void (*port_leave)(struct team *team, struct team_port *port);
void (*port_change_dev_addr)(struct team *team, struct team_port *port);
void (*port_enabled)(struct team *team, struct team_port *port);
void (*port_disabled)(struct team *team, struct team_port *port);
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
};
enum team_option_type {
TEAM_OPTION_TYPE_U32,
TEAM_OPTION_TYPE_STRING,
TEAM_OPTION_TYPE_BINARY,
TEAM_OPTION_TYPE_BOOL,
TEAM_OPTION_TYPE_S32,
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
};
struct team_option_inst_info {
u32 array_index;
struct team_port *port; /* != NULL if per-port */
};
struct team_gsetter_ctx {
union {
u32 u32_val;
const char *str_val;
struct {
const void *ptr;
u32 len;
} bin_val;
bool bool_val;
s32 s32_val;
} data;
struct team_option_inst_info *info;
};
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
struct team_option {
struct list_head list;
const char *name;
bool per_port;
unsigned int array_size; /* != 0 means the option is array */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
enum team_option_type type;
int (*init)(struct team *team, struct team_option_inst_info *info);
int (*getter)(struct team *team, struct team_gsetter_ctx *ctx);
int (*setter)(struct team *team, struct team_gsetter_ctx *ctx);
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
};
extern void team_option_inst_set_change(struct team_option_inst_info *opt_inst_info);
extern void team_options_change_check(struct team *team);
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
struct team_mode {
const char *kind;
struct module *owner;
size_t priv_size;
size_t port_priv_size;
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
const struct team_mode_ops *ops;
};
#define TEAM_PORT_HASHBITS 4
#define TEAM_PORT_HASHENTRIES (1 << TEAM_PORT_HASHBITS)
#define TEAM_MODE_PRIV_LONGS 4
#define TEAM_MODE_PRIV_SIZE (sizeof(long) * TEAM_MODE_PRIV_LONGS)
struct team {
struct net_device *dev; /* associated netdevice */
struct team_pcpu_stats __percpu *pcpu_stats;
struct mutex lock; /* used for overall locking, e.g. port lists write */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
/*
* List of enabled ports and their count
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
*/
int en_port_count;
struct hlist_head en_port_hlist[TEAM_PORT_HASHENTRIES];
struct list_head port_list; /* list of all ports */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
struct list_head option_list;
struct list_head option_inst_list; /* list of option instances */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
const struct team_mode *mode;
struct team_mode_ops ops;
bool queue_override_enabled;
struct list_head *qom_lists; /* array of queue override mapping lists */
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
long mode_priv[TEAM_MODE_PRIV_LONGS];
};
static inline int team_dev_queue_xmit(struct team *team, struct team_port *port,
struct sk_buff *skb)
{
BUILD_BUG_ON(sizeof(skb->queue_mapping) !=
sizeof(qdisc_skb_cb(skb)->slave_dev_queue_mapping));
skb_set_queue_mapping(skb, qdisc_skb_cb(skb)->slave_dev_queue_mapping);
skb->dev = port->dev;
if (unlikely(netpoll_tx_running(team->dev))) {
team_netpoll_send_skb(port, skb);
return 0;
}
return dev_queue_xmit(skb);
}
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
static inline struct hlist_head *team_port_index_hash(struct team *team,
int port_index)
{
return &team->en_port_hlist[port_index & (TEAM_PORT_HASHENTRIES - 1)];
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
}
static inline struct team_port *team_get_port_by_index(struct team *team,
int port_index)
{
struct hlist_node *p;
struct team_port *port;
struct hlist_head *head = team_port_index_hash(team, port_index);
hlist_for_each_entry(port, p, head, hlist)
if (port->index == port_index)
return port;
return NULL;
}
static inline struct team_port *team_get_port_by_index_rcu(struct team *team,
int port_index)
{
struct hlist_node *p;
struct team_port *port;
struct hlist_head *head = team_port_index_hash(team, port_index);
hlist_for_each_entry_rcu(port, p, head, hlist)
if (port->index == port_index)
return port;
return NULL;
}
extern int team_port_set_team_dev_addr(struct team_port *port);
extern int team_options_register(struct team *team,
const struct team_option *option,
size_t option_count);
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
extern void team_options_unregister(struct team *team,
const struct team_option *option,
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
size_t option_count);
extern int team_mode_register(const struct team_mode *mode);
extern void team_mode_unregister(const struct team_mode *mode);
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
#define TEAM_DEFAULT_NUM_TX_QUEUES 16
#define TEAM_DEFAULT_NUM_RX_QUEUES 16
net: introduce ethernet teaming device This patch introduces new network device called team. It supposes to be very fast, simple, userspace-driven alternative to existing bonding driver. Userspace library called libteam with couple of demo apps is available here: https://github.com/jpirko/libteam Note it's still in its dipers atm. team<->libteam use generic netlink for communication. That and rtnl suppose to be the only way to configure team device, no sysfs etc. Python binding of libteam was recently introduced. Daemon providing arpmon/miimon active-backup functionality will be introduced shortly. All what's necessary is already implemented in kernel team driver. v7->v8: - check ndo_ndo_vlan_rx_[add/kill]_vid functions before calling them. - use dev_kfree_skb_any() instead of dev_kfree_skb() v6->v7: - transmit and receive functions are not checked in hot paths. That also resolves memory leak on transmit when no port is present v5->v6: - changed couple of _rcu calls to non _rcu ones in non-readers v4->v5: - team_change_mtu() uses team->lock while travesing though port list - mac address changes are moved completely to jurisdiction of userspace daemon. This way the daemon can do FOM1, FOM2 and possibly other weird things with mac addresses. Only round-robin mode sets up all ports to bond's address then enslaved. - Extended Kconfig text v3->v4: - remove redundant synchronize_rcu from __team_change_mode() - revert "set and clear of mode_ops happens per pointer, not per byte" - extend comment of function __team_change_mode() v2->v3: - team_change_mtu() uses rcu version of list traversal to unwind - set and clear of mode_ops happens per pointer, not per byte - port hashlist changed to be embedded into team structure - error branch in team_port_enter() does cleanup now - fixed rtln->rtnl v1->v2: - modes are made as modules. Makes team more modular and extendable. - several commenters' nitpicks found on v1 were fixed - several other bugs were fixed. - note I ignored Eric's comment about roundrobin port selector as Eric's way may be easily implemented as another mode (mode "random") in future. Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2011-11-12 06:16:48 +08:00
#endif /* _LINUX_IF_TEAM_H_ */