Subversion Repositories shark

Rev

Go to most recent revision | Blame | Compare with Previous | Last modification | View Log | RSS feed

/*
 * INET         An implementation of the TCP/IP protocol suite for the LINUX
 *              operating system.  INET is implemented using the  BSD Socket
 *              interface as the means of communication with the user level.
 *
 *              Definitions for the Interfaces handler.
 *
 * Version:     @(#)dev.h       1.0.10  08/12/93
 *
 * Authors:     Ross Biro, <bir7@leland.Stanford.Edu>
 *              Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *              Corey Minyard <wf-rch!minyard@relay.EU.net>
 *              Donald J. Becker, <becker@cesdis.gsfc.nasa.gov>
 *              Alan Cox, <Alan.Cox@linux.org>
 *              Bjorn Ekwall. <bj0rn@blox.se>
 *              Pekka Riikonen <priikone@poseidon.pspt.fi>
 *
 *              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.
 *
 *              Moved to /usr/include/linux for NET3
 */

#ifndef _LINUX_NETDEVICE_H
#define _LINUX_NETDEVICE_H

#include <linux/if.h>
#include <linux/if_ether.h>
#include <linux/if_packet.h>

#include <asm/atomic.h>
#include <asm/cache.h>
#include <asm/byteorder.h>

#ifdef __KERNEL__
#include <linux/config.h>
#include <linux/device.h>
#include <linux/percpu.h>

struct divert_blk;
struct vlan_group;
struct ethtool_ops;

                                        /* source back-compat hook */
#define SET_ETHTOOL_OPS(netdev,ops) \
        ( (netdev)->ethtool_ops = (ops) )


#define HAVE_ALLOC_NETDEV               /* feature macro: alloc_xxxdev
                                           functions are available. */
#define HAVE_FREE_NETDEV

#define NET_XMIT_SUCCESS        0
#define NET_XMIT_DROP           1       /* skb dropped                  */
#define NET_XMIT_CN             2       /* congestion notification      */
#define NET_XMIT_POLICED        3       /* skb is shot by police        */
#define NET_XMIT_BYPASS         4       /* packet does not leave via dequeue;
                                           (TC use only - dev_queue_xmit
                                           returns this as NET_XMIT_SUCCESS) */

/* Backlog congestion levels */
#define NET_RX_SUCCESS          0   /* keep 'em coming, baby */
#define NET_RX_DROP             1  /* packet dropped */
#define NET_RX_CN_LOW           2   /* storm alert, just in case */
#define NET_RX_CN_MOD           3   /* Storm on its way! */
#define NET_RX_CN_HIGH          4   /* The storm is here */
#define NET_RX_BAD              5  /* packet dropped due to kernel error */

#define net_xmit_errno(e)       ((e) != NET_XMIT_CN ? -ENOBUFS : 0)

#endif

#define MAX_ADDR_LEN    32              /* Largest hardware address length */

/*
 *      Compute the worst case header length according to the protocols
 *      used.
 */

 
#if !defined(CONFIG_AX25) && !defined(CONFIG_AX25_MODULE) && !defined(CONFIG_TR)
#define LL_MAX_HEADER   32
#else
#if defined(CONFIG_AX25) || defined(CONFIG_AX25_MODULE)
#define LL_MAX_HEADER   96
#else
#define LL_MAX_HEADER   48
#endif
#endif

#if !defined(CONFIG_NET_IPIP) && \
    !defined(CONFIG_IPV6) && !defined(CONFIG_IPV6_MODULE)

#define MAX_HEADER LL_MAX_HEADER
#else
#define MAX_HEADER (LL_MAX_HEADER + 48)
#endif

/*
 *      Network device statistics. Akin to the 2.0 ether stats but
 *      with byte counters.
 */

 
struct net_device_stats
{
        unsigned long   rx_packets;             /* total packets received       */
        unsigned long   tx_packets;             /* total packets transmitted    */
        unsigned long   rx_bytes;               /* total bytes received         */
        unsigned long   tx_bytes;               /* total bytes transmitted      */
        unsigned long   rx_errors;              /* bad packets received         */
        unsigned long   tx_errors;              /* packet transmit problems     */
        unsigned long   rx_dropped;             /* no space in linux buffers    */
        unsigned long   tx_dropped;             /* no space available in linux  */
        unsigned long   multicast;              /* multicast packets received   */
        unsigned long   collisions;

        /* detailed rx_errors: */
        unsigned long   rx_length_errors;
        unsigned long   rx_over_errors;         /* receiver ring buff overflow  */
        unsigned long   rx_crc_errors;          /* recved pkt with crc error    */
        unsigned long   rx_frame_errors;        /* recv'd frame alignment error */
        unsigned long   rx_fifo_errors;         /* recv'r fifo overrun          */
        unsigned long   rx_missed_errors;       /* receiver missed packet       */

        /* detailed tx_errors */
        unsigned long   tx_aborted_errors;
        unsigned long   tx_carrier_errors;
        unsigned long   tx_fifo_errors;
        unsigned long   tx_heartbeat_errors;
        unsigned long   tx_window_errors;
       
        /* for cslip etc */
        unsigned long   rx_compressed;
        unsigned long   tx_compressed;
};


/* Media selection options. */
enum {
        IF_PORT_UNKNOWN = 0,
        IF_PORT_10BASE2,
        IF_PORT_10BASET,
        IF_PORT_AUI,
        IF_PORT_100BASET,
        IF_PORT_100BASETX,
        IF_PORT_100BASEFX
};

#ifdef __KERNEL__

#include <linux/cache.h>
#include <linux/skbuff.h>

struct neighbour;
struct neigh_parms;
struct sk_buff;

struct netif_rx_stats
{
        unsigned total;
        unsigned dropped;
        unsigned time_squeeze;
        unsigned throttled;
        unsigned fastroute_hit;
        unsigned fastroute_success;
        unsigned fastroute_defer;
        unsigned fastroute_deferred_out;
        unsigned fastroute_latency_reduction;
        unsigned cpu_collision;
};

DECLARE_PER_CPU(struct netif_rx_stats, netdev_rx_stat);


/*
 *      We tag multicasts with these structures.
 */

 
struct dev_mc_list
{      
        struct dev_mc_list      *next;
        __u8                    dmi_addr[MAX_ADDR_LEN];
        unsigned char           dmi_addrlen;
        int                     dmi_users;
        int                     dmi_gusers;
};

struct hh_cache
{
        struct hh_cache *hh_next;       /* Next entry                        */
        atomic_t        hh_refcnt;      /* number of users                   */
        unsigned short  hh_type;        /* protocol identifier, f.e ETH_P_IP
                                         *  NOTE:  For VLANs, this will be the
                                         *  encapuslated type. --BLG
                                         */

        int             hh_len;         /* length of header */
        int             (*hh_output)(struct sk_buff *skb);
        rwlock_t        hh_lock;

        /* cached hardware header; allow for machine alignment needs.        */
#define HH_DATA_MOD     16
#define HH_DATA_OFF(__len) \
        (HH_DATA_MOD - ((__len) & (HH_DATA_MOD - 1)))

#define HH_DATA_ALIGN(__len) \
        (((__len)+(HH_DATA_MOD-1))&~(HH_DATA_MOD - 1))

        unsigned long   hh_data[HH_DATA_ALIGN(LL_MAX_HEADER) / sizeof(long)];
};

/* Reserve HH_DATA_MOD byte aligned hard_header_len, but at least that much.
 * Alternative is:
 *   dev->hard_header_len ? (dev->hard_header_len +
 *                           (HH_DATA_MOD - 1)) & ~(HH_DATA_MOD - 1) : 0
 *
 * We could use other alignment values, but we must maintain the
 * relationship HH alignment <= LL alignment.
 */

#define LL_RESERVED_SPACE(dev) \
        (((dev)->hard_header_len&~(HH_DATA_MOD - 1)) + HH_DATA_MOD)


/* These flag bits are private to the generic network queueing
 * layer, they may not be explicitly referenced by any other
 * code.
 */


enum netdev_state_t
{
        __LINK_STATE_XOFF=0,
        __LINK_STATE_START,
        __LINK_STATE_PRESENT,
        __LINK_STATE_SCHED,
        __LINK_STATE_NOCARRIER,
        __LINK_STATE_RX_SCHED,
        __LINK_STATE_LINKWATCH_PENDING
};


/*
 * This structure holds at boot time configured netdevice settings. They
 * are then used in the device probing.
 */

struct netdev_boot_setup {
        char name[IFNAMSIZ];
        struct ifmap map;
};
#define NETDEV_BOOT_SETUP_MAX 8


/*
 *      The DEVICE structure.
 *      Actually, this whole structure is a big mistake.  It mixes I/O
 *      data with strictly "high-level" data, and it has to know about
 *      almost every data structure used in the INET module.
 *
 *      FIXME: cleanup struct net_device such that network protocol info
 *      moves out.
 */


struct net_device
{

        /*
         * This is the first field of the "visible" part of this structure
         * (i.e. as seen by users in the "Space.c" file).  It is the name
         * the interface.
         */

        char                    name[IFNAMSIZ];

        /*
         *      I/O specific fields
         *      FIXME: Merge these and struct ifmap into one
         */

        unsigned long           mem_end;        /* shared mem end       */
        unsigned long           mem_start;      /* shared mem start     */
        unsigned long           base_addr;      /* device I/O address   */
        unsigned int            irq;            /* device IRQ number    */

        /*
         *      Some hardware also needs these fields, but they are not
         *      part of the usual set specified in Space.c.
         */


        unsigned char           if_port;        /* Selectable AUI, TP,..*/
        unsigned char           dma;            /* DMA channel          */

        unsigned long           state;

        struct net_device       *next;
       
        /* The device initialization function. Called only once. */
        int                     (*init)(struct net_device *dev);

        /* ------- Fields preinitialized in Space.c finish here ------- */

        struct net_device       *next_sched;

        /* Interface index. Unique device identifier    */
        int                     ifindex;
        int                     iflink;


        struct net_device_stats* (*get_stats)(struct net_device *dev);
        struct iw_statistics*   (*get_wireless_stats)(struct net_device *dev);

        /* List of functions to handle Wireless Extensions (instead of ioctl).
         * See <net/iw_handler.h> for details. Jean II */

        struct iw_handler_def * wireless_handlers;

        struct ethtool_ops *ethtool_ops;

        /*
         * This marks the end of the "visible" part of the structure. All
         * fields hereafter are internal to the system, and may change at
         * will (read: may be cleaned up at will).
         */


        /* These may be needed for future network-power-down code. */
        unsigned long           trans_start;    /* Time (in jiffies) of last Tx */
        unsigned long           last_rx;        /* Time of last Rx      */

        unsigned short          flags;  /* interface flags (a la BSD)   */
        unsigned short          gflags;
        unsigned short          priv_flags; /* Like 'flags' but invisible to userspace. */
        unsigned short          unused_alignment_fixer; /* Because we need priv_flags,
                                                         * and we want to be 32-bit aligned.
                                                         */


        unsigned                mtu;    /* interface MTU value          */
        unsigned short          type;   /* interface hardware type      */
        unsigned short          hard_header_len;        /* hardware hdr length  */
        void                    *priv;  /* pointer to private data      */

        struct net_device       *master; /* Pointer to master device of a group,
                                          * which this device is member of.
                                          */


        /* Interface address info. */
        unsigned char           broadcast[MAX_ADDR_LEN];        /* hw bcast add */
        unsigned char           dev_addr[MAX_ADDR_LEN]; /* hw address   */
        unsigned char           addr_len;       /* hardware address length      */

        struct dev_mc_list      *mc_list;       /* Multicast mac addresses      */
        int                     mc_count;       /* Number of installed mcasts   */
        int                     promiscuity;
        int                     allmulti;

        int                     watchdog_timeo;
        struct timer_list       watchdog_timer;

        /* Protocol specific pointers */
       
        void                    *atalk_ptr;     /* AppleTalk link       */
        void                    *ip_ptr;        /* IPv4 specific data   */  
        void                    *dn_ptr;        /* DECnet specific data */
        void                    *ip6_ptr;       /* IPv6 specific data */
        void                    *ec_ptr;        /* Econet specific data */
        void                    *ax25_ptr;      /* AX.25 specific data */

        struct list_head        poll_list;      /* Link to poll list    */
        int                     quota;
        int                     weight;

        struct Qdisc            *qdisc;
        struct Qdisc            *qdisc_sleeping;
        struct Qdisc            *qdisc_list;
        struct Qdisc            *qdisc_ingress;
        unsigned long           tx_queue_len;   /* Max frames per queue allowed */

        /* hard_start_xmit synchronizer */
        spinlock_t              xmit_lock;
        /* cpu id of processor entered to hard_start_xmit or -1,
           if nobody entered there.
         */

        int                     xmit_lock_owner;
        /* device queue lock */
        spinlock_t              queue_lock;
        /* Number of references to this device */
        atomic_t                refcnt;
        /* delayed register/unregister */
        struct list_head        todo_list;

        /* register/unregister state machine */
        enum { NETREG_UNINITIALIZED=0,
               NETREG_REGISTERING,      /* called register_netdevice */
               NETREG_REGISTERED,       /* completed register todo */
               NETREG_UNREGISTERING,    /* called unregister_netdevice */
               NETREG_UNREGISTERED,     /* completed unregister todo */
               NETREG_RELEASED,         /* called free_netdev */
        } reg_state;

        /* Net device features */
        int                     features;
#define NETIF_F_SG              1       /* Scatter/gather IO. */
#define NETIF_F_IP_CSUM         2       /* Can checksum only TCP/UDP over IPv4. */
#define NETIF_F_NO_CSUM         4       /* Does not require checksum. F.e. loopack. */
#define NETIF_F_HW_CSUM         8       /* Can checksum all the packets. */
#define NETIF_F_HIGHDMA         32      /* Can DMA to high memory. */
#define NETIF_F_FRAGLIST        64      /* Scatter/gather IO. */
#define NETIF_F_HW_VLAN_TX      128     /* Transmit VLAN hw acceleration */
#define NETIF_F_HW_VLAN_RX      256     /* Receive VLAN hw acceleration */
#define NETIF_F_HW_VLAN_FILTER  512     /* Receive filtering on VLAN */
#define NETIF_F_VLAN_CHALLENGED 1024    /* Device cannot handle VLAN packets */
#define NETIF_F_TSO             2048    /* Can offload TCP/IP segmentation */

        /* Called after device is detached from network. */
        void                    (*uninit)(struct net_device *dev);
        /* Called after last user reference disappears. */
        void                    (*destructor)(struct net_device *dev);

        /* Pointers to interface service routines.      */
        int                     (*open)(struct net_device *dev);
        int                     (*stop)(struct net_device *dev);
        int                     (*hard_start_xmit) (struct sk_buff *skb,
                                                    struct net_device *dev);
#define HAVE_NETDEV_POLL
        int                     (*poll) (struct net_device *dev, int *quota);
        int                     (*hard_header) (struct sk_buff *skb,
                                                struct net_device *dev,
                                                unsigned short type,
                                                void *daddr,
                                                void *saddr,
                                                unsigned len);
        int                     (*rebuild_header)(struct sk_buff *skb);
#define HAVE_MULTICAST                   
        void                    (*set_multicast_list)(struct net_device *dev);
#define HAVE_SET_MAC_ADDR                
        int                     (*set_mac_address)(struct net_device *dev,
                                                   void *addr);
#define HAVE_PRIVATE_IOCTL
        int                     (*do_ioctl)(struct net_device *dev,
                                            struct ifreq *ifr, int cmd);
#define HAVE_SET_CONFIG
        int                     (*set_config)(struct net_device *dev,
                                              struct ifmap *map);
#define HAVE_HEADER_CACHE
        int                     (*hard_header_cache)(struct neighbour *neigh,
                                                     struct hh_cache *hh);
        void                    (*header_cache_update)(struct hh_cache *hh,
                                                       struct net_device *dev,
                                                       unsigned char *  haddr);
#define HAVE_CHANGE_MTU
        int                     (*change_mtu)(struct net_device *dev, int new_mtu);

#define HAVE_TX_TIMEOUT
        void                    (*tx_timeout) (struct net_device *dev);

        void                    (*vlan_rx_register)(struct net_device *dev,
                                                    struct vlan_group *grp);
        void                    (*vlan_rx_add_vid)(struct net_device *dev,
                                                   unsigned short vid);
        void                    (*vlan_rx_kill_vid)(struct net_device *dev,
                                                    unsigned short vid);

        int                     (*hard_header_parse)(struct sk_buff *skb,
                                                     unsigned char *haddr);
        int                     (*neigh_setup)(struct net_device *dev, struct neigh_parms *);
        int                     (*accept_fastpath)(struct net_device *, struct dst_entry*);

        /* bridge stuff */
        struct net_bridge_port  *br_port;

#ifdef CONFIG_NET_FASTROUTE
#define NETDEV_FASTROUTE_HMASK 0xF
        /* Semi-private data. Keep it at the end of device struct. */
        rwlock_t                fastpath_lock;
        struct dst_entry        *fastpath[NETDEV_FASTROUTE_HMASK+1];
#endif
#ifdef CONFIG_NET_DIVERT
        /* this will get initialized at each interface type init routine */
        struct divert_blk       *divert;
#endif /* CONFIG_NET_DIVERT */

        /* class/net/name entry */
        struct class_device     class_dev;
        struct net_device_stats* (*last_stats)(struct net_device *);
};

#define SET_MODULE_OWNER(dev) do { } while (0)
/* Set the sysfs physical device reference for the network logical device
 * if set prior to registration will cause a symlink during initialization.
 */

#define SET_NETDEV_DEV(net, pdev)       ((net)->class_dev.dev = (pdev))

struct packet_type {
        unsigned short          type;   /* This is really htons(ether_type).    */
        struct net_device               *dev;   /* NULL is wildcarded here              */
        int                     (*func) (struct sk_buff *, struct net_device *,
                                         struct packet_type *);
        void                    *af_packet_priv;
        struct list_head        list;
};

#include <linux/interrupt.h>
#include <linux/notifier.h>

extern struct net_device                loopback_dev;           /* The loopback */
extern struct net_device                *dev_base;              /* All devices */
extern rwlock_t                         dev_base_lock;          /* Device list lock */

extern int                      netdev_boot_setup_add(char *name, struct ifmap *map);
extern int                      netdev_boot_setup_check(struct net_device *dev);
extern struct net_device    *dev_getbyhwaddr(unsigned short type, char *hwaddr);
extern struct net_device *__dev_getfirstbyhwtype(unsigned short type);
extern struct net_device *dev_getfirstbyhwtype(unsigned short type);
extern void             dev_add_pack(struct packet_type *pt);
extern void             dev_remove_pack(struct packet_type *pt);
extern void             __dev_remove_pack(struct packet_type *pt);
extern int              __dev_get(const char *name);
static inline int __deprecated dev_get(const char *name)
{
        return __dev_get(name);
}
extern struct net_device        *dev_get_by_flags(unsigned short flags,
                                                  unsigned short mask);
extern struct net_device        *__dev_get_by_flags(unsigned short flags,
                                                    unsigned short mask);
extern struct net_device        *dev_get_by_name(const char *name);
extern struct net_device        *__dev_get_by_name(const char *name);
extern struct net_device        *__dev_alloc(const char *name, int *err);
static inline __deprecated struct net_device *dev_alloc(const char *name, int *err)
{
        return __dev_alloc(name, err);
}
extern int              dev_alloc_name(struct net_device *dev, const char *name);
extern int              dev_open(struct net_device *dev);
extern int              dev_close(struct net_device *dev);
extern int              dev_queue_xmit(struct sk_buff *skb);
extern int              register_netdevice(struct net_device *dev);
extern int              unregister_netdevice(struct net_device *dev);
extern void             free_netdev(struct net_device *dev);
extern void             synchronize_net(void);
extern int              register_netdevice_notifier(struct notifier_block *nb);
extern int              unregister_netdevice_notifier(struct notifier_block *nb);
extern int              call_netdevice_notifiers(unsigned long val, void *v);
extern int              dev_new_index(void);
extern struct net_device        *dev_get_by_index(int ifindex);
extern struct net_device        *__dev_get_by_index(int ifindex);
extern int              dev_restart(struct net_device *dev);

typedef int gifconf_func_t(struct net_device * dev, char * bufptr, int len);
extern int              register_gifconf(unsigned int family, gifconf_func_t * gifconf);
static inline int unregister_gifconf(unsigned int family)
{
        return register_gifconf(family, 0);
}

/*
 * Incoming packets are placed on per-cpu queues so that
 * no locking is needed.
 */


struct softnet_data
{
        int                     throttle;
        int                     cng_level;
        int                     avg_blog;
        struct sk_buff_head     input_pkt_queue;
        struct list_head        poll_list;
        struct net_device       *output_queue;
        struct sk_buff          *completion_queue;

        struct net_device       backlog_dev;    /* Sorry. 8) */
};

DECLARE_PER_CPU(struct softnet_data,softnet_data);

#define HAVE_NETIF_QUEUE

static inline void __netif_schedule(struct net_device *dev)
{
        if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) {
                unsigned long flags;
                struct softnet_data *sd;

                local_irq_save(flags);
                sd = &__get_cpu_var(softnet_data);
                dev->next_sched = sd->output_queue;
                sd->output_queue = dev;
                raise_softirq_irqoff(NET_TX_SOFTIRQ);
                local_irq_restore(flags);
        }
}

static inline void netif_schedule(struct net_device *dev)
{
        if (!test_bit(__LINK_STATE_XOFF, &dev->state))
                __netif_schedule(dev);
}

static inline void netif_start_queue(struct net_device *dev)
{
        clear_bit(__LINK_STATE_XOFF, &dev->state);
}

static inline void netif_wake_queue(struct net_device *dev)
{
        if (test_and_clear_bit(__LINK_STATE_XOFF, &dev->state))
                __netif_schedule(dev);
}

static inline void netif_stop_queue(struct net_device *dev)
{
        set_bit(__LINK_STATE_XOFF, &dev->state);
}

static inline int netif_queue_stopped(const struct net_device *dev)
{
        return test_bit(__LINK_STATE_XOFF, &dev->state);
}

static inline int netif_running(const struct net_device *dev)
{
        return test_bit(__LINK_STATE_START, &dev->state);
}


/* Use this variant when it is known for sure that it
 * is executing from interrupt context.
 */

static inline void dev_kfree_skb_irq(struct sk_buff *skb)
{
        if (atomic_dec_and_test(&skb->users)) {
                struct softnet_data *sd;
                unsigned long flags;

                local_irq_save(flags);
                sd = &__get_cpu_var(softnet_data);
                skb->next = sd->completion_queue;
                sd->completion_queue = skb;
                raise_softirq_irqoff(NET_TX_SOFTIRQ);
                local_irq_restore(flags);
        }
}

/* Use this variant in places where it could be invoked
 * either from interrupt or non-interrupt context.
 */

static inline void dev_kfree_skb_any(struct sk_buff *skb)
{
        if (in_irq())
                dev_kfree_skb_irq(skb);
        else
                dev_kfree_skb(skb);
}

#define HAVE_NETIF_RX 1
extern int              netif_rx(struct sk_buff *skb);
#define HAVE_NETIF_RECEIVE_SKB 1
extern int              netif_receive_skb(struct sk_buff *skb);
extern int              dev_ioctl(unsigned int cmd, void *);
extern int              dev_ethtool(struct ifreq *);
extern unsigned         dev_get_flags(const struct net_device *);
extern int              dev_change_flags(struct net_device *, unsigned);
extern int              dev_set_mtu(struct net_device *, int);
extern void             dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev);

extern void             dev_init(void);

extern int              netdev_nit;

/* Post buffer to the network code from _non interrupt_ context.
 * see net/core/dev.c for netif_rx description.
 */

static inline int netif_rx_ni(struct sk_buff *skb)
{
       int err = netif_rx(skb);
       if (softirq_pending(smp_processor_id()))
               do_softirq();
       return err;
}

/* Called by rtnetlink.c:rtnl_unlock() */
extern void netdev_run_todo(void);

static inline void dev_put(struct net_device *dev)
{
        atomic_dec(&dev->refcnt);
}

#define __dev_put(dev) atomic_dec(&(dev)->refcnt)
#define dev_hold(dev) atomic_inc(&(dev)->refcnt)

/* Carrier loss detection, dial on demand. The functions netif_carrier_on
 * and _off may be called from IRQ context, but it is caller
 * who is responsible for serialization of these calls.
 */


extern void linkwatch_fire_event(struct net_device *dev);

static inline int netif_carrier_ok(const struct net_device *dev)
{
        return !test_bit(__LINK_STATE_NOCARRIER, &dev->state);
}

extern void __netdev_watchdog_up(struct net_device *dev);

static inline void netif_carrier_on(struct net_device *dev)
{
        if (test_and_clear_bit(__LINK_STATE_NOCARRIER, &dev->state))
                linkwatch_fire_event(dev);
        if (netif_running(dev))
                __netdev_watchdog_up(dev);
}

static inline void netif_carrier_off(struct net_device *dev)
{
        if (!test_and_set_bit(__LINK_STATE_NOCARRIER, &dev->state))
                linkwatch_fire_event(dev);
}

/* Hot-plugging. */
static inline int netif_device_present(struct net_device *dev)
{
        return test_bit(__LINK_STATE_PRESENT, &dev->state);
}

static inline void netif_device_detach(struct net_device *dev)
{
        if (test_and_clear_bit(__LINK_STATE_PRESENT, &dev->state) &&
            netif_running(dev)) {
                netif_stop_queue(dev);
        }
}

static inline void netif_device_attach(struct net_device *dev)
{
        if (!test_and_set_bit(__LINK_STATE_PRESENT, &dev->state) &&
            netif_running(dev)) {
                netif_wake_queue(dev);
                __netdev_watchdog_up(dev);
        }
}

/*
 * Network interface message level settings
 */

#define HAVE_NETIF_MSG 1

enum {
        NETIF_MSG_DRV           = 0x0001,
        NETIF_MSG_PROBE         = 0x0002,
        NETIF_MSG_LINK          = 0x0004,
        NETIF_MSG_TIMER         = 0x0008,
        NETIF_MSG_IFDOWN        = 0x0010,
        NETIF_MSG_IFUP          = 0x0020,
        NETIF_MSG_RX_ERR        = 0x0040,
        NETIF_MSG_TX_ERR        = 0x0080,
        NETIF_MSG_TX_QUEUED     = 0x0100,
        NETIF_MSG_INTR          = 0x0200,
        NETIF_MSG_TX_DONE       = 0x0400,
        NETIF_MSG_RX_STATUS     = 0x0800,
        NETIF_MSG_PKTDATA       = 0x1000,
        NETIF_MSG_HW            = 0x2000,
        NETIF_MSG_WOL           = 0x4000,
};

#define netif_msg_drv(p)        ((p)->msg_enable & NETIF_MSG_DRV)
#define netif_msg_probe(p)      ((p)->msg_enable & NETIF_MSG_PROBE)
#define netif_msg_link(p)       ((p)->msg_enable & NETIF_MSG_LINK)
#define netif_msg_timer(p)      ((p)->msg_enable & NETIF_MSG_TIMER)
#define netif_msg_ifdown(p)     ((p)->msg_enable & NETIF_MSG_IFDOWN)
#define netif_msg_ifup(p)       ((p)->msg_enable & NETIF_MSG_IFUP)
#define netif_msg_rx_err(p)     ((p)->msg_enable & NETIF_MSG_RX_ERR)
#define netif_msg_tx_err(p)     ((p)->msg_enable & NETIF_MSG_TX_ERR)
#define netif_msg_tx_queued(p)  ((p)->msg_enable & NETIF_MSG_TX_QUEUED)
#define netif_msg_intr(p)       ((p)->msg_enable & NETIF_MSG_INTR)
#define netif_msg_tx_done(p)    ((p)->msg_enable & NETIF_MSG_TX_DONE)
#define netif_msg_rx_status(p)  ((p)->msg_enable & NETIF_MSG_RX_STATUS)
#define netif_msg_pktdata(p)    ((p)->msg_enable & NETIF_MSG_PKTDATA)
#define netif_msg_hw(p)         ((p)->msg_enable & NETIF_MSG_HW)
#define netif_msg_wol(p)        ((p)->msg_enable & NETIF_MSG_WOL)

/* Schedule rx intr now? */

static inline int netif_rx_schedule_prep(struct net_device *dev)
{
        return netif_running(dev) &&
                !test_and_set_bit(__LINK_STATE_RX_SCHED, &dev->state);
}

/* Add interface to tail of rx poll list. This assumes that _prep has
 * already been called and returned 1.
 */


static inline void __netif_rx_schedule(struct net_device *dev)
{
        unsigned long flags;

        local_irq_save(flags);
        dev_hold(dev);
        list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list);
        if (dev->quota < 0)
                dev->quota += dev->weight;
        else
                dev->quota = dev->weight;
        __raise_softirq_irqoff(NET_RX_SOFTIRQ);
        local_irq_restore(flags);
}

/* Try to reschedule poll. Called by irq handler. */

static inline void netif_rx_schedule(struct net_device *dev)
{
        if (netif_rx_schedule_prep(dev))
                __netif_rx_schedule(dev);
}

/* Try to reschedule poll. Called by dev->poll() after netif_rx_complete().
 * Do not inline this?
 */

static inline int netif_rx_reschedule(struct net_device *dev, int undo)
{
        if (netif_rx_schedule_prep(dev)) {
                unsigned long flags;

                dev->quota += undo;

                local_irq_save(flags);
                list_add_tail(&dev->poll_list, &__get_cpu_var(softnet_data).poll_list);
                __raise_softirq_irqoff(NET_RX_SOFTIRQ);
                local_irq_restore(flags);
                return 1;
        }
        return 0;
}

/* Remove interface from poll list: it must be in the poll list
 * on current cpu. This primitive is called by dev->poll(), when
 * it completes the work. The device cannot be out of poll list at this
 * moment, it is BUG().
 */

static inline void netif_rx_complete(struct net_device *dev)
{
        unsigned long flags;

        local_irq_save(flags);
        if (!test_bit(__LINK_STATE_RX_SCHED, &dev->state)) BUG();
        list_del(&dev->poll_list);
        smp_mb__before_clear_bit();
        clear_bit(__LINK_STATE_RX_SCHED, &dev->state);
        local_irq_restore(flags);
}

static inline void netif_poll_disable(struct net_device *dev)
{
        while (test_and_set_bit(__LINK_STATE_RX_SCHED, &dev->state)) {
                /* No hurry. */
                current->state = TASK_INTERRUPTIBLE;
                schedule_timeout(1);
        }
}

static inline void netif_poll_enable(struct net_device *dev)
{
        clear_bit(__LINK_STATE_RX_SCHED, &dev->state);
}

/* same as netif_rx_complete, except that local_irq_save(flags)
 * has already been issued
 */

static inline void __netif_rx_complete(struct net_device *dev)
{
        if (!test_bit(__LINK_STATE_RX_SCHED, &dev->state)) BUG();
        list_del(&dev->poll_list);
        smp_mb__before_clear_bit();
        clear_bit(__LINK_STATE_RX_SCHED, &dev->state);
}

static inline void netif_tx_disable(struct net_device *dev)
{
        spin_lock_bh(&dev->xmit_lock);
        netif_stop_queue(dev);
        spin_unlock_bh(&dev->xmit_lock);
}

/* These functions live elsewhere (drivers/net/net_init.c, but related) */

extern void             ether_setup(struct net_device *dev);
extern void             fddi_setup(struct net_device *dev);
extern void             tr_setup(struct net_device *dev);
extern void             fc_setup(struct net_device *dev);
extern void             fc_freedev(struct net_device *dev);
/* Support for loadable net-drivers */
extern struct net_device *alloc_netdev(int sizeof_priv, const char *name,
                                       void (*setup)(struct net_device *));
extern int              register_netdev(struct net_device *dev);
extern void             unregister_netdev(struct net_device *dev);
/* Functions used for multicast support */
extern void             dev_mc_upload(struct net_device *dev);
extern int              dev_mc_delete(struct net_device *dev, void *addr, int alen, int all);
extern int              dev_mc_add(struct net_device *dev, void *addr, int alen, int newonly);
extern void             dev_mc_discard(struct net_device *dev);
extern void             dev_set_promiscuity(struct net_device *dev, int inc);
extern void             dev_set_allmulti(struct net_device *dev, int inc);
extern void             netdev_state_change(struct net_device *dev);
/* Load a device via the kmod */
extern void             dev_load(const char *name);
extern void             dev_mcast_init(void);
extern int              netdev_register_fc(struct net_device *dev, void (*stimul)(struct net_device *dev));
extern void             netdev_unregister_fc(int bit);
extern int              netdev_max_backlog;
extern int              weight_p;
extern unsigned long    netdev_fc_xoff;
extern atomic_t netdev_dropping;
extern int              netdev_set_master(struct net_device *dev, struct net_device *master);
extern struct sk_buff * skb_checksum_help(struct sk_buff *skb);
#ifdef CONFIG_NET_FASTROUTE
extern int              netdev_fastroute;
extern int              netdev_fastroute_obstacles;
extern void             dev_clear_fastroute(struct net_device *dev);
#endif

#ifdef CONFIG_SYSCTL
extern char *net_sysctl_strdup(const char *s);
#endif

#endif /* __KERNEL__ */

#endif  /* _LINUX_DEV_H */