dev.c

GNU Mach 微内核源代码, 基于美国卡内基美隆大学的 Mach 研究项目

/*
 * 	NET3	Protocol independent device support routines.
 *
 *		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.
 *
 *	Derived from the non IP parts of dev.c 1.0.19
 * 		Authors:	Ross Biro, <[e-mail]>
 *				Fred N. van Kempen, <[e-mail]>
 *				Mark Evans, <[e-mail]>
 *
 *	Additional Authors:
 *		Florian la Roche <[e-mail]>
 *		Alan Cox <[e-mail]>
 *		David Hinds <[e-mail]>
 *
 *	Changes:
 *		Alan Cox	:	device private ioctl copies fields back.
 *		Alan Cox	:	Transmit queue code does relevant stunts to
 *					keep the queue safe.
 *		Alan Cox	:	Fixed double lock.
 *		Alan Cox	:	Fixed promisc NULL pointer trap
 *		????????	:	Support the full private ioctl range
 *		Alan Cox	:	Moved ioctl permission check into drivers
 *		Tim Kordas	:	SIOCADDMULTI/SIOCDELMULTI
 *		Alan Cox	:	100 backlog just doesn't cut it when
 *					you start doing multicast video 8)
 *		Alan Cox	:	Rewrote net_bh and list manager.
 *		Alan Cox	: 	Fix ETH_P_ALL echoback lengths.
 *		Alan Cox	:	Took out transmit every packet pass
 *					Saved a few bytes in the ioctl handler
 *		Alan Cox	:	Network driver sets packet type before calling netif_rx. Saves
 *					a function call a packet.
 *		Alan Cox	:	Hashed net_bh()
 *		Richard Kooijman:	Timestamp fixes.
 *		Alan Cox	:	Wrong field in SIOCGIFDSTADDR
 *		Alan Cox	:	Device lock protection.
 *		Alan Cox	: 	Fixed nasty side effect of device close changes.
 *		Rudi Cilibrasi	:	Pass the right thing to set_mac_address()
 *		Dave Miller	:	32bit quantity for the device lock to make it work out
 *					on a Sparc.
 *		Bjorn Ekwall	:	Added KERNELD hack.
 *		Alan Cox	:	Cleaned up the backlog initialise.
 *		Craig Metz	:	SIOCGIFCONF fix if space for under
 *					1 device.
 *	    Thomas Bogendoerfer :	Return ENODEV for dev_open, if there
 *					is no device open function.
 *	    Lawrence V. Stefani	:	Changed set MTU ioctl to not assume
 *					min MTU of 68 bytes for devices
 *					that have change MTU functions.
 *
 */

#include <asm/segment.h>
#include <asm/system.h>
#include <asm/bitops.h>
#include <linux/config.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/arp.h>
#include <net/slhc.h>
#include <linux/proc_fs.h>
#include <linux/stat.h>
#include <net/br.h>
#ifdef CONFIG_NET_ALIAS
#include <linux/net_alias.h>
#endif
#ifdef CONFIG_KERNELD
#include <linux/kerneld.h>
#endif
#ifdef CONFIG_NET_RADIO
#include <linux/wireless.h>
#endif	/* CONFIG_NET_RADIO */

#ifndef MACH
/*
 *	The list of packet types we will receive (as opposed to discard)
 *	and the routines to invoke.
 */

struct packet_type *ptype_base[16];
struct packet_type *ptype_all = NULL;		/* Taps */

/*
 *	Device list lock
 */
 
int dev_lockct=0;
 
/*
 *	Our notifier list
 */
 
struct notifier_block *netdev_chain=NULL;

/*
 *	Device drivers call our routines to queue packets here. We empty the
 *	queue in the bottom half handler.
 */

static struct sk_buff_head backlog;

/* 
 *	We don't overdo the queue or we will thrash memory badly.
 */
 
static int backlog_size = 0;

/*
 *	Return the lesser of the two values. 
 */
 
static __inline__ unsigned long min(unsigned long a, unsigned long b)
{
	return (a < b)? a : b;
}


/******************************************************************************************

		Protocol management and registration routines

*******************************************************************************************/

/*
 *	For efficiency
 */

static int dev_nit=0;

/*
 *	Add a protocol ID to the list. Now that the input handler is
 *	smarter we can dispense with all the messy stuff that used to be
 *	here.
 */
 
void dev_add_pack(struct packet_type *pt)
{
	int hash;
	if(pt->type==htons(ETH_P_ALL))
	{
		dev_nit++;
		pt->next=ptype_all;
		ptype_all=pt;
	}
	else
	{	
		hash=ntohs(pt->type)&15;
		pt->next = ptype_base[hash];
		ptype_base[hash] = pt;
	}
}


/*
 *	Remove a protocol ID from the list.
 */
 
void dev_remove_pack(struct packet_type *pt)
{
	struct packet_type **pt1;
	if(pt->type==htons(ETH_P_ALL))
	{
		dev_nit--;
		pt1=&ptype_all;
	}
	else
		pt1=&ptype_base[ntohs(pt->type)&15];
	for(; (*pt1)!=NULL; pt1=&((*pt1)->next))
	{
		if(pt==(*pt1))
		{
			*pt1=pt->next;
			return;
		}
	}
	printk(KERN_WARNING "dev_remove_pack: %p not found.\n", pt);
}

/*****************************************************************************************

			    Device Interface Subroutines

******************************************************************************************/

/* 
 *	Find an interface by name.
 */
 
struct device *dev_get(const char *name)
{
	struct device *dev;

	for (dev = dev_base; dev != NULL; dev = dev->next) 
	{
		if (strcmp(dev->name, name) == 0)
			return(dev);
	}
	return NULL;
}
	
/*
 *	Find and possibly load an interface.
 */
 
#ifdef CONFIG_KERNELD

extern __inline__ void dev_load(const char *name)
{
	if(!dev_get(name) && suser()) {
#ifdef CONFIG_NET_ALIAS
		const char *sptr;
 
		for (sptr=name ; *sptr ; sptr++) if(*sptr==':') break;
		if (!(*sptr && *(sptr+1)))
#endif
		request_module(name);
	}
}

#endif
 
/*
 *	Prepare an interface for use. 
 */
 
int dev_open(struct device *dev)
{
	int ret = -ENODEV;

	/*
	 *	Call device private open method
	 */
	if (dev->open) 
  		ret = dev->open(dev);

	/*
	 *	If it went open OK then set the flags
	 */
	 
	if (ret == 0) 
	{
		dev->flags |= (IFF_UP | IFF_RUNNING);
		/*
		 *	Initialise multicasting status 
		 */
		dev_mc_upload(dev);
		notifier_call_chain(&netdev_chain, NETDEV_UP, dev);
	}
	return(ret);
}


/*
 *	Completely shutdown an interface.
 */
 
int dev_close(struct device *dev)
{
	int ct=0;

	/*
	 *	Call the device specific close. This cannot fail.
	 *	Only if device is UP
	 */
	 
	if ((dev->flags & IFF_UP) && dev->stop)
		dev->stop(dev);

	/*
	 *	Device is now down.
	 */
	 
	dev->flags&=~(IFF_UP|IFF_RUNNING);

	/*
	 *	Tell people we are going down
	 */
	notifier_call_chain(&netdev_chain, NETDEV_DOWN, dev);
	/*
	 *	Flush the multicast chain
	 */
	dev_mc_discard(dev);

	/*
	 *	Purge any queued packets when we down the link 
	 */
	while(ct<DEV_NUMBUFFS)
	{
		struct sk_buff *skb;
		while((skb=skb_dequeue(&dev->buffs[ct]))!=NULL)
			if(skb->free)
				kfree_skb(skb,FREE_WRITE);
		ct++;
	}
	return(0);
}


/*
 *	Device change register/unregister. These are not inline or static
 *	as we export them to the world.
 */

int register_netdevice_notifier(struct notifier_block *nb)
{
	return notifier_chain_register(&netdev_chain, nb);
}

int unregister_netdevice_notifier(struct notifier_block *nb)
{
	return notifier_chain_unregister(&netdev_chain,nb);
}

/*
 *	Send (or queue for sending) a packet. 
 *
 *	IMPORTANT: When this is called to resend frames. The caller MUST
 *	already have locked the sk_buff. Apart from that we do the
 *	rest of the magic.
 */

static void do_dev_queue_xmit(struct sk_buff *skb, struct device *dev, int pri)
{
	unsigned long flags;
	struct sk_buff_head *list;
	int retransmission = 0;	/* used to say if the packet should go	*/
				/* at the front or the back of the	*/
				/* queue - front is a retransmit try	*/

	if(pri>=0 && !skb_device_locked(skb))
		skb_device_lock(skb);	/* Shove a lock on the frame */
#if CONFIG_SKB_CHECK 
	IS_SKB(skb);
#endif    
	skb->dev = dev;

	/*
	 *	Negative priority is used to flag a frame that is being pulled from the
	 *	queue front as a retransmit attempt. It therefore goes back on the queue
	 *	start on a failure.
	 */
	 
  	if (pri < 0) 
  	{
		pri = -pri-1;
		retransmission = 1;
  	}

#ifdef CONFIG_NET_DEBUG
	if (pri >= DEV_NUMBUFFS) 
	{
		printk(KERN_WARNING "bad priority in dev_queue_xmit.\n");
		pri = 1;
	}
#endif

	/*
	 *	If the address has not been resolved. Call the device header rebuilder.
	 *	This can cover all protocols and technically not just ARP either.
	 */
	 
	if (!skb->arp && dev->rebuild_header(skb->data, dev, skb->raddr, skb)) {
		return;
	}

	/*
	 *
	 * 	If dev is an alias, switch to its main device.
	 *	"arp" resolution has been made with alias device, so
	 *	arp entries refer to alias, not main.
	 *
	 */

#ifdef CONFIG_NET_ALIAS
	if (net_alias_is(dev))
	  	skb->dev = dev = net_alias_dev_tx(dev);
#endif

	/*
	 *	If we are bridging and this is directly generated output
	 *	pass the frame via the bridge.
	 */

#ifdef CONFIG_BRIDGE
	if(skb->pkt_bridged!=IS_BRIDGED && br_stats.flags & BR_UP)
	{
		if(br_tx_frame(skb))
			return;
	}
#endif

	list = dev->buffs + pri;

	save_flags(flags);
	/* if this isn't a retransmission, use the first packet instead... */
	if (!retransmission) {
		if (skb_queue_len(list)) {
			/* avoid overrunning the device queue.. */
			if (skb_queue_len(list) > dev->tx_queue_len) {
				dev_kfree_skb(skb, FREE_WRITE);
				return;
			}
		}

		/* copy outgoing packets to any sniffer packet handlers */
		if (dev_nit) {
			struct packet_type *ptype;
			skb->stamp=xtime;
			for (ptype = ptype_all; ptype!=NULL; ptype = ptype->next) 
			{
				/* Never send packets back to the socket
				 * they originated from - MvS ([e-mail])
				 */
				if ((ptype->dev == dev || !ptype->dev) &&
				   ((struct sock *)ptype->data != skb->sk))
				{
					struct sk_buff *skb2;
					if ((skb2 = skb_clone(skb, GFP_ATOMIC)) == NULL)
						break;
					/* FIXME?: Wrong when the hard_header_len
					 * is an upper bound. Is this even
					 * used anywhere?
					 */
					skb2->h.raw = skb2->data + dev->hard_header_len;
					/* On soft header devices we
					 * yank the header before mac.raw
					 * back off. This is set by
					 * dev->hard_header().
					 */
					if (dev->flags&IFF_SOFTHEADERS)
						skb_pull(skb2,skb2->mac.raw-skb2->data);
					skb2->mac.raw = skb2->data;
					ptype->func(skb2, skb->dev, ptype);
				}
			}
		}

		if (skb_queue_len(list)) {
			cli();
			skb_device_unlock(skb);		/* Buffer is on the device queue and can be freed safely */
			__skb_queue_tail(list, skb);
			skb = __skb_dequeue(list);
			skb_device_lock(skb);		/* New buffer needs locking down */
			restore_flags(flags);
		}
	}
	if (dev->hard_start_xmit(skb, dev) == 0) {
		/*
		 *	Packet is now solely the responsibility of the driver
		 */
		return;
	}

	/*
	 *	Transmission failed, put skb back into a list. Once on the list it's safe and
	 *	no longer device locked (it can be freed safely from the device queue)
	 */
	cli();
	skb_device_unlock(skb);
	__skb_queue_head(list,skb);
	restore_flags(flags);
}

void dev_queue_xmit(struct sk_buff *skb, struct device *dev, int pri)
{
	start_bh_atomic();
	do_dev_queue_xmit(skb, dev, pri);
	end_bh_atomic();
}

/*
 *	Receive a packet from a device driver and queue it for the upper
 *	(protocol) levels.  It always succeeds. This is the recommended 
 *	interface to use.
 */

void netif_rx(struct sk_buff *skb)
{
	static int dropping = 0;

	/*
	 *	Any received buffers are un-owned and should be discarded
	 *	when freed. These will be updated later as the frames get
	 *	owners.
	 */

	skb->sk = NULL;
	skb->free = 1;
	if(skb->stamp.tv_sec==0)
		skb->stamp = xtime;

	/*
	 *	Check that we aren't overdoing things.
	 */

	if (!backlog_size)
  		dropping = 0;
	else if (backlog_size > 300)
		dropping = 1;

	if (dropping) 
	{
		kfree_skb(skb, FREE_READ);
		return;
	}

	/*
	 *	Add it to the "backlog" queue. 
	 */
#if CONFIG_SKB_CHECK
	IS_SKB(skb);
#endif	
	skb_queue_tail(&backlog,skb);
	backlog_size++;
  
	/*
	 *	If any packet arrived, mark it for processing after the
	 *	hardware interrupt returns.
	 */

	mark_bh(NET_BH);