Merge bk://kernel.bkbits.net/davem/sparc-2.5

into home.transmeta.com:/home/torvalds/v2.5/linux

arch/i386/kernel/cpu/mtrr/if.c

@@ -70,7 +70,7 @@ mtrr_file_del(unsigned long base, unsigned long size,
 
 /* RED-PEN: seq_file can seek now. this is ignored. */
 static ssize_t
-mtrr_write(struct file *file, const char *buf, size_t len, loff_t * ppos)
+mtrr_write(struct file *file, const char __user *buf, size_t len, loff_t * ppos)
 /*  Format of control line:
     "base=%Lx size=%Lx type=%s"     OR:
     "disable=%d"
@@ -133,12 +133,13 @@ mtrr_write(struct file *file, const char *buf, size_t len, loff_t * ppos)
 
 static int
 mtrr_ioctl(struct inode *inode, struct file *file,
-	   unsigned int cmd, unsigned long arg)
+	   unsigned int cmd, unsigned long __arg)
 {
 	int err;
 	mtrr_type type;
 	struct mtrr_sentry sentry;
 	struct mtrr_gentry gentry;
+	void __user *arg = (void __user *) __arg;
 
 	switch (cmd) {
 	default:
@@ -146,7 +147,7 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_ADD_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err =
 		    mtrr_file_add(sentry.base, sentry.size, sentry.type, 1,
@@ -157,7 +158,7 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_SET_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err = mtrr_add(sentry.base, sentry.size, sentry.type, 0);
 		if (err < 0)
@@ -166,7 +167,7 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_DEL_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err = mtrr_file_del(sentry.base, sentry.size, file, 0);
 		if (err < 0)
@@ -175,14 +176,14 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_KILL_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err = mtrr_del(-1, sentry.base, sentry.size);
 		if (err < 0)
 			return err;
 		break;
 	case MTRRIOC_GET_ENTRY:
-		if (copy_from_user(&gentry, (void *) arg, sizeof gentry))
+		if (copy_from_user(&gentry, arg, sizeof gentry))
 			return -EFAULT;
 		if (gentry.regnum >= num_var_ranges)
 			return -EINVAL;
@@ -198,13 +199,13 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 			gentry.type = type;
 		}
 
-		if (copy_to_user((void *) arg, &gentry, sizeof gentry))
+		if (copy_to_user(arg, &gentry, sizeof gentry))
 			return -EFAULT;
 		break;
 	case MTRRIOC_ADD_PAGE_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err =
 		    mtrr_file_add(sentry.base, sentry.size, sentry.type, 1,
@@ -215,7 +216,7 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_SET_PAGE_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err = mtrr_add_page(sentry.base, sentry.size, sentry.type, 0);
 		if (err < 0)
@@ -224,7 +225,7 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_DEL_PAGE_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err = mtrr_file_del(sentry.base, sentry.size, file, 1);
 		if (err < 0)
@@ -233,21 +234,21 @@ mtrr_ioctl(struct inode *inode, struct file *file,
 	case MTRRIOC_KILL_PAGE_ENTRY:
 		if (!capable(CAP_SYS_ADMIN))
 			return -EPERM;
-		if (copy_from_user(&sentry, (void *) arg, sizeof sentry))
+		if (copy_from_user(&sentry, arg, sizeof sentry))
 			return -EFAULT;
 		err = mtrr_del_page(-1, sentry.base, sentry.size);
 		if (err < 0)
 			return err;
 		break;
 	case MTRRIOC_GET_PAGE_ENTRY:
-		if (copy_from_user(&gentry, (void *) arg, sizeof gentry))
+		if (copy_from_user(&gentry, arg, sizeof gentry))
 			return -EFAULT;
 		if (gentry.regnum >= num_var_ranges)
 			return -EINVAL;
 		mtrr_if->get(gentry.regnum, &gentry.base, &gentry.size, &type);
 		gentry.type = type;
 
-		if (copy_to_user((void *) arg, &gentry, sizeof gentry))
+		if (copy_to_user(arg, &gentry, sizeof gentry))
 			return -EFAULT;
 		break;
 	}

crypto/Kconfig

@@ -6,14 +6,16 @@ menu "Cryptographic options"
 
 config CRYPTO
 	bool "Cryptographic API"
-	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m
+	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m || INET6_AH=y || INET6_AH=m || \
+		     INET6_ESP=y || INET6_ESP=m
 	help
 	  This option provides the core Cryptographic API.
 
 config CRYPTO_HMAC
 	bool "HMAC support"
 	depends on CRYPTO
-	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m
+	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m || INET6_AH=y || INET6_AH=m || \
+		     INET6_ESP=y || INET6_ESP=m
 	help
 	  HMAC: Keyed-Hashing for Message Authentication (RFC2104).
 	  This is required for IPSec.
@@ -33,14 +35,16 @@ config CRYPTO_MD4
 config CRYPTO_MD5
 	tristate "MD5 digest algorithm"
 	depends on CRYPTO
-	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m
+	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m || INET6_AH=y || INET6_AH=m || \
+	             INET6_ESP=y || INET6_ESP=m
 	help
 	  MD5 message digest algorithm (RFC1321).
 
 config CRYPTO_SHA1
 	tristate "SHA1 digest algorithm"
 	depends on CRYPTO
-	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m
+	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m  || INET6_AH=y || INET6_AH=m || \
+	             INET6_ESP=y || INET6_ESP=m
 	help
 	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2).
 
@@ -68,7 +72,7 @@ config CRYPTO_SHA512
 config CRYPTO_DES
 	tristate "DES and Triple DES EDE cipher algorithms"
 	depends on CRYPTO
-	default y if INET_AH=y || INET_AH=m || INET_ESP=y || INET_ESP=m
+	default y if INET_ESP=y || INET_ESP=m || INET6_ESP=y || INET6_ESP=m
 	help
 	  DES cipher algorithm (FIPS 46-2), and Triple DES EDE (FIPS 46-3).
 

fs/open.c

@@ -280,7 +280,7 @@ asmlinkage long sys_utime(char __user * filename, struct utimbuf __user * times)
  * must be owner or have write permission.
  * Else, update from *times, must be owner or super user.
  */
-long do_utimes(char __user * filename, struct timeval __user * times)
+long do_utimes(char __user * filename, struct timeval * times)
 {
 	int error;
 	struct nameidata nd;

include/asm-sparc/hardirq.h

@@ -9,7 +9,6 @@
 
 #include <linux/config.h>
 #include <linux/threads.h>
-#include <linux/brlock.h>
 #include <linux/spinlock.h>
 #include <linux/cache.h>
 

include/asm-sparc64/hardirq.h

@@ -8,7 +8,6 @@
 
 #include <linux/config.h>
 #include <linux/threads.h>
-#include <linux/brlock.h>
 #include <linux/spinlock.h>
 #include <linux/cache.h>
 

include/linux/brlock.h

-#ifndef __LINUX_BRLOCK_H
-#define __LINUX_BRLOCK_H
-
-/*
- * 'Big Reader' read-write spinlocks.
- *
- * super-fast read/write locks, with write-side penalty. The point
- * is to have a per-CPU read/write lock. Readers lock their CPU-local
- * readlock, writers must lock all locks to get write access. These
- * CPU-read-write locks are semantically identical to normal rwlocks.
- * Memory usage is higher as well. (NR_CPUS*L1_CACHE_BYTES bytes)
- *
- * The most important feature is that these spinlocks do not cause
- * cacheline ping-pong in the 'most readonly data' case.
- *
- * Copyright 2000, Ingo Molnar <mingo@redhat.com>
- *
- * Registry idea and naming [ crutial! :-) ] by:
- *
- *                 David S. Miller <davem@redhat.com>
- *
- * David has an implementation that doesn't use atomic operations in
- * the read branch via memory ordering tricks - i guess we need to
- * split this up into a per-arch thing? The atomicity issue is a
- * secondary item in profiles, at least on x86 platforms.
- *
- * The atomic op version overhead is indeed a big deal on
- * load-locked/store-conditional cpus (ALPHA/MIPS/PPC) and
- * compare-and-swap cpus (Sparc64).  So we control which
- * implementation to use with a __BRLOCK_USE_ATOMICS define. -DaveM
- *
- */
-
-/* Register bigreader lock indices here. */
-enum brlock_indices {
-	BR_NETPROTO_LOCK,
-	__BR_END
-};
-
-#include <linux/config.h>
-
-#ifdef CONFIG_SMP
-
-#include <linux/cache.h>
-#include <linux/spinlock.h>
-
-#if defined(__i386__) || defined(__ia64__) || defined(__x86_64__)
-#define __BRLOCK_USE_ATOMICS
-#else
-#undef __BRLOCK_USE_ATOMICS
-#endif
-
-#ifdef __BRLOCK_USE_ATOMICS
-typedef rwlock_t	brlock_read_lock_t;
-#else
-typedef unsigned int	brlock_read_lock_t;
-#endif
-
-/*
- * align last allocated index to the next cacheline:
- */
-#define __BR_IDX_MAX \
-	(((sizeof(brlock_read_lock_t)*__BR_END + SMP_CACHE_BYTES-1) & ~(SMP_CACHE_BYTES-1)) / sizeof(brlock_read_lock_t))
-
-extern brlock_read_lock_t __brlock_array[NR_CPUS][__BR_IDX_MAX];
-
-#ifndef __BRLOCK_USE_ATOMICS
-struct br_wrlock {
-	spinlock_t lock;
-} __attribute__ ((__aligned__(SMP_CACHE_BYTES)));
-
-extern struct br_wrlock __br_write_locks[__BR_IDX_MAX];
-#endif
-
-extern void __br_lock_usage_bug (void);
-
-#ifdef __BRLOCK_USE_ATOMICS
-
-static inline void br_read_lock (enum brlock_indices idx)
-{
-	/*
-	 * This causes a link-time bug message if an
-	 * invalid index is used:
-	 */
-	if (idx >= __BR_END)
-		__br_lock_usage_bug();
-
-	preempt_disable();
-	_raw_read_lock(&__brlock_array[smp_processor_id()][idx]);
-}
-
-static inline void br_read_unlock (enum brlock_indices idx)
-{
-	if (idx >= __BR_END)
-		__br_lock_usage_bug();
-
-	read_unlock(&__brlock_array[smp_processor_id()][idx]);
-}
-
-#else /* ! __BRLOCK_USE_ATOMICS */
-static inline void br_read_lock (enum brlock_indices idx)
-{
-	unsigned int *ctr;
-	spinlock_t *lock;
-
-	/*
-	 * This causes a link-time bug message if an
-	 * invalid index is used:
-	 */
-	if (idx >= __BR_END)
-		__br_lock_usage_bug();
-
-	preempt_disable();
-	ctr = &__brlock_array[smp_processor_id()][idx];
-	lock = &__br_write_locks[idx].lock;
-again:
-	(*ctr)++;
-	mb();
-	if (spin_is_locked(lock)) {
-		(*ctr)--;
-		wmb(); /*
-			* The release of the ctr must become visible
-			* to the other cpus eventually thus wmb(),
-			* we don't care if spin_is_locked is reordered
-			* before the releasing of the ctr.
-			* However IMHO this wmb() is superflous even in theory.
-			* It would not be superflous only if on the
-			* other CPUs doing a ldl_l instead of an ldl
-			* would make a difference and I don't think this is
-			* the case.
-			* I'd like to clarify this issue further
-			* but for now this is a slow path so adding the
-			* wmb() will keep us on the safe side.
-			*/
-		while (spin_is_locked(lock))
-			barrier();
-		goto again;
-	}
-}
-
-static inline void br_read_unlock (enum brlock_indices idx)
-{
-	unsigned int *ctr;
-
-	if (idx >= __BR_END)
-		__br_lock_usage_bug();
-
-	ctr = &__brlock_array[smp_processor_id()][idx];
-
-	wmb();
-	(*ctr)--;
-	preempt_enable();
-}
-#endif /* __BRLOCK_USE_ATOMICS */
-
-/* write path not inlined - it's rare and larger */
-
-extern void FASTCALL(__br_write_lock (enum brlock_indices idx));
-extern void FASTCALL(__br_write_unlock (enum brlock_indices idx));
-
-static inline void br_write_lock (enum brlock_indices idx)
-{
-	if (idx >= __BR_END)
-		__br_lock_usage_bug();
-	__br_write_lock(idx);
-}
-
-static inline void br_write_unlock (enum brlock_indices idx)
-{
-	if (idx >= __BR_END)
-		__br_lock_usage_bug();
-	__br_write_unlock(idx);
-}
-
-#else
-# define br_read_lock(idx)	({ (void)(idx); preempt_disable(); })
-# define br_read_unlock(idx)	({ (void)(idx); preempt_enable(); })
-# define br_write_lock(idx)	({ (void)(idx); preempt_disable(); })
-# define br_write_unlock(idx)	({ (void)(idx); preempt_enable(); })
-#endif	/* CONFIG_SMP */
-
-/*
- * Now enumerate all of the possible sw/hw IRQ protected
- * versions of the interfaces.
- */
-#define br_read_lock_irqsave(idx, flags) \
-	do { local_irq_save(flags); br_read_lock(idx); } while (0)
-
-#define br_read_lock_irq(idx) \
-	do { local_irq_disable(); br_read_lock(idx); } while (0)
-
-#define br_read_lock_bh(idx) \
-	do { local_bh_disable(); br_read_lock(idx); } while (0)
-
-#define br_write_lock_irqsave(idx, flags) \
-	do { local_irq_save(flags); br_write_lock(idx); } while (0)
-
-#define br_write_lock_irq(idx) \
-	do { local_irq_disable(); br_write_lock(idx); } while (0)
-
-#define br_write_lock_bh(idx) \
-	do { local_bh_disable(); br_write_lock(idx); } while (0)
-
-#define br_read_unlock_irqrestore(idx, flags) \
-	do { br_read_unlock(irx); local_irq_restore(flags); } while (0)
-
-#define br_read_unlock_irq(idx) \
-	do { br_read_unlock(idx); local_irq_enable(); } while (0)
-
-#define br_read_unlock_bh(idx) \
-	do { br_read_unlock(idx); local_bh_enable(); } while (0)
-
-#define br_write_unlock_irqrestore(idx, flags) \
-	do { br_write_unlock(irx); local_irq_restore(flags); } while (0)
-
-#define br_write_unlock_irq(idx) \
-	do { br_write_unlock(idx); local_irq_enable(); } while (0)
-
-#define br_write_unlock_bh(idx) \
-	do { br_write_unlock(idx); local_bh_enable(); } while (0)
-
-#endif /* __LINUX_BRLOCK_H */

include/linux/fs.h

@@ -1017,7 +1017,7 @@ static inline int break_lease(struct inode *inode, unsigned int mode)
 
 /* fs/open.c */
 
-asmlinkage long sys_open(const char *, int, int);
+asmlinkage long sys_open(const char __user *, int, int);
 asmlinkage long sys_close(unsigned int);	/* yes, it's really unsigned */
 extern int do_truncate(struct dentry *, loff_t start);
 

include/linux/ipv6.h

@@ -121,6 +121,7 @@ struct ipv6hdr {
 #include <linux/icmpv6.h>
 #include <net/if_inet6.h>       /* struct ipv6_mc_socklist */
 #include <linux/tcp.h>
+#include <linux/udp.h>
 
 /* 
    This structure contains results of exthdrs parsing
@@ -178,6 +179,11 @@ struct ipv6_pinfo {
 
 	struct ipv6_txoptions	*opt;
 	struct sk_buff		*pktoptions;
+	struct {
+		struct ipv6_txoptions *opt;
+		struct rt6_info	*rt;
+		struct flowi *fl;
+	} cork;
 };
 
 struct raw6_opt {
@@ -200,6 +206,7 @@ struct udp6_sock {
 	struct sock	  sk;
 	struct ipv6_pinfo *pinet6;
 	struct inet_opt   inet;
+	struct udp_opt	  udp;
 	struct ipv6_pinfo inet6;
 };
 

include/linux/netdevice.h

@@ -456,7 +456,7 @@ struct packet_type
 	int			(*func) (struct sk_buff *, struct net_device *,
 					 struct packet_type *);
 	void			*data;	/* Private to the packet type		*/
-	struct packet_type	*next;
+	struct list_head	list;
 };
 
 
@@ -472,6 +472,7 @@ extern int 			netdev_boot_setup_check(struct net_device *dev);
 extern struct net_device    *dev_getbyhwaddr(unsigned short type, char *hwaddr);
 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);
 extern struct net_device	*dev_get_by_flags(unsigned short flags,
 						  unsigned short mask);

include/linux/skbuff.h

@@ -792,6 +792,15 @@ static inline int skb_pagelen(const struct sk_buff *skb)
 	return len + skb_headlen(skb);
 }
 
+static inline void skb_fill_page_desc(struct sk_buff *skb, int i, struct page *page, int off, int size)
+{
+	skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+	frag->page = page;
+	frag->page_offset = off;
+	frag->size = size;
+	skb_shinfo(skb)->nr_frags = i+1;
+}
+
 #define SKB_PAGE_ASSERT(skb) do { if (skb_shinfo(skb)->nr_frags) \
 					BUG(); } while (0)
 #define SKB_FRAG_ASSERT(skb) do { if (skb_shinfo(skb)->frag_list) \

include/linux/time.h

@@ -204,7 +204,7 @@ extern void do_settimeofday(struct timeval *tv);
 extern int do_sys_settimeofday(struct timeval *tv, struct timezone *tz);
 extern void clock_was_set(void); // call when ever the clock is set
 extern long do_nanosleep(struct timespec *t);
-extern long do_utimes(char * filename, struct timeval * times);
+extern long do_utimes(char __user * filename, struct timeval * times);
 #endif
 
 #define FD_SETSIZE		__FD_SETSIZE

include/net/ipv6.h

@@ -316,6 +316,26 @@ extern int			ip6_build_xmit(struct sock *sk,
 					       struct ipv6_txoptions *opt,
 					       int hlimit, int flags);
 
+extern int			ip6_append_data(struct sock *sk,
+						int getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb),
+		    				void *from,
+						int length,
+						int transhdrlen,
+		      				int hlimit,
+						struct ipv6_txoptions *opt,
+						struct flowi *fl,
+						struct rt6_info *rt,
+						unsigned int flags);
+
+extern int			ip6_push_pending_frames(struct sock *sk);
+
+extern void			ip6_flush_pending_frames(struct sock *sk);
+
+extern int			ip6_dst_lookup(struct sock *sk,
+					       struct dst_entry **dst,
+					       struct flowi *fl,
+					       struct in6_addr **saddr);
+
 /*
  *	skb processing functions
  */

include/net/xfrm.h

@@ -123,6 +123,12 @@ struct xfrm_state
 	/* Data for encapsulator */
 	struct xfrm_encap_tmpl	*encap;
 
+	/* IPComp needs an IPIP tunnel for handling uncompressed packets */
+	struct xfrm_state	*tunnel;
+
+	/* If a tunnel, number of users + 1 */
+	atomic_t		tunnel_users;
+
 	/* State for replay detection */
 	struct xfrm_replay_state replay;
 
@@ -196,6 +202,8 @@ extern int xfrm_state_unregister_afinfo(struct xfrm_state_afinfo *afinfo);
 extern struct xfrm_state_afinfo *xfrm_state_get_afinfo(unsigned short family);
 extern void xfrm_state_put_afinfo(struct xfrm_state_afinfo *afinfo);
 
+extern void xfrm_state_delete_tunnel(struct xfrm_state *x);
+
 struct xfrm_decap_state;
 struct xfrm_type
 {
@@ -699,6 +707,11 @@ xfrm_state_addr_check(struct xfrm_state *x,
 	return 0;
 }
 
+static inline int xfrm_state_kern(struct xfrm_state *x)
+{
+	return atomic_read(&x->tunnel_users);
+}
+
 /*
  * xfrm algorithm information
  */

kernel/ksyms.c

@@ -40,7 +40,6 @@
 #include <linux/mm.h>
 #include <linux/capability.h>
 #include <linux/highuid.h>
-#include <linux/brlock.h>
 #include <linux/fs.h>
 #include <linux/uio.h>
 #include <linux/tty.h>
@@ -429,17 +428,6 @@ EXPORT_SYMBOL(del_timer_sync);
 #endif
 EXPORT_SYMBOL(mod_timer);
 
-#ifdef CONFIG_SMP
-
-/* Big-Reader lock implementation */
-EXPORT_SYMBOL(__brlock_array);
-#ifndef __BRLOCK_USE_ATOMICS
-EXPORT_SYMBOL(__br_write_locks);
-#endif
-EXPORT_SYMBOL(__br_write_lock);
-EXPORT_SYMBOL(__br_write_unlock);
-#endif
-
 #ifdef HAVE_DISABLE_HLT
 EXPORT_SYMBOL(disable_hlt);
 EXPORT_SYMBOL(enable_hlt);

lib/Makefile

@@ -8,7 +8,7 @@
 
 L_TARGET := lib.a
 
-obj-y := errno.o ctype.o string.o vsprintf.o brlock.o cmdline.o \
+obj-y := errno.o ctype.o string.o vsprintf.o cmdline.o \
 	 bust_spinlocks.o rbtree.o radix-tree.o dump_stack.o \
 	 kobject.o idr.o
 

lib/brlock.c

-/*
- *
- * linux/lib/brlock.c
- *
- * 'Big Reader' read-write spinlocks.  See linux/brlock.h for details.
- *
- * Copyright 2000, Ingo Molnar <mingo@redhat.com>
- * Copyright 2000, David S. Miller <davem@redhat.com>
- */
-
-#include <linux/config.h>
-
-#ifdef CONFIG_SMP
-
-#include <linux/sched.h>
-#include <linux/brlock.h>
-
-#ifdef __BRLOCK_USE_ATOMICS
-
-brlock_read_lock_t __brlock_array[NR_CPUS][__BR_IDX_MAX] =
-   { [0 ... NR_CPUS-1] = { [0 ... __BR_IDX_MAX-1] = RW_LOCK_UNLOCKED } };
-
-void __br_write_lock (enum brlock_indices idx)
-{
-	int i;
-
-	preempt_disable();
-	for (i = 0; i < NR_CPUS; i++)
-		_raw_write_lock(&__brlock_array[i][idx]);
-}
-
-void __br_write_unlock (enum brlock_indices idx)
-{
-	int i;
-
-	for (i = 0; i < NR_CPUS; i++)
-		_raw_write_unlock(&__brlock_array[i][idx]);
-	preempt_enable();
-}
-
-#else /* ! __BRLOCK_USE_ATOMICS */
-
-brlock_read_lock_t __brlock_array[NR_CPUS][__BR_IDX_MAX] =
-   { [0 ... NR_CPUS-1] = { [0 ... __BR_IDX_MAX-1] = 0 } };
-
-struct br_wrlock __br_write_locks[__BR_IDX_MAX] =
-   { [0 ... __BR_IDX_MAX-1] = { SPIN_LOCK_UNLOCKED } };
-
-void __br_write_lock (enum brlock_indices idx)
-{
-	int i;
-
-	preempt_disable();
-again:
-	_raw_spin_lock(&__br_write_locks[idx].lock);
-	for (i = 0; i < NR_CPUS; i++)
-		if (__brlock_array[i][idx] != 0) {
-			_raw_spin_unlock(&__br_write_locks[idx].lock);
-			barrier();
-			cpu_relax();
-			goto again;
-		}
-}
-
-void __br_write_unlock (enum brlock_indices idx)
-{
-	spin_unlock(&__br_write_locks[idx].lock);
-}
-
-#endif /* __BRLOCK_USE_ATOMICS */
-
-#endif /* CONFIG_SMP */

mm/page_alloc.c

@@ -792,7 +792,7 @@ unsigned int nr_free_pagecache_pages(void)
 	return nr_free_zone_pages(GFP_HIGHUSER & GFP_ZONEMASK);
 }
 
-#if CONFIG_HIGHMEM
+#ifdef CONFIG_HIGHMEM
 unsigned int nr_free_highpages (void)
 {
 	pg_data_t *pgdat;

net/bridge/br_fdb.c

@@ -20,25 +20,19 @@
 #include <asm/uaccess.h>
 #include "br_private.h"
 
-static __inline__ unsigned long __timeout(struct net_bridge *br)
+/* if topology_changing then use forward_delay (default 15 sec)
+ * otherwise keep longer (default 5 minutes)
+ */
+static __inline__ unsigned long hold_time(const struct net_bridge *br)
 {
-	unsigned long timeout;
-
-	timeout = jiffies - br->ageing_time;
-	if (br->topology_change)
-		timeout = jiffies - br->forward_delay;
-
-	return timeout;
+	return br->topology_change ? br->forward_delay : br->ageing_time;
 }
 
-static __inline__ int has_expired(struct net_bridge *br,
-				  struct net_bridge_fdb_entry *fdb)
+static __inline__ int has_expired(const struct net_bridge *br,
+				  const struct net_bridge_fdb_entry *fdb)
 {
-	if (!fdb->is_static &&
-	    time_before_eq(fdb->ageing_timer, __timeout(br)))
-		return 1;
-
-	return 0;
+	return !fdb->is_static 
+		&& time_before_eq(fdb->ageing_timer + hold_time(br), jiffies);
 }
 
 static __inline__ void copy_fdb(struct __fdb_entry *ent, 
@@ -52,7 +46,7 @@ static __inline__ void copy_fdb(struct __fdb_entry *ent,
 		: ((jiffies - f->ageing_timer) * USER_HZ) / HZ;
 }
 
-static __inline__ int br_mac_hash(unsigned char *mac)
+static __inline__ int br_mac_hash(const unsigned char *mac)
 {
 	unsigned long x;
 
@@ -68,7 +62,14 @@ static __inline__ int br_mac_hash(unsigned char *mac)
 	return x & (BR_HASH_SIZE - 1);
 }
 
-void br_fdb_changeaddr(struct net_bridge_port *p, unsigned char *newaddr)
+static __inline__ void fdb_delete(struct net_bridge_fdb_entry *f)
+{
+	hlist_del(&f->hlist);
+	list_del(&f->age_list);
+	br_fdb_put(f);
+}
+
+void br_fdb_changeaddr(struct net_bridge_port *p, const unsigned char *newaddr)
 {
 	struct net_bridge *br;
 	int i;
@@ -98,25 +99,29 @@ void br_fdb_changeaddr(struct net_bridge_port *p, unsigned char *newaddr)
 	write_unlock_bh(&br->hash_lock);
 }
 
-void br_fdb_cleanup(struct net_bridge *br)
+void br_fdb_cleanup(unsigned long _data)
 {
-	int i;
-	unsigned long timeout;
-
-	timeout = __timeout(br);
+	struct net_bridge *br = (struct net_bridge *)_data;
+	struct list_head *l, *n;
+	unsigned long delay;
 
 	write_lock_bh(&br->hash_lock);
-	for (i=0;i<BR_HASH_SIZE;i++) {
-		struct hlist_node *h, *g;
-		
-		hlist_for_each_safe(h, g, &br->hash[i]) {
-			struct net_bridge_fdb_entry *f
-				= hlist_entry(h, struct net_bridge_fdb_entry, hlist);
-			if (!f->is_static &&
-			    time_before_eq(f->ageing_timer, timeout)) {
-				hlist_del(&f->hlist);
-				br_fdb_put(f);
+	delay = hold_time(br);
+
+	list_for_each_safe(l, n, &br->age_list) {
+		struct net_bridge_fdb_entry *f
+			= list_entry(l, struct net_bridge_fdb_entry, age_list);
+		unsigned long expires = f->ageing_timer + delay;
+
+		if (time_before_eq(expires, jiffies)) {
+			if (!f->is_static) {
+				pr_debug("expire age %lu jiffies %lu\n",
+					 f->ageing_timer, jiffies);
+				fdb_delete(f);
 			}
+		} else {
+			mod_timer(&br->gc_timer, expires);
+			break;
 		}
 	}
 	write_unlock_bh(&br->hash_lock);
@@ -134,8 +139,7 @@ void br_fdb_delete_by_port(struct net_bridge *br, struct net_bridge_port *p)
 			struct net_bridge_fdb_entry *f
 				= hlist_entry(h, struct net_bridge_fdb_entry, hlist);
 			if (f->dst == p) {
-				hlist_del(&f->hlist);
-				br_fdb_put(f);
+				fdb_delete(f);
 			}
 		}
 	}
@@ -237,55 +241,46 @@ int br_fdb_get_entries(struct net_bridge *br,
 	return num;
 }
 
-static __inline__ void __fdb_possibly_replace(struct net_bridge_fdb_entry *fdb,
-					      struct net_bridge_port *source,
-					      int is_local)
-{
-	if (!fdb->is_static || is_local) {
-		fdb->dst = source;
-		fdb->is_local = is_local;
-		fdb->is_static = is_local;
-		fdb->ageing_timer = jiffies;
-	}
-}
-
-void br_fdb_insert(struct net_bridge *br,
-		   struct net_bridge_port *source,
-		   unsigned char *addr,
-		   int is_local)
+void br_fdb_insert(struct net_bridge *br, struct net_bridge_port *source,
+		   const unsigned char *addr, int is_local)
 {
 	struct hlist_node *h;
 	struct net_bridge_fdb_entry *fdb;
-	int hash;
-
-	hash = br_mac_hash(addr);
+	int hash = br_mac_hash(addr);
 
 	write_lock_bh(&br->hash_lock);
 	hlist_for_each(h, &br->hash[hash]) {
 		fdb = hlist_entry(h, struct net_bridge_fdb_entry, hlist);
 		if (!fdb->is_local &&
 		    !memcmp(fdb->addr.addr, addr, ETH_ALEN)) {
-			__fdb_possibly_replace(fdb, source, is_local);
-			write_unlock_bh(&br->hash_lock);
-			return;
+			if (likely(!fdb->is_static || is_local)) {
+				/* move to end of age list */
+				list_del(&fdb->age_list);
+				goto update;
+			}
+			goto out;
 		}
-
 	}
 
 	fdb = kmalloc(sizeof(*fdb), GFP_ATOMIC);
-	if (fdb == NULL) {
-		write_unlock_bh(&br->hash_lock);
-		return;
-	}
+	if (fdb == NULL) 
+		goto out;
 
 	memcpy(fdb->addr.addr, addr, ETH_ALEN);
 	atomic_set(&fdb->use_count, 1);
+	hlist_add_head(&fdb->hlist, &br->hash[hash]);
+
+	if (!timer_pending(&br->gc_timer)) {
+		br->gc_timer.expires = jiffies + hold_time(br);
+		add_timer(&br->gc_timer);
+	}
+
+ update:
 	fdb->dst = source;
 	fdb->is_local = is_local;
 	fdb->is_static = is_local;
 	fdb->ageing_timer = jiffies;
-
-	hlist_add_head(&fdb->hlist, &br->hash[hash]);
-
+	list_add_tail(&fdb->age_list, &br->age_list);
+ out:
 	write_unlock_bh(&br->hash_lock);
 }

net/bridge/br_if.c

@@ -84,8 +84,6 @@ static struct net_bridge *new_nb(const char *name)
 	memset(br, 0, sizeof(*br));
 	dev = &br->dev;
 
-	init_timer(&br->tick);
-
 	strncpy(dev->name, name, IFNAMSIZ);
 	dev->priv = br;
 	dev->priv_flags = IFF_EBRIDGE;
@@ -109,12 +107,10 @@ static struct net_bridge *new_nb(const char *name)
 	br->bridge_forward_delay = br->forward_delay = 15 * HZ;
 	br->topology_change = 0;
 	br->topology_change_detected = 0;
-	br_timer_clear(&br->hello_timer);
-	br_timer_clear(&br->tcn_timer);
-	br_timer_clear(&br->topology_change_timer);
-
 	br->ageing_time = 300 * HZ;
-	br->gc_interval = 4 * HZ;
+	INIT_LIST_HEAD(&br->age_list);
+
+	br_stp_timer_init(br);
 
 	return br;
 }

net/bridge/br_ioctl.c

@@ -32,9 +32,10 @@ static inline unsigned long ticks_to_user(unsigned long tick)
 }
 
 /* Report time remaining in user HZ  */
-static unsigned long timer_residue(const struct br_timer *timer)
+static unsigned long timer_residue(const struct timer_list *timer)
 {
-	return ticks_to_user(timer->running ? (jiffies - timer->expires) : 0);
+	return ticks_to_user(timer_pending(timer) 
+			     ? (timer->expires - jiffies) : 0);
 }
 
 static int br_ioctl_device(struct net_bridge *br,
@@ -87,7 +88,6 @@ static int br_ioctl_device(struct net_bridge *br,
 		b.root_port = br->root_port;
 		b.stp_enabled = br->stp_enabled;
 		b.ageing_time = ticks_to_user(br->ageing_time);
-		b.gc_interval = ticks_to_user(br->gc_interval);
 		b.hello_timer_value = timer_residue(&br->hello_timer);
 		b.tcn_timer_value = timer_residue(&br->tcn_timer);
 		b.topology_change_timer_value = timer_residue(&br->topology_change_timer);
@@ -146,8 +146,7 @@ static int br_ioctl_device(struct net_bridge *br,
 		br->ageing_time = user_to_ticks(arg0);
 		return 0;
 
-	case BRCTL_SET_GC_INTERVAL:
-		br->gc_interval = user_to_ticks(arg0);
+	case BRCTL_SET_GC_INTERVAL:	 /* no longer used */
 		return 0;
 
 	case BRCTL_GET_PORT_INFO:

net/bridge/br_private.h

@@ -18,7 +18,6 @@
 #include <linux/netdevice.h>
 #include <linux/miscdevice.h>
 #include <linux/if_bridge.h>
-#include "br_private_timer.h"
 
 #define BR_HASH_BITS 8
 #define BR_HASH_SIZE (1 << BR_HASH_BITS)
@@ -44,10 +43,11 @@ struct mac_addr
 struct net_bridge_fdb_entry
 {
 	struct hlist_node		hlist;
-	atomic_t			use_count;
-	mac_addr			addr;
 	struct net_bridge_port		*dst;
+	struct list_head		age_list;
+	atomic_t			use_count;
 	unsigned long			ageing_timer;
+	mac_addr			addr;
 	unsigned			is_local:1;
 	unsigned			is_static:1;
 };
@@ -71,10 +71,9 @@ struct net_bridge_port
 	unsigned			config_pending:1;
 	int				priority;
 
-	struct br_timer			forward_delay_timer;
-	struct br_timer			hold_timer;
-	struct br_timer			message_age_timer;
-
+	struct timer_list		forward_delay_timer;
+	struct timer_list		hold_timer;
+	struct timer_list		message_age_timer;
 	struct rcu_head			rcu;
 };
 
@@ -86,7 +85,7 @@ struct net_bridge
 	struct net_device_stats		statistics;
 	rwlock_t			hash_lock;
 	struct hlist_head		hash[BR_HASH_SIZE];
-	struct timer_list		tick;
+	struct list_head		age_list;
 
 	/* STP */
 	bridge_id			designated_root;
@@ -103,13 +102,12 @@ struct net_bridge
 	unsigned			topology_change:1;
 	unsigned			topology_change_detected:1;
 
-	struct br_timer			hello_timer;
-	struct br_timer			tcn_timer;
-	struct br_timer			topology_change_timer;
-	struct br_timer			gc_timer;
+	struct timer_list		hello_timer;
+	struct timer_list		tcn_timer;
+	struct timer_list		topology_change_timer;
+	struct timer_list		gc_timer;
 
 	int				ageing_time;
-	int				gc_interval;
 };
 
 extern struct notifier_block br_device_notifier;
@@ -128,8 +126,8 @@ extern int br_dev_xmit(struct sk_buff *skb, struct net_device *dev);
 
 /* br_fdb.c */
 extern void br_fdb_changeaddr(struct net_bridge_port *p,
-		       unsigned char *newaddr);
-extern void br_fdb_cleanup(struct net_bridge *br);
+			      const unsigned char *newaddr);
+extern void br_fdb_cleanup(unsigned long arg);
 extern void br_fdb_delete_by_port(struct net_bridge *br,
 			   struct net_bridge_port *p);
 extern struct net_bridge_fdb_entry *br_fdb_get(struct net_bridge *br,
@@ -140,9 +138,9 @@ extern int  br_fdb_get_entries(struct net_bridge *br,
 			int maxnum,
 			int offset);
 extern void br_fdb_insert(struct net_bridge *br,
-		   struct net_bridge_port *source,
-		   unsigned char *addr,
-		   int is_local);
+			  struct net_bridge_port *source,
+			  const unsigned char *addr,
+			  int is_local);
 
 /* br_forward.c */
 extern void br_deliver(const struct net_bridge_port *to,
@@ -188,10 +186,10 @@ extern int br_netfilter_init(void);
 extern void br_netfilter_fini(void);
 
 /* br_stp.c */
+extern void br_log_state(const struct net_bridge_port *p);
 extern struct net_bridge_port *br_get_port(struct net_bridge *br,
 				    int port_no);
 extern void br_init_port(struct net_bridge_port *p);
-extern port_id br_make_port_id(struct net_bridge_port *p);
 extern void br_become_designated_port(struct net_bridge_port *p);
 
 /* br_stp_if.c */
@@ -210,4 +208,8 @@ extern void br_stp_set_path_cost(struct net_bridge_port *p,
 /* br_stp_bpdu.c */
 extern void br_stp_handle_bpdu(struct sk_buff *skb);
 
+/* br_stp_timer.c */
+extern void br_stp_timer_init(struct net_bridge *br);
+extern void br_stp_port_timer_init(struct net_bridge_port *p);
+
 #endif

net/bridge/br_private_stp.h

@@ -47,7 +47,6 @@ extern void br_configuration_update(struct net_bridge *);
 extern void br_port_state_selection(struct net_bridge *);
 extern void br_received_config_bpdu(struct net_bridge_port *p, struct br_config_bpdu *bpdu);
 extern void br_received_tcn_bpdu(struct net_bridge_port *p);
-extern void br_tick(unsigned long __data);
 extern void br_transmit_config(struct net_bridge_port *p);
 extern void br_transmit_tcn(struct net_bridge *br);
 extern void br_topology_change_detection(struct net_bridge *br);

net/bridge/br_private_timer.h

-/*
- *	Linux ethernet bridge
- *
- *	Authors:
- *	Lennert Buytenhek		<buytenh@gnu.org>
- *
- *	$Id: br_private_timer.h,v 1.1 2000/02/18 16:47:13 davem Exp $
- *
- *	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 _BR_PRIVATE_TIMER_H
-#define _BR_PRIVATE_TIMER_H
-
-struct br_timer
-{
-	int running;
-	unsigned long expires;
-};
-
-extern __inline__ void br_timer_clear(struct br_timer *t)
-{
-	t->running = 0;
-}
-
-extern __inline__ unsigned long br_timer_get_residue(struct br_timer *t)
-{
-	if (t->running)
-		return jiffies - t->expires;
-
-	return 0;
-}
-
-extern __inline__ void br_timer_set(struct br_timer *t, unsigned long x)
-{
-	t->expires = x;
-	t->running = 1;
-}
-
-extern __inline__ int br_timer_is_running(struct br_timer *t)
-{
-	return t->running;
-}
-
-extern __inline__ int br_timer_has_expired(struct br_timer *t, unsigned long to)
-{
-	return t->running && time_after_eq(jiffies, t->expires + to);
-}
-
-
-#endif

net/bridge/br_stp.c

@@ -12,7 +12,6 @@
  *	as published by the Free Software Foundation; either version
  *	2 of the License, or (at your option) any later version.
  */
-
 #include <linux/kernel.h>
 #include <linux/if_bridge.h>
 #include <linux/smp_lock.h>
@@ -20,6 +19,18 @@
 #include "br_private.h"
 #include "br_private_stp.h"
 
+static const char *br_port_state_names[] = {
+	"disabled", "learning", "forwarding", "blocking",
+};
+
+void br_log_state(const struct net_bridge_port *p)
+{
+	pr_info("%s: port %d(%s) entering %s state\n",
+		p->br->dev.name, p->port_no, p->dev->name, 
+		br_port_state_names[p->state]);
+
+}
+
 /* called under bridge lock */
 struct net_bridge_port *br_get_port(struct net_bridge *br, int port_no)
 {
@@ -34,7 +45,8 @@ struct net_bridge_port *br_get_port(struct net_bridge *br, int port_no)
 }
 
 /* called under bridge lock */
-static int br_should_become_root_port(struct net_bridge_port *p, int root_port)
+static int br_should_become_root_port(const struct net_bridge_port *p, 
+				      int root_port)
 {
 	struct net_bridge *br;
 	struct net_bridge_port *rp;
@@ -116,9 +128,12 @@ void br_become_root_bridge(struct net_bridge *br)
 	br->hello_time = br->bridge_hello_time;
 	br->forward_delay = br->bridge_forward_delay;
 	br_topology_change_detection(br);
-	br_timer_clear(&br->tcn_timer);
-	br_config_bpdu_generation(br);
-	br_timer_set(&br->hello_timer, jiffies);
+	del_timer(&br->tcn_timer);
+
+	if (br->dev.flags & IFF_UP) {
+		br_config_bpdu_generation(br);
+		mod_timer(&br->hello_timer, jiffies + br->hello_time);
+	}
 }
 
 /* called under bridge lock */
@@ -127,7 +142,8 @@ void br_transmit_config(struct net_bridge_port *p)
 	struct br_config_bpdu bpdu;
 	struct net_bridge *br;
 
-	if (br_timer_is_running(&p->hold_timer)) {
+
+	if (timer_pending(&p->hold_timer)) {
 		p->config_pending = 1;
 		return;
 	}
@@ -142,12 +158,11 @@ void br_transmit_config(struct net_bridge_port *p)
 	bpdu.port_id = p->port_id;
 	bpdu.message_age = 0;
 	if (!br_is_root_bridge(br)) {
-		struct net_bridge_port *root;
-		unsigned long age;
+		struct net_bridge_port *root
+			= br_get_port(br, br->root_port);
+		bpdu.max_age = root->message_age_timer.expires - jiffies;
 
-		root = br_get_port(br, br->root_port);
-		age = br_timer_get_residue(&root->message_age_timer) + 1;
-		bpdu.message_age = age;
+		if (bpdu.max_age <= 0) bpdu.max_age = 1;
 	}
 	bpdu.max_age = br->max_age;
 	bpdu.hello_time = br->hello_time;
@@ -157,22 +172,26 @@ void br_transmit_config(struct net_bridge_port *p)
 
 	p->topology_change_ack = 0;
 	p->config_pending = 0;
-	br_timer_set(&p->hold_timer, jiffies);
+	
+	mod_timer(&p->hold_timer, jiffies + BR_HOLD_TIME);
 }
 
 /* called under bridge lock */
-static void br_record_config_information(struct net_bridge_port *p, struct br_config_bpdu *bpdu)
+static inline void br_record_config_information(struct net_bridge_port *p, 
+						const struct br_config_bpdu *bpdu)
 {
 	p->designated_root = bpdu->root;
 	p->designated_cost = bpdu->root_path_cost;
 	p->designated_bridge = bpdu->bridge_id;
 	p->designated_port = bpdu->port_id;
 
-	br_timer_set(&p->message_age_timer, jiffies - bpdu->message_age);
+	mod_timer(&p->message_age_timer, jiffies 
+		  + (p->br->max_age - bpdu->message_age));
 }
 
 /* called under bridge lock */
-static void br_record_config_timeout_values(struct net_bridge *br, struct br_config_bpdu *bpdu)
+static inline void br_record_config_timeout_values(struct net_bridge *br, 
+					    const struct br_config_bpdu *bpdu)
 {
 	br->max_age = bpdu->max_age;
 	br->hello_time = bpdu->hello_time;
@@ -187,7 +206,7 @@ void br_transmit_tcn(struct net_bridge *br)
 }
 
 /* called under bridge lock */
-static int br_should_become_designated_port(struct net_bridge_port *p)
+static int br_should_become_designated_port(const struct net_bridge_port *p)
 {
 	struct net_bridge *br;
 	int t;
@@ -261,25 +280,28 @@ static int br_supersedes_port_info(struct net_bridge_port *p, struct br_config_b
 }
 
 /* called under bridge lock */
-static void br_topology_change_acknowledged(struct net_bridge *br)
+static inline void br_topology_change_acknowledged(struct net_bridge *br)
 {
 	br->topology_change_detected = 0;
-	br_timer_clear(&br->tcn_timer);
+	del_timer(&br->tcn_timer);
 }
 
 /* called under bridge lock */
 void br_topology_change_detection(struct net_bridge *br)
 {
-	printk(KERN_INFO "%s: topology change detected", br->dev.name);
+	if (!(br->dev.flags & IFF_UP))
+		return;
 
+	pr_info("%s: topology change detected", br->dev.name);
 	if (br_is_root_bridge(br)) {
 		printk(", propagating");
 		br->topology_change = 1;
-		br_timer_set(&br->topology_change_timer, jiffies);
+		mod_timer(&br->topology_change_timer, jiffies
+			  + br->bridge_forward_delay + br->bridge_max_age);
 	} else if (!br->topology_change_detected) {
 		printk(", sending tcn bpdu");
 		br_transmit_tcn(br);
-		br_timer_set(&br->tcn_timer, jiffies);
+		mod_timer(&br->tcn_timer, jiffies + br->bridge_hello_time);
 	}
 
 	printk("\n");
@@ -299,7 +321,7 @@ void br_config_bpdu_generation(struct net_bridge *br)
 }
 
 /* called under bridge lock */
-static void br_reply(struct net_bridge_port *p)
+static inline void br_reply(struct net_bridge_port *p)
 {
 	br_transmit_config(p);
 }
@@ -323,6 +345,7 @@ void br_become_designated_port(struct net_bridge_port *p)
 	p->designated_port = p->port_id;
 }
 
+
 /* called under bridge lock */
 static void br_make_blocking(struct net_bridge_port *p)
 {
@@ -332,11 +355,9 @@ static void br_make_blocking(struct net_bridge_port *p)
 		    p->state == BR_STATE_LEARNING)
 			br_topology_change_detection(p->br);
 
-		printk(KERN_INFO "%s: port %i(%s) entering %s state\n",
-		       p->br->dev.name, p->port_no, p->dev->name, "blocking");
-
 		p->state = BR_STATE_BLOCKING;
-		br_timer_clear(&p->forward_delay_timer);
+		br_log_state(p);
+		del_timer(&p->forward_delay_timer);
 	}
 }
 
@@ -345,20 +366,12 @@ static void br_make_forwarding(struct net_bridge_port *p)
 {
 	if (p->state == BR_STATE_BLOCKING) {
 		if (p->br->stp_enabled) {
-			printk(KERN_INFO "%s: port %i(%s) entering %s state\n",
-			       p->br->dev.name, p->port_no, p->dev->name,
-			       "listening");
-
 			p->state = BR_STATE_LISTENING;
 		} else {
-			printk(KERN_INFO "%s: port %i(%s) entering %s state\n",
-			       p->br->dev.name, p->port_no, p->dev->name,
-			       "learning");
-
 			p->state = BR_STATE_LEARNING;
 		}
-		br_timer_set(&p->forward_delay_timer, jiffies);
-	}
+		br_log_state(p);
+		mod_timer(&p->forward_delay_timer, jiffies + p->br->forward_delay);	}
 }
 
 /* called under bridge lock */
@@ -373,7 +386,7 @@ void br_port_state_selection(struct net_bridge *br)
 				p->topology_change_ack = 0;
 				br_make_forwarding(p);
 			} else if (br_is_designated_port(p)) {
-				br_timer_clear(&p->message_age_timer);
+				del_timer(&p->message_age_timer);
 				br_make_forwarding(p);
 			} else {
 				p->config_pending = 0;
@@ -381,11 +394,12 @@ void br_port_state_selection(struct net_bridge *br)
 				br_make_blocking(p);
 			}
 		}
+
 	}
 }
 
 /* called under bridge lock */
-static void br_topology_change_acknowledge(struct net_bridge_port *p)
+static inline void br_topology_change_acknowledge(struct net_bridge_port *p)
 {
 	p->topology_change_ack = 1;
 	br_transmit_config(p);
@@ -396,20 +410,23 @@ void br_received_config_bpdu(struct net_bridge_port *p, struct br_config_bpdu *b
 {
 	struct net_bridge *br;
 	int was_root;
-
+ 
 	br = p->br;
 	was_root = br_is_root_bridge(br);
+
 	if (br_supersedes_port_info(p, bpdu)) {
 		br_record_config_information(p, bpdu);
 		br_configuration_update(br);
 		br_port_state_selection(br);
 
 		if (!br_is_root_bridge(br) && was_root) {
-			br_timer_clear(&br->hello_timer);
+			del_timer(&br->hello_timer);
 			if (br->topology_change_detected) {
-				br_timer_clear(&br->topology_change_timer);
+				del_timer(&br->topology_change_timer);
 				br_transmit_tcn(br);
-				br_timer_set(&br->tcn_timer, jiffies);
+
+				mod_timer(&br->tcn_timer, 
+					  jiffies + br->bridge_hello_time);
 			}
 		}
 
@@ -428,7 +445,7 @@ void br_received_config_bpdu(struct net_bridge_port *p, struct br_config_bpdu *b
 void br_received_tcn_bpdu(struct net_bridge_port *p)
 {
 	if (br_is_designated_port(p)) {
-		printk(KERN_INFO "%s: received tcn bpdu on port %i(%s)\n",
+		pr_info("%s: received tcn bpdu on port %i(%s)\n",
 		       p->br->dev.name, p->port_no, p->dev->name);
 
 		br_topology_change_detection(p->br);

net/bridge/br_stp_if.c

@@ -20,7 +20,7 @@
 #include "br_private.h"
 #include "br_private_stp.h"
 
-__u16 br_make_port_id(struct net_bridge_port *p)
+static inline __u16 br_make_port_id(const struct net_bridge_port *p)
 {
 	return (p->priority << 8) | p->port_no;
 }
@@ -33,33 +33,25 @@ void br_init_port(struct net_bridge_port *p)
 	p->state = BR_STATE_BLOCKING;
 	p->topology_change_ack = 0;
 	p->config_pending = 0;
-	br_timer_clear(&p->message_age_timer);
-	br_timer_clear(&p->forward_delay_timer);
-	br_timer_clear(&p->hold_timer);
+
+	br_stp_port_timer_init(p);
 }
 
 /* called under bridge lock */
 void br_stp_enable_bridge(struct net_bridge *br)
 {
 	struct net_bridge_port *p;
-	struct timer_list *timer = &br->tick;
 
 	spin_lock_bh(&br->lock);
-	init_timer(timer);
-	timer->data = (unsigned long) br;
-	timer->function = br_tick;
-	timer->expires = jiffies + 1;
-	add_timer(timer);
-
-	br_timer_set(&br->hello_timer, jiffies);
+	br->hello_timer.expires = jiffies + br->hello_time;
+	add_timer(&br->hello_timer);
 	br_config_bpdu_generation(br);
 
 	list_for_each_entry(p, &br->port_list, list) {
 		if (p->dev->flags & IFF_UP)
 			br_stp_enable_port(p);
-	}
 
-	br_timer_set(&br->gc_timer, jiffies);
+	}
 	spin_unlock_bh(&br->lock);
 }
 
@@ -68,22 +60,22 @@ void br_stp_disable_bridge(struct net_bridge *br)
 {
 	struct net_bridge_port *p;
 
-	spin_lock_bh(&br->lock);
-	br->topology_change = 0;
-	br->topology_change_detected = 0;
-	br_timer_clear(&br->hello_timer);
-	br_timer_clear(&br->topology_change_timer);
-	br_timer_clear(&br->tcn_timer);
-	br_timer_clear(&br->gc_timer);
-	br_fdb_cleanup(br);
-
+	spin_lock(&br->lock);
 	list_for_each_entry(p, &br->port_list, list) {
 		if (p->state != BR_STATE_DISABLED)
 			br_stp_disable_port(p);
+
 	}
-	spin_unlock_bh(&br->lock);
 
-	del_timer_sync(&br->tick);
+	br->topology_change = 0;
+	br->topology_change_detected = 0;
+	spin_unlock(&br->lock);
+
+	del_timer_sync(&br->hello_timer);
+	del_timer_sync(&br->topology_change_timer);
+	del_timer_sync(&br->tcn_timer);
+	del_timer_sync(&br->gc_timer);
+
 }
 
 /* called under bridge lock */
@@ -108,10 +100,13 @@ void br_stp_disable_port(struct net_bridge_port *p)
 	p->state = BR_STATE_DISABLED;
 	p->topology_change_ack = 0;
 	p->config_pending = 0;
-	br_timer_clear(&p->message_age_timer);
-	br_timer_clear(&p->forward_delay_timer);
-	br_timer_clear(&p->hold_timer);
+
+	del_timer(&p->message_age_timer);
+	del_timer(&p->forward_delay_timer);
+	del_timer(&p->hold_timer);
+
 	br_configuration_update(br);
+
 	br_port_state_selection(br);
 
 	if (br_is_root_bridge(br) && !wasroot)

net/bridge/br_stp_timer.c

@@ -20,51 +20,59 @@
 #include "br_private.h"
 #include "br_private_stp.h"
 
-static void dump_bridge_id(bridge_id *id)
-{
-	printk("%.2x%.2x.%.2x:%.2x:%.2x:%.2x:%.2x:%.2x", id->prio[0],
-	       id->prio[1], id->addr[0], id->addr[1], id->addr[2], id->addr[3],
-	       id->addr[4], id->addr[5]);
-}
-
 /* called under bridge lock */
-static int br_is_designated_for_some_port(struct net_bridge *br)
+static int br_is_designated_for_some_port(const struct net_bridge *br)
 {
 	struct net_bridge_port *p;
 
 	list_for_each_entry(p, &br->port_list, list) {
 		if (p->state != BR_STATE_DISABLED &&
-		    !memcmp(&p->designated_bridge, &br->bridge_id, 8))
+		    !memcmp(&p->designated_bridge, &br->bridge_id, 8)) 
 			return 1;
 	}
 
 	return 0;
 }
 
-/* called under bridge lock */
-static void br_hello_timer_expired(struct net_bridge *br)
+static void br_hello_timer_expired(unsigned long arg)
 {
-	br_config_bpdu_generation(br);
-	br_timer_set(&br->hello_timer, jiffies);
+	struct net_bridge *br = (struct net_bridge *)arg;
+	
+	pr_debug("%s: hello timer expired\n", br->dev.name);
+	spin_lock_bh(&br->lock);
+	if (br->dev.flags & IFF_UP) {
+		br_config_bpdu_generation(br);
+
+		br->hello_timer.expires = jiffies + br->hello_time;
+		add_timer(&br->hello_timer);
+	}
+	spin_unlock_bh(&br->lock);
 }
 
-/* called under bridge lock */
-static void br_message_age_timer_expired(struct net_bridge_port *p)
+static void br_message_age_timer_expired(unsigned long arg)
 {
-	struct net_bridge *br;
+	struct net_bridge_port *p = (struct net_bridge_port *) arg;
+	struct net_bridge *br = p->br;
+	const bridge_id *id = &p->designated_bridge;
 	int was_root;
 
-	br = p->br;
-	printk(KERN_INFO "%s: ", br->dev.name);
-	printk("neighbour ");
-	dump_bridge_id(&p->designated_bridge);
-	printk(" lost on port %i(%s)\n", p->port_no, p->dev->name);
+	if (p->state == BR_STATE_DISABLED)
+		return;
+
+	
+	pr_info("%s: neighbor %.2x%.2x.%.2x:%.2x:%.2x:%.2x:%.2x:%.2x lost on port %d(%s)\n",
+		br->dev.name, 
+		id->prio[0], id->prio[1], 
+		id->addr[0], id->addr[1], id->addr[2], 
+		id->addr[3], id->addr[4], id->addr[5],
+		p->port_no, p->dev->name);
 
 	/*
 	 * According to the spec, the message age timer cannot be
 	 * running when we are the root bridge. So..  this was_root
 	 * check is redundant. I'm leaving it in for now, though.
 	 */
+	spin_lock_bh(&br->lock);
 	was_root = br_is_root_bridge(br);
 
 	br_become_designated_port(p);
@@ -72,107 +80,101 @@ static void br_message_age_timer_expired(struct net_bridge_port *p)
 	br_port_state_selection(br);
 	if (br_is_root_bridge(br) && !was_root)
 		br_become_root_bridge(br);
+	spin_unlock_bh(&br->lock);
 }
 
-/* called under bridge lock */
-static void br_forward_delay_timer_expired(struct net_bridge_port *p)
+static void br_forward_delay_timer_expired(unsigned long arg)
 {
-	if (p->state == BR_STATE_LISTENING) {
-		printk(KERN_INFO "%s: port %i(%s) entering %s state\n",
-		       p->br->dev.name, p->port_no, p->dev->name, "learning");
+	struct net_bridge_port *p = (struct net_bridge_port *) arg;
+	struct net_bridge *br = p->br;
 
+	pr_debug("%s: %d(%s) forward delay timer\n",
+		 br->dev.name, p->port_no, p->dev->name);
+	spin_lock_bh(&br->lock);
+	if (p->state == BR_STATE_LISTENING) {
 		p->state = BR_STATE_LEARNING;
-		br_timer_set(&p->forward_delay_timer, jiffies);
+		p->forward_delay_timer.expires = jiffies + br->forward_delay;
+		add_timer(&p->forward_delay_timer);
 	} else if (p->state == BR_STATE_LEARNING) {
-		printk(KERN_INFO "%s: port %i(%s) entering %s state\n",
-		       p->br->dev.name, p->port_no, p->dev->name, "forwarding");
-
 		p->state = BR_STATE_FORWARDING;
-		if (br_is_designated_for_some_port(p->br))
-			br_topology_change_detection(p->br);
+		if (br_is_designated_for_some_port(br))
+			br_topology_change_detection(br);
 	}
+	br_log_state(p);
+	spin_unlock_bh(&br->lock);
 }
 
-/* called under bridge lock */
-static void br_tcn_timer_expired(struct net_bridge *br)
+static void br_tcn_timer_expired(unsigned long arg)
 {
-	printk(KERN_INFO "%s: retransmitting tcn bpdu\n", br->dev.name);
-	br_transmit_tcn(br);
-	br_timer_set(&br->tcn_timer, jiffies);
+	struct net_bridge *br = (struct net_bridge *) arg;
+
+	pr_debug("%s: tcn timer expired\n", br->dev.name);
+	spin_lock_bh(&br->lock);
+	if (br->dev.flags & IFF_UP) {
+		br_transmit_tcn(br);
+	
+		br->tcn_timer.expires = jiffies + br->bridge_hello_time;
+		add_timer(&br->tcn_timer);
+	}
+	spin_unlock_bh(&br->lock);
 }
 
-/* called under bridge lock */
-static void br_topology_change_timer_expired(struct net_bridge *br)
+static void br_topology_change_timer_expired(unsigned long arg)
 {
+	struct net_bridge *br = (struct net_bridge *) arg;
+
+	pr_debug("%s: topo change timer expired\n", br->dev.name);
+	spin_lock_bh(&br->lock);
 	br->topology_change_detected = 0;
 	br->topology_change = 0;
+	spin_unlock_bh(&br->lock);
 }
 
-/* called under bridge lock */
-static void br_hold_timer_expired(struct net_bridge_port *p)
+static void br_hold_timer_expired(unsigned long arg)
 {
+	struct net_bridge_port *p = (struct net_bridge_port *) arg;
+
+	pr_debug("%s: %d(%s) hold timer expired\n", 
+		 p->br->dev.name,  p->port_no, p->dev->name);
+
+	spin_lock_bh(&p->br->lock);
 	if (p->config_pending)
 		br_transmit_config(p);
+	spin_unlock_bh(&p->br->lock);
 }
 
-/* called under bridge lock */
-static void br_check_port_timers(struct net_bridge_port *p)
+static inline void br_timer_init(struct timer_list *timer,
+			  void (*_function)(unsigned long),
+			  unsigned long _data)
 {
-	if (br_timer_has_expired(&p->message_age_timer, p->br->max_age)) {
-		br_timer_clear(&p->message_age_timer);
-		br_message_age_timer_expired(p);
-	}
-
-	if (br_timer_has_expired(&p->forward_delay_timer, p->br->forward_delay)) {
-		br_timer_clear(&p->forward_delay_timer);
-		br_forward_delay_timer_expired(p);
-	}
-
-	if (br_timer_has_expired(&p->hold_timer, BR_HOLD_TIME)) {
-		br_timer_clear(&p->hold_timer);
-		br_hold_timer_expired(p);
-	}
+	init_timer(timer);
+	timer->function = _function;
+	timer->data = _data;
 }
 
-/* called under bridge lock */
-static void br_check_timers(struct net_bridge *br)
+void br_stp_timer_init(struct net_bridge *br)
 {
-	struct net_bridge_port *p;
-
-	if (br_timer_has_expired(&br->gc_timer, br->gc_interval)) {
-		br_timer_set(&br->gc_timer, jiffies);
-		br_fdb_cleanup(br);
-	}
+	br_timer_init(&br->hello_timer, br_hello_timer_expired,
+		      (unsigned long) br);
 
-	if (br_timer_has_expired(&br->hello_timer, br->hello_time)) {
-		br_timer_clear(&br->hello_timer);
-		br_hello_timer_expired(br);
-	}
+	br_timer_init(&br->tcn_timer, br_tcn_timer_expired, 
+		      (unsigned long) br);
 
-	if (br_timer_has_expired(&br->tcn_timer, br->bridge_hello_time)) {
-		br_timer_clear(&br->tcn_timer);
-		br_tcn_timer_expired(br);
-	}
+	br_timer_init(&br->topology_change_timer,
+		      br_topology_change_timer_expired,
+		      (unsigned long) br);
 
-	if (br_timer_has_expired(&br->topology_change_timer, br->bridge_forward_delay + br->bridge_max_age)) {
-		br_timer_clear(&br->topology_change_timer);
-		br_topology_change_timer_expired(br);
-	}
-
-	list_for_each_entry(p, &br->port_list, list) {
-		if (p->state != BR_STATE_DISABLED)
-			br_check_port_timers(p);
-	}
+	br_timer_init(&br->gc_timer, br_fdb_cleanup, (unsigned long) br);
 }
 
-void br_tick(unsigned long __data)
+void br_stp_port_timer_init(struct net_bridge_port *p)
 {
-	struct net_bridge *br = (struct net_bridge *)__data;
-
-	if (spin_trylock_bh(&br->lock)) {
-		br_check_timers(br);
-		spin_unlock_bh(&br->lock);
-	}
-	br->tick.expires = jiffies + 1;
-	add_timer(&br->tick);
-}
+	br_timer_init(&p->message_age_timer, br_message_age_timer_expired,
+		      (unsigned long) p);
+
+	br_timer_init(&p->forward_delay_timer, br_forward_delay_timer_expired,
+		      (unsigned long) p);
+		      
+	br_timer_init(&p->hold_timer, br_hold_timer_expired,
+		      (unsigned long) p);
+}	

net/core/dev.c

@@ -90,7 +90,6 @@
 #include <linux/etherdevice.h>
 #include <linux/notifier.h>
 #include <linux/skbuff.h>
-#include <linux/brlock.h>
 #include <net/sock.h>
 #include <linux/rtnetlink.h>
 #include <linux/proc_fs.h>
@@ -170,8 +169,9 @@ const char *if_port_text[] = {
  *		86DD	IPv6
  */
 
-static struct packet_type *ptype_base[16];	/* 16 way hashed list */
-static struct packet_type *ptype_all;		/* Taps */
+static spinlock_t ptype_lock = SPIN_LOCK_UNLOCKED;
+static struct list_head ptype_base[16];	/* 16 way hashed list */
+static struct list_head ptype_all;		/* Taps */
 
 #ifdef OFFLINE_SAMPLE
 static void sample_queue(unsigned long dummy);
@@ -239,14 +239,17 @@ int netdev_nit;
  *	Add a protocol handler to the networking stack. The passed &packet_type
  *	is linked into kernel lists and may not be freed until it has been
  *	removed from the kernel lists.
+ *
+ *	This call does not sleep therefore it can not 
+ *	guarantee all CPU's that are in middle of receiving packets
+ *	will see the new packet type (until the next received packet).
  */
 
 void dev_add_pack(struct packet_type *pt)
 {
 	int hash;
 
-	br_write_lock_bh(BR_NETPROTO_LOCK);
-
+	spin_lock_bh(&ptype_lock);
 #ifdef CONFIG_NET_FASTROUTE
 	/* Hack to detect packet socket */
 	if (pt->data && (long)(pt->data) != 1) {
@@ -256,52 +259,76 @@ void dev_add_pack(struct packet_type *pt)
 #endif
 	if (pt->type == htons(ETH_P_ALL)) {
 		netdev_nit++;
-		pt->next  = ptype_all;
-		ptype_all = pt;
+		list_add_rcu(&pt->list, &ptype_all);
 	} else {
 		hash = ntohs(pt->type) & 15;
-		pt->next = ptype_base[hash];
-		ptype_base[hash] = pt;
+		list_add_rcu(&pt->list, &ptype_base[hash]);
 	}
-	br_write_unlock_bh(BR_NETPROTO_LOCK);
+	spin_unlock_bh(&ptype_lock);
 }
 
 extern void linkwatch_run_queue(void);
 
+
+
 /**
- *	dev_remove_pack	 - remove packet handler
+ *	__dev_remove_pack	 - remove packet handler
  *	@pt: packet type declaration
  *
  *	Remove a protocol handler that was previously added to the kernel
  *	protocol handlers by dev_add_pack(). The passed &packet_type is removed
  *	from the kernel lists and can be freed or reused once this function
- *	returns.
+ *	returns. 
+ *
+ *      The packet type might still be in use by receivers
+ *	and must not be freed until after all the CPU's have gone
+ *	through a quiescent state.
  */
-void dev_remove_pack(struct packet_type *pt)
+void __dev_remove_pack(struct packet_type *pt)
 {
-	struct packet_type **pt1;
+	struct list_head *head;
+	struct packet_type *pt1;
 
-	br_write_lock_bh(BR_NETPROTO_LOCK);
+	spin_lock_bh(&ptype_lock);
 
 	if (pt->type == htons(ETH_P_ALL)) {
 		netdev_nit--;
-		pt1 = &ptype_all;
+		head = &ptype_all;
 	} else
-		pt1 = &ptype_base[ntohs(pt->type) & 15];
+		head = &ptype_base[ntohs(pt->type) & 15];
 
-	for (; *pt1; pt1 = &((*pt1)->next)) {
-		if (pt == *pt1) {
-			*pt1 = pt->next;
+	list_for_each_entry(pt1, head, list) {
+		if (pt == pt1) {
 #ifdef CONFIG_NET_FASTROUTE
 			if (pt->data)
 				netdev_fastroute_obstacles--;
 #endif
+			list_del_rcu(&pt->list);
 			goto out;
 		}
 	}
+
 	printk(KERN_WARNING "dev_remove_pack: %p not found.\n", pt);
 out:
-	br_write_unlock_bh(BR_NETPROTO_LOCK);
+	spin_unlock_bh(&ptype_lock);
+}
+/**
+ *	dev_remove_pack	 - remove packet handler
+ *	@pt: packet type declaration
+ *
+ *	Remove a protocol handler that was previously added to the kernel
+ *	protocol handlers by dev_add_pack(). The passed &packet_type is removed
+ *	from the kernel lists and can be freed or reused once this function
+ *	returns.
+ *
+ *	This call sleeps to guarantee that no CPU is looking at the packet
+ *	type after return.
+ */
+void dev_remove_pack(struct packet_type *pt)
+{
+	__dev_remove_pack(pt);
+	
+	synchronize_net();
 }
 
 /******************************************************************************
@@ -943,8 +970,8 @@ void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev)
 	struct packet_type *ptype;
 	do_gettimeofday(&skb->stamp);
 
-	br_read_lock(BR_NETPROTO_LOCK);
-	for (ptype = ptype_all; ptype; ptype = ptype->next) {
+	rcu_read_lock();
+	list_for_each_entry_rcu(ptype, &ptype_all, list) {
 		/* Never send packets back to the socket
 		 * they originated from - MvS (miquels@drinkel.ow.org)
 		 */
@@ -974,7 +1001,7 @@ void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev)
 			ptype->func(skb2, skb->dev, ptype);
 		}
 	}
-	br_read_unlock(BR_NETPROTO_LOCK);
+	rcu_read_unlock();
 }
 
 /* Calculate csum in the case, when packet is misrouted.
@@ -1488,7 +1515,8 @@ int netif_receive_skb(struct sk_buff *skb)
 	skb->h.raw = skb->nh.raw = skb->data;
 
 	pt_prev = NULL;
-	for (ptype = ptype_all; ptype; ptype = ptype->next) {
+	rcu_read_lock();
+	list_for_each_entry_rcu(ptype, &ptype_all, list) {
 		if (!ptype->dev || ptype->dev == skb->dev) {
 			if (pt_prev) {
 				if (!pt_prev->data) {
@@ -1511,17 +1539,15 @@ int netif_receive_skb(struct sk_buff *skb)
 
 #if defined(CONFIG_BRIDGE) || defined(CONFIG_BRIDGE_MODULE)
 	if (skb->dev->br_port) {
-		int ret;
-
 		ret = handle_bridge(skb, pt_prev);
 		if (br_handle_frame_hook(skb) == 0) 
-			return ret;
+			goto out;
 
 		pt_prev = NULL;
 	}
 #endif
 
-	for (ptype = ptype_base[ntohs(type) & 15]; ptype; ptype = ptype->next) {
+	list_for_each_entry_rcu(ptype, &ptype_base[ntohs(type)&15], list) {
 		if (ptype->type == type &&
 		    (!ptype->dev || ptype->dev == skb->dev)) {
 			if (pt_prev) {
@@ -1552,6 +1578,8 @@ int netif_receive_skb(struct sk_buff *skb)
 		ret = NET_RX_DROP;
 	}
 
+ out:
+	rcu_read_unlock();
 	return ret;
 }
 
@@ -1625,7 +1653,8 @@ static void net_rx_action(struct softirq_action *h)
 	unsigned long start_time = jiffies;
 	int budget = netdev_max_backlog;
 
-	br_read_lock(BR_NETPROTO_LOCK);
+	
+	preempt_disable();
 	local_irq_disable();
 
 	while (!list_empty(&queue->poll_list)) {
@@ -1654,7 +1683,7 @@ static void net_rx_action(struct softirq_action *h)
 	}
 out:
 	local_irq_enable();
-	br_read_unlock(BR_NETPROTO_LOCK);
+	preempt_enable();
 	return;
 
 softnet_break:
@@ -1997,9 +2026,9 @@ int netdev_set_master(struct net_device *slave, struct net_device *master)
 		dev_hold(master);
 	}
 
-	br_write_lock_bh(BR_NETPROTO_LOCK);
 	slave->master = master;
-	br_write_unlock_bh(BR_NETPROTO_LOCK);
+	
+	synchronize_net();
 
 	if (old)
 		dev_put(old);
@@ -2663,8 +2692,8 @@ int netdev_finish_unregister(struct net_device *dev)
 /* Synchronize with packet receive processing. */
 void synchronize_net(void) 
 {
-	br_write_lock_bh(BR_NETPROTO_LOCK);
-	br_write_unlock_bh(BR_NETPROTO_LOCK);
+	might_sleep();
+	synchronize_kernel();
 }
 
 /**
@@ -2848,6 +2877,10 @@ static int __init net_dev_init(void)
 
 	subsystem_register(&net_subsys);
 
+	INIT_LIST_HEAD(&ptype_all);
+	for (i = 0; i < 16; i++) 
+		INIT_LIST_HEAD(&ptype_base[i]);
+
 #ifdef CONFIG_NET_DIVERT
 	dv_init();
 #endif /* CONFIG_NET_DIVERT */

net/core/wireless.c

@@ -53,6 +53,7 @@
 /***************************** INCLUDES *****************************/
 
 #include <linux/config.h>		/* Not needed ??? */
+#include <linux/module.h>
 #include <linux/types.h>		/* off_t */
 #include <linux/netdevice.h>		/* struct ifreq, dev_get_by_name() */
 #include <linux/proc_fs.h>

net/ipv4/af_inet.c

@@ -1013,7 +1013,7 @@ void inet_register_protosw(struct inet_protosw *p)
 out:
 	spin_unlock_bh(&inetsw_lock);
 
-	synchronize_kernel();
+	synchronize_net();
 
 	return;
 
@@ -1040,7 +1040,7 @@ void inet_unregister_protosw(struct inet_protosw *p)
 		list_del_rcu(&p->list);
 		spin_unlock_bh(&inetsw_lock);
 
-		synchronize_kernel();
+		synchronize_net();
 	}
 }
 

net/ipv4/ip_output.c

@@ -685,16 +685,6 @@ skb_can_coalesce(struct sk_buff *skb, int i, struct page *page, int off)
 	return 0;
 }
 
-static void
-skb_fill_page_desc(struct sk_buff *skb, int i, struct page *page, int off, int size)
-{
-	skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-	frag->page = page;
-	frag->page_offset = off;
-	frag->size = size;
-	skb_shinfo(skb)->nr_frags = i+1;
-}
-
 static inline unsigned int
 csum_page(struct page *page, int offset, int copy)
 {

net/ipv4/ipcomp.c

@@ -269,6 +269,67 @@ static void ipcomp4_err(struct sk_buff *skb, u32 info)
 	xfrm_state_put(x);
 }
 
+/* We always hold one tunnel user reference to indicate a tunnel */ 
+static struct xfrm_state *ipcomp_tunnel_create(struct xfrm_state *x)
+{
+	struct xfrm_state *t;
+	
+	t = xfrm_state_alloc();
+	if (t == NULL)
+		goto out;
+
+	t->id.proto = IPPROTO_IPIP;
+	t->id.spi = x->props.saddr.a4;
+	t->id.daddr.a4 = x->id.daddr.a4;
+	memcpy(&t->sel, &x->sel, sizeof(t->sel));
+	t->props.family = AF_INET;
+	t->props.mode = 1;
+	t->props.saddr.a4 = x->props.saddr.a4;
+	
+	t->type = xfrm_get_type(IPPROTO_IPIP, t->props.family);
+	if (t->type == NULL)
+		goto error;
+		
+	if (t->type->init_state(t, NULL))
+		goto error;
+
+	t->km.state = XFRM_STATE_VALID;
+	atomic_set(&t->tunnel_users, 1);
+out:
+	return t;
+
+error:
+	xfrm_state_put(t);
+	t = NULL;
+	goto out;
+}
+
+/*
+ * Must be protected by xfrm_cfg_sem.  State and tunnel user references are
+ * always incremented on success.
+ */
+static int ipcomp_tunnel_attach(struct xfrm_state *x)
+{
+	int err = 0;
+	struct xfrm_state *t;
+
+	t = xfrm_state_lookup((xfrm_address_t *)&x->id.daddr.a4,
+	                      x->props.saddr.a4, IPPROTO_IPIP, AF_INET);
+	if (!t) {
+		t = ipcomp_tunnel_create(x);
+		if (!t) {
+			err = -EINVAL;
+			goto out;
+		}
+		xfrm_state_insert(t);
+		xfrm_state_hold(t);
+	}
+	x->tunnel = t;
+	atomic_inc(&t->tunnel_users);
+out:
+	return err;
+}
+
 static void ipcomp_free_data(struct ipcomp_data *ipcd)
 {
 	if (ipcd->tfm)
@@ -308,6 +369,12 @@ static int ipcomp_init_state(struct xfrm_state *x, void *args)
 	if (!ipcd->tfm)
 		goto error;
 
+	if (x->props.mode) {
+		err = ipcomp_tunnel_attach(x);
+		if (err)
+			goto error;
+	}
+
 	calg_desc = xfrm_calg_get_byname(x->calg->alg_name);
 	BUG_ON(!calg_desc);
 	ipcd->threshold = calg_desc->uinfo.comp.threshold;

net/ipv4/netfilter/ip_conntrack_core.c

@@ -1188,12 +1188,9 @@ ip_ct_gather_frags(struct sk_buff *skb)
 	local_bh_enable();
 
 	if (!skb) {
-		if (sk) sock_put(sk);
+		if (sk)
+			sock_put(sk);
 		return skb;
-	} else if (skb_is_nonlinear(skb) && skb_linearize(skb, GFP_ATOMIC) != 0) {
-		kfree_skb(skb);
-		if (sk) sock_put(sk);
-		return NULL;
 	}
 
 	if (sk) {

net/ipv4/netfilter/ip_fw_compat.c

@@ -15,34 +15,10 @@ struct notifier_block;
 #include <linux/netfilter_ipv4/compat_firewall.h>
 #include <linux/netfilter_ipv4/ip_conntrack.h>
 #include <linux/netfilter_ipv4/ip_conntrack_core.h>
+#include "ip_fw_compat.h"
 
 static struct firewall_ops *fwops;
 
-/* From ip_fw_compat_redir.c */
-extern unsigned int
-do_redirect(struct sk_buff *skb,
-	    const struct net_device *dev,
-	    u_int16_t redirpt);
-
-extern void
-check_for_redirect(struct sk_buff *skb);
-
-extern void
-check_for_unredirect(struct sk_buff *skb);
-
-/* From ip_fw_compat_masq.c */
-extern unsigned int
-do_masquerade(struct sk_buff **pskb, const struct net_device *dev);
-
-extern unsigned int
-check_for_masq_error(struct sk_buff **pskb);
-
-extern unsigned int
-check_for_demasq(struct sk_buff **pskb);
-
-extern int __init masq_init(void);
-extern void masq_cleanup(void);
-
 /* They call these; we do what they want. */
 int register_firewall(int pf, struct firewall_ops *fw)
 {

net/ipv4/netfilter/ip_fw_compat.h

+#ifndef _LINUX_IP_FW_COMPAT_H
+#define _LINUX_IP_FW_COMPAT_H
+
+/* From ip_fw_compat_redir.c */
+extern unsigned int
+do_redirect(struct sk_buff *skb,
+	    const struct net_device *dev,
+	    u_int16_t redirpt);
+
+extern void
+check_for_redirect(struct sk_buff *skb);
+
+extern void
+check_for_unredirect(struct sk_buff *skb);
+
+/* From ip_fw_compat_masq.c */
+extern unsigned int
+do_masquerade(struct sk_buff **pskb, const struct net_device *dev);
+
+extern void check_for_masq_error(struct sk_buff **pskb);
+
+extern unsigned int
+check_for_demasq(struct sk_buff **pskb);
+
+extern int __init masq_init(void);
+extern void masq_cleanup(void);
+
+#endif /* _LINUX_IP_FW_COMPAT_H */

net/ipv4/netfilter/ip_fw_compat_masq.c

@@ -25,6 +25,7 @@
 #include <linux/netfilter_ipv4/ip_nat.h>
 #include <linux/netfilter_ipv4/ip_nat_core.h>
 #include <linux/netfilter_ipv4/listhelp.h>
+#include "ip_fw_compat.h"
 
 #if 0
 #define DEBUGP printk

net/ipv4/netfilter/ip_fw_compat_redir.c

@@ -28,6 +28,7 @@ static DECLARE_LOCK(redir_lock);
 #define ASSERT_WRITE_LOCK(x) MUST_BE_LOCKED(&redir_lock)
 
 #include <linux/netfilter_ipv4/listhelp.h>
+#include "ip_fw_compat.h"
 
 #if 0
 #define DEBUGP printk

net/ipv4/xfrm4_tunnel.c

@@ -163,36 +163,32 @@ static int ipip_rcv(struct sk_buff *skb)
 			      skb->nh.iph->saddr,
 			      IPPROTO_IPIP, AF_INET);
 
-	if (x) {
-		spin_lock(&x->lock);
+	if (!x)
+		goto drop;
 
-		if (unlikely(x->km.state != XFRM_STATE_VALID))
-			goto drop_unlock;
-	}
+	spin_lock(&x->lock);
+
+	if (unlikely(x->km.state != XFRM_STATE_VALID))
+		goto drop_unlock;
 
 	err = ipip_xfrm_rcv(x, NULL, skb);
 	if (err)
 		goto drop_unlock;
 
-	if (x) {
-		x->curlft.bytes += skb->len;
-		x->curlft.packets++;
-
-		spin_unlock(&x->lock);
-
-		xfrm_state_put(x);
-	}
-
-	return 0;
+	x->curlft.bytes += skb->len;
+	x->curlft.packets++;
+	spin_unlock(&x->lock);
+	xfrm_state_put(x);
+out:	
+	return err;
 
 drop_unlock:
-	if (x) {
-		spin_unlock(&x->lock);
-		xfrm_state_put(x);
-	}
+	spin_unlock(&x->lock);
+	xfrm_state_put(x);
+drop:
+	err = NET_RX_DROP;
 	kfree_skb(skb);
-out:
-	return 0;
+	goto out;
 }
 
 static void ipip_err(struct sk_buff *skb, u32 info)

net/ipv6/af_inet6.c

@@ -637,7 +637,7 @@ inet6_unregister_protosw(struct inet_protosw *p)
 		list_del_rcu(&p->list);
 		spin_unlock_bh(&inetsw6_lock);
 
-		synchronize_kernel();
+		synchronize_net();
 	}
 }
 

net/ipv6/icmp.c

@@ -28,6 +28,7 @@
  *	YOSHIFUJI Hideaki @USAGI:	added sysctl for icmp rate limit.
  *	Randy Dunlap and
  *	YOSHIFUJI Hideaki @USAGI:	Per-interface statistics support
+ *	Kazunori MIYAZAWA @USAGI:       change output process to use ip6_append_data
  */
 
 #include <linux/module.h>
@@ -104,42 +105,6 @@ static __inline__ void icmpv6_xmit_unlock(void)
 	spin_unlock_bh(&icmpv6_socket->sk->lock.slock);
 }
 
-
-
-/*
- *	getfrag callback
- */
-
-static int icmpv6_getfrag(const void *data, struct in6_addr *saddr, 
-			   char *buff, unsigned int offset, unsigned int len)
-{
-	struct icmpv6_msg *msg = (struct icmpv6_msg *) data;
-	struct icmp6hdr *icmph;
-	__u32 csum;
-
-	if (offset) {
-		csum = skb_copy_and_csum_bits(msg->skb, msg->offset +
-					      (offset - sizeof(struct icmp6hdr)),
-					      buff, len, msg->csum);
-		msg->csum = csum;
-		return 0;
-	}
-
-	csum = csum_partial_copy_nocheck((void *) &msg->icmph, buff,
-					 sizeof(struct icmp6hdr), msg->csum);
-
-	csum = skb_copy_and_csum_bits(msg->skb, msg->offset,
-				      buff + sizeof(struct icmp6hdr),
-				      len - sizeof(struct icmp6hdr), csum);
-
-	icmph = (struct icmp6hdr *) buff;
-
-	icmph->icmp6_cksum = csum_ipv6_magic(saddr, msg->daddr, msg->len,
-					     IPPROTO_ICMPV6, csum);
-	return 0; 
-}
-
-
 /* 
  * Slightly more convenient version of icmpv6_send.
  */
@@ -242,22 +207,74 @@ static __inline__ int opt_unrec(struct sk_buff *skb, __u32 offset)
 	return (optval&0xC0) == 0x80;
 }
 
+int icmpv6_push_pending_frames(struct sock *sk, struct flowi *fl, struct icmp6hdr *thdr, int len)
+{
+	struct sk_buff *skb;
+	struct icmp6hdr *icmp6h;
+	int err = 0;
+
+	if ((skb = skb_peek(&sk->write_queue)) == NULL)
+		goto out;
+
+	icmp6h = (struct icmp6hdr*) skb->h.raw;
+	memcpy(icmp6h, thdr, sizeof(struct icmp6hdr));
+	icmp6h->icmp6_cksum = 0;
+
+	if (skb_queue_len(&sk->write_queue) == 1) {
+		skb->csum = csum_partial((char *)icmp6h,
+					sizeof(struct icmp6hdr), skb->csum);
+		icmp6h->icmp6_cksum = csum_ipv6_magic(fl->fl6_src,
+					     fl->fl6_dst,
+					     len, fl->proto, skb->csum);
+	} else {
+		u32 tmp_csum = 0;
+
+		skb_queue_walk(&sk->write_queue, skb) {
+			tmp_csum = csum_add(tmp_csum, skb->csum);
+		}
+
+		tmp_csum = csum_partial((char *)icmp6h,
+					sizeof(struct icmp6hdr), tmp_csum);
+		tmp_csum = csum_ipv6_magic(fl->fl6_src,
+					   fl->fl6_dst,
+					   len, fl->proto, tmp_csum);
+		icmp6h->icmp6_cksum = tmp_csum;
+	}
+	if (icmp6h->icmp6_cksum == 0)
+		icmp6h->icmp6_cksum = -1;
+	ip6_push_pending_frames(sk);
+out:
+	return err;
+}
+
+static int icmpv6_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb)
+{
+	struct sk_buff *org_skb = (struct sk_buff *)from;
+	__u32 csum = 0;
+	csum = skb_copy_and_csum_bits(org_skb, offset, to, len, csum);
+	skb->csum = csum_block_add(skb->csum, csum, odd);
+	return 0;
+}
+
 /*
  *	Send an ICMP message in response to a packet in error
  */
-
 void icmpv6_send(struct sk_buff *skb, int type, int code, __u32 info, 
 		 struct net_device *dev)
 {
 	struct inet6_dev *idev;
 	struct ipv6hdr *hdr = skb->nh.ipv6h;
 	struct sock *sk = icmpv6_socket->sk;
-	struct in6_addr *saddr = NULL;
-	int iif = 0;
-	struct icmpv6_msg msg;
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct in6_addr *saddr = NULL, *tmp_saddr = NULL;
+	struct dst_entry *dst;
+	struct icmp6hdr tmp_hdr;
 	struct flowi fl;
+	int iif = 0;
 	int addr_type = 0;
-	int len;
+	int len, plen;
+	int hlimit = -1;
+	int err = 0;
 
 	if ((u8*)hdr < skb->head || (u8*)(hdr+1) > skb->tail)
 		return;
@@ -328,36 +345,48 @@ void icmpv6_send(struct sk_buff *skb, int type, int code, __u32 info,
 	if (!icmpv6_xrlim_allow(sk, type, &fl))
 		goto out;
 
-	/*
-	 *	ok. kick it. checksum will be provided by the 
-	 *	getfrag_t callback.
-	 */
+	tmp_hdr.icmp6_type = type;
+	tmp_hdr.icmp6_code = code;
+	tmp_hdr.icmp6_cksum = 0;
+	tmp_hdr.icmp6_pointer = htonl(info);
 
-	msg.icmph.icmp6_type = type;
-	msg.icmph.icmp6_code = code;
-	msg.icmph.icmp6_cksum = 0;
-	msg.icmph.icmp6_pointer = htonl(info);
+	if (!fl.oif && ipv6_addr_is_multicast(fl.fl6_dst))
+		fl.oif = np->mcast_oif;
 
-	msg.skb = skb;
-	msg.offset = skb->nh.raw - skb->data;
-	msg.csum = 0;
-	msg.daddr = &hdr->saddr;
+	err = ip6_dst_lookup(sk, &dst, &fl, &tmp_saddr);
+	if (err) goto out;
 
-	len = skb->len - msg.offset + sizeof(struct icmp6hdr);
-	len = min_t(unsigned int, len, IPV6_MIN_MTU - sizeof(struct ipv6hdr));
+	if (hlimit < 0) {
+		if (ipv6_addr_is_multicast(fl.fl6_dst))
+			hlimit = np->mcast_hops;
+		else
+			hlimit = np->hop_limit;
+		if (hlimit < 0)
+			hlimit = ((struct rt6_info*)dst)->rt6i_hoplimit;
+	}
 
+	plen = skb->nh.raw - skb->data;
+	__skb_pull(skb, plen);
+	len = skb->len;
+	len = min_t(unsigned int, len, IPV6_MIN_MTU - sizeof(struct ipv6hdr) -sizeof(struct icmp6hdr));
 	if (len < 0) {
 		if (net_ratelimit())
 			printk(KERN_DEBUG "icmp: len problem\n");
+		__skb_push(skb, plen);
 		goto out;
 	}
 
-	msg.len = len;
-
 	idev = in6_dev_get(skb->dev);
-	
-	ip6_build_xmit(sk, icmpv6_getfrag, &msg, &fl, len, NULL, -1,
-		       MSG_DONTWAIT);
+
+	err = ip6_append_data(sk, icmpv6_getfrag, skb, len + sizeof(struct icmp6hdr), sizeof(struct icmp6hdr),
+				hlimit, NULL, &fl, (struct rt6_info*)dst, MSG_DONTWAIT);
+	if (err) {
+		ip6_flush_pending_frames(sk);
+		goto out;
+	}
+	err = icmpv6_push_pending_frames(sk, &fl, &tmp_hdr, len + sizeof(struct icmp6hdr));
+	__skb_push(skb, plen);
+
 	if (type >= ICMPV6_DEST_UNREACH && type <= ICMPV6_PARAMPROB)
 		ICMP6_INC_STATS_OFFSET_BH(idev, Icmp6OutDestUnreachs, type - ICMPV6_DEST_UNREACH);
 	ICMP6_INC_STATS_BH(idev, Icmp6OutMsgs);
@@ -365,6 +394,7 @@ void icmpv6_send(struct sk_buff *skb, int type, int code, __u32 info,
 	if (likely(idev != NULL))
 		in6_dev_put(idev);
 out:
+	if (tmp_saddr) kfree(tmp_saddr);
 	icmpv6_xmit_unlock();
 }
 
@@ -372,10 +402,14 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
 {
 	struct sock *sk = icmpv6_socket->sk;
 	struct inet6_dev *idev;
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct in6_addr *saddr = NULL, *tmp_saddr = NULL;
 	struct icmp6hdr *icmph = (struct icmp6hdr *) skb->h.raw;
-	struct in6_addr *saddr;
-	struct icmpv6_msg msg;
+	struct icmp6hdr tmp_hdr;
 	struct flowi fl;
+	struct dst_entry *dst;
+	int err = 0;
+	int hlimit = -1;
 
 	saddr = &skb->nh.ipv6h->daddr;
 
@@ -383,39 +417,55 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
 	    ipv6_chk_acast_addr(0, saddr)) 
 		saddr = NULL;
 
-	msg.icmph.icmp6_type = ICMPV6_ECHO_REPLY;
-	msg.icmph.icmp6_code = 0;
-	msg.icmph.icmp6_cksum = 0;
-	msg.icmph.icmp6_identifier = icmph->icmp6_identifier;
-	msg.icmph.icmp6_sequence = icmph->icmp6_sequence;
-
-	msg.skb = skb;
-	msg.offset = 0;
-	msg.csum = 0;
-	msg.len = skb->len + sizeof(struct icmp6hdr);
-	msg.daddr =  &skb->nh.ipv6h->saddr;
+	memcpy(&tmp_hdr, icmph, sizeof(tmp_hdr));
+	tmp_hdr.icmp6_type = ICMPV6_ECHO_REPLY;
 
 	fl.proto = IPPROTO_ICMPV6;
-	fl.fl6_dst = msg.daddr;
+	fl.fl6_dst = &skb->nh.ipv6h->saddr;
 	fl.fl6_src = saddr;
 	fl.oif = skb->dev->ifindex;
 	fl.fl6_flowlabel = 0;
 	fl.fl_icmp_type = ICMPV6_ECHO_REPLY;
 	fl.fl_icmp_code = 0;
 
+	icmpv6_xmit_lock();
+
+	if (!fl.oif && ipv6_addr_is_multicast(fl.nl_u.ip6_u.daddr))
+		fl.oif = np->mcast_oif;
+
+	err = ip6_dst_lookup(sk, &dst, &fl, &tmp_saddr);
+
+	if (err) goto out;
+
+	if (hlimit < 0) {
+		if (ipv6_addr_is_multicast(fl.fl6_dst))
+			hlimit = np->mcast_hops;
+		else
+			hlimit = np->hop_limit;
+		if (hlimit < 0)
+			hlimit = ((struct rt6_info*)dst)->rt6i_hoplimit;
+	}
+
 	idev = in6_dev_get(skb->dev);
 
-	icmpv6_xmit_lock();
+	err = ip6_append_data(sk, icmpv6_getfrag, skb, skb->len + sizeof(struct icmp6hdr),
+				sizeof(struct icmp6hdr), hlimit, NULL, &fl,
+				(struct rt6_info*)dst, MSG_DONTWAIT);
 
-	ip6_build_xmit(sk, icmpv6_getfrag, &msg, &fl, msg.len, NULL, -1,
-		       MSG_DONTWAIT);
-	ICMP6_INC_STATS_BH(idev, Icmp6OutEchoReplies);
-	ICMP6_INC_STATS_BH(idev, Icmp6OutMsgs);
+	if (err) {
+		ip6_flush_pending_frames(sk);
+		goto out;
+	}
+	err = icmpv6_push_pending_frames(sk, &fl, &tmp_hdr, skb->len + sizeof(struct icmp6hdr));
 
-	icmpv6_xmit_unlock();
+        ICMP6_INC_STATS_BH(idev, Icmp6OutEchoReplies);
+        ICMP6_INC_STATS_BH(idev, Icmp6OutMsgs);
 
 	if (likely(idev != NULL))
 		in6_dev_put(idev);
+out: 
+	if (tmp_saddr) kfree(tmp_saddr);
+	icmpv6_xmit_unlock();
 }
 
 static void icmpv6_notify(struct sk_buff *skb, int type, int code, u32 info)

net/ipv6/ip6_output.c

@@ -23,6 +23,9 @@
  *
  *      H. von Brand    :       Added missing #include <linux/string.h>
  *	Imran Patel	: 	frag id should be in NBO
+ *      Kazunori MIYAZAWA @USAGI
+ *			:       add ip6_append_data and related functions
+ *				for datagram xmit
  */
 
 #include <linux/config.h>
@@ -52,6 +55,8 @@
 #include <net/icmp.h>
 #include <net/xfrm.h>
 
+static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff*));
+
 static __inline__ void ipv6_select_ident(struct sk_buff *skb, struct frag_hdr *fhdr)
 {
 	static u32 ipv6_fragmentation_id = 1;
@@ -98,7 +103,7 @@ static int ip6_dev_loopback_xmit(struct sk_buff *newskb)
 }
 
 
-int ip6_output(struct sk_buff *skb)
+int ip6_output2(struct sk_buff *skb)
 {
 	struct dst_entry *dst = skb->dst;
 	struct net_device *dev = dst->dev;
@@ -134,6 +139,13 @@ int ip6_output(struct sk_buff *skb)
 	return NF_HOOK(PF_INET6, NF_IP6_POST_ROUTING, skb,NULL, skb->dev,ip6_output_finish);
 }
 
+int ip6_output(struct sk_buff *skb)
+{
+	if ((skb->len > skb->dst->dev->mtu || skb_shinfo(skb)->frag_list))
+		return ip6_fragment(skb, ip6_output2);
+	else
+		return ip6_output2(skb);
+}
 
 #ifdef CONFIG_NETFILTER
 int ip6_route_me_harder(struct sk_buff *skb)
@@ -847,3 +859,658 @@ int ip6_forward(struct sk_buff *skb)
 	kfree_skb(skb);
 	return -EINVAL;
 }
+
+static void ip6_copy_metadata(struct sk_buff *to, struct sk_buff *from)
+{
+	to->pkt_type = from->pkt_type;
+	to->priority = from->priority;
+	to->protocol = from->protocol;
+	to->security = from->security;
+	to->dst = dst_clone(from->dst);
+	to->dev = from->dev;
+
+#ifdef CONFIG_NET_SCHED
+	to->tc_index = from->tc_index;
+#endif
+#ifdef CONFIG_NETFILTER
+	to->nfmark = from->nfmark;
+	/* Connection association is same as pre-frag packet */
+	to->nfct = from->nfct;
+	nf_conntrack_get(to->nfct);
+#if defined(CONFIG_BRIDGE) || defined(CONFIG_BRIDGE_MODULE)
+	to->nf_bridge = from->nf_bridge;
+	nf_bridge_get(to->nf_bridge);
+#endif
+#ifdef CONFIG_NETFILTER_DEBUG
+	to->nf_debug = from->nf_debug;
+#endif
+#endif
+}
+
+static int ip6_found_nexthdr(struct sk_buff *skb, u8 **nexthdr)
+{
+	u16 offset = sizeof(struct ipv6hdr);
+	struct ipv6_opt_hdr *exthdr = (struct ipv6_opt_hdr*)(skb->nh.ipv6h + 1);
+	unsigned int packet_len = skb->tail - skb->nh.raw;
+	int found_rhdr = 0;
+	*nexthdr = &skb->nh.ipv6h->nexthdr;
+
+	while (offset + 1 <= packet_len) {
+
+		switch (**nexthdr) {
+
+		case NEXTHDR_HOP:
+		case NEXTHDR_ROUTING:
+		case NEXTHDR_DEST:
+			if (**nexthdr == NEXTHDR_ROUTING) found_rhdr = 1;
+			if (**nexthdr == NEXTHDR_DEST && found_rhdr) return offset;
+			offset += ipv6_optlen(exthdr);
+			*nexthdr = &exthdr->nexthdr;
+			exthdr = (struct ipv6_opt_hdr*)(skb->nh.raw + offset);
+			break;
+		default :
+			return offset;
+		}
+	}
+
+	return offset;
+}
+
+static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff*))
+{
+	struct net_device *dev;
+	struct rt6_info *rt = (struct rt6_info*)skb->dst;
+	struct sk_buff *frag;
+	struct ipv6hdr *tmp_hdr;
+	struct frag_hdr *fh;
+	unsigned int mtu, hlen, left, len;
+	u32 frag_id = 0;
+	int ptr, offset = 0, err=0;
+	u8 *prevhdr, nexthdr = 0;
+
+	dev = rt->u.dst.dev;
+	hlen = ip6_found_nexthdr(skb, &prevhdr);
+	nexthdr = *prevhdr;
+
+	mtu = dst_pmtu(&rt->u.dst) - hlen - sizeof(struct frag_hdr);
+
+	if (skb_shinfo(skb)->frag_list) {
+		int first_len = 0;
+
+		if (first_len - hlen > mtu ||
+		    ((first_len - hlen) & 7) ||
+		    skb_cloned(skb))
+			goto slow_path;
+
+		for (frag = skb_shinfo(skb)->frag_list; frag; frag = frag->next) {
+			/* Correct geometry. */
+			if (frag->len > mtu ||
+			    ((frag->len & 7) && frag->next) ||
+			    skb_headroom(frag) < hlen)
+			    goto slow_path;
+
+			/* Correct socket ownership. */
+			if (frag->sk == NULL)
+				goto slow_path;
+
+			/* Partially cloned skb? */
+			if (skb_shared(frag))
+				goto slow_path;
+		}
+
+		err = 0;
+		offset = 0;
+		frag = skb_shinfo(skb)->frag_list;
+		skb_shinfo(skb)->frag_list = 0;
+		/* BUILD HEADER */
+
+		tmp_hdr = kmalloc(hlen, GFP_ATOMIC);
+		if (!tmp_hdr) {
+			IP6_INC_STATS(Ip6FragFails);
+			return -ENOMEM;
+		}
+
+		*prevhdr = NEXTHDR_FRAGMENT;
+		memcpy(tmp_hdr, skb->nh.raw, hlen);
+		__skb_pull(skb, hlen);
+		fh = (struct frag_hdr*)__skb_push(skb, sizeof(struct frag_hdr));
+		skb->nh.raw = __skb_push(skb, hlen);
+		memcpy(skb->nh.raw, tmp_hdr, hlen);
+
+		ipv6_select_ident(skb, fh);
+		fh->nexthdr = nexthdr;
+		fh->reserved = 0;
+		fh->frag_off = htons(0x0001);
+		frag_id = fh->identification;
+
+		first_len = skb_pagelen(skb);
+		skb->data_len = first_len - skb_headlen(skb);
+		skb->len = first_len;
+		skb->nh.ipv6h->payload_len = htons(first_len - sizeof(struct ipv6hdr));
+ 
+
+		for (;;) {
+			/* Prepare header of the next frame,
+			 * before previous one went down. */
+			if (frag) {
+				frag->h.raw = frag->data;
+				fh = (struct frag_hdr*)__skb_push(frag, sizeof(struct frag_hdr));
+				frag->nh.raw = __skb_push(frag, hlen);
+				memcpy(frag->nh.raw, tmp_hdr, hlen);
+				offset += skb->len - hlen - sizeof(struct frag_hdr);
+				fh->nexthdr = nexthdr;
+				fh->reserved = 0;
+				if (frag->next != NULL)
+					offset |= 0x0001;
+				fh->frag_off = htons(offset);
+				fh->identification = frag_id;
+				frag->nh.ipv6h->payload_len = htons(frag->len - sizeof(struct ipv6hdr));
+				ip6_copy_metadata(frag, skb);
+			}
+			err = output(skb);
+
+			if (err || !frag)
+				break;
+
+			skb = frag;
+			frag = skb->next;
+			skb->next = NULL;
+		}
+
+		if (tmp_hdr)
+			kfree(tmp_hdr);
+
+		if (err == 0) {
+			IP6_INC_STATS(Ip6FragOKs);
+			return 0;
+		}
+
+		while (frag) {
+			skb = frag->next;
+			kfree_skb(frag);
+			frag = skb;
+		}
+
+		IP6_INC_STATS(Ip6FragFails);
+		return err;
+	}
+
+slow_path:
+	left = skb->len - hlen;		/* Space per frame */
+	ptr = hlen;			/* Where to start from */
+
+	/*
+	 *	Fragment the datagram.
+	 */
+
+	*prevhdr = NEXTHDR_FRAGMENT;
+
+	/*
+	 *	Keep copying data until we run out.
+	 */
+	while(left > 0)	{
+		len = left;
+		/* IF: it doesn't fit, use 'mtu' - the data space left */
+		if (len > mtu)
+			len = mtu;
+		/* IF: we are not sending upto and including the packet end
+		   then align the next start on an eight byte boundary */
+		if (len < left)	{
+			len &= ~7;
+		}
+		/*
+		 *	Allocate buffer.
+		 */
+
+		if ((frag = alloc_skb(len+hlen+sizeof(struct frag_hdr)+LL_RESERVED_SPACE(rt->u.dst.dev), GFP_ATOMIC)) == NULL) {
+			NETDEBUG(printk(KERN_INFO "IPv6: frag: no memory for new fragment!\n"));
+			err = -ENOMEM;
+			goto fail;
+		}
+
+		/*
+		 *	Set up data on packet
+		 */
+
+		ip6_copy_metadata(frag, skb);
+		skb_reserve(frag, LL_RESERVED_SPACE(rt->u.dst.dev));
+		skb_put(frag, len + hlen + sizeof(struct frag_hdr));
+		frag->nh.raw = frag->data;
+		fh = (struct frag_hdr*)(frag->data + hlen);
+		frag->h.raw = frag->data + hlen + sizeof(struct frag_hdr);
+
+		/*
+		 *	Charge the memory for the fragment to any owner
+		 *	it might possess
+		 */
+		if (skb->sk)
+			skb_set_owner_w(frag, skb->sk);
+
+		/*
+		 *	Copy the packet header into the new buffer.
+		 */
+		memcpy(frag->nh.raw, skb->data, hlen);
+
+		/*
+		 *	Build fragment header.
+		 */
+		fh->nexthdr = nexthdr;
+		fh->reserved = 0;
+		if (frag_id) {
+			ipv6_select_ident(skb, fh);
+			frag_id = fh->identification;
+		} else
+			fh->identification = frag_id;
+
+		/*
+		 *	Copy a block of the IP datagram.
+		 */
+		if (skb_copy_bits(skb, ptr, frag->h.raw, len))
+			BUG();
+		left -= len;
+
+		fh->frag_off = htons( left > 0 ?  (offset | 0x0001) : offset);
+		frag->nh.ipv6h->payload_len = htons(frag->len - sizeof(struct ipv6hdr));
+
+		ptr += len;
+		offset += len;
+
+		/*
+		 *	Put this fragment into the sending queue.
+		 */
+
+		IP6_INC_STATS(Ip6FragCreates);
+
+		err = output(frag);
+		if (err)
+			goto fail;
+	}
+	kfree_skb(skb);
+	IP6_INC_STATS(Ip6FragOKs);
+	return err;
+
+fail:
+	kfree_skb(skb); 
+	IP6_INC_STATS(Ip6FragFails);
+	return err;
+}
+
+int ip6_dst_lookup(struct sock *sk, struct dst_entry **dst, struct flowi *fl, struct in6_addr **saddr)
+{
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	int err = 0;
+
+	*dst = __sk_dst_check(sk, np->dst_cookie);
+	if (*dst) {
+		struct rt6_info *rt = (struct rt6_info*)*dst;
+
+			/* Yes, checking route validity in not connected
+			   case is not very simple. Take into account,
+			   that we do not support routing by source, TOS,
+			   and MSG_DONTROUTE 		--ANK (980726)
+
+			   1. If route was host route, check that
+			      cached destination is current.
+			      If it is network route, we still may
+			      check its validity using saved pointer
+			      to the last used address: daddr_cache.
+			      We do not want to save whole address now,
+			      (because main consumer of this service
+			       is tcp, which has not this problem),
+			      so that the last trick works only on connected
+			      sockets.
+			   2. oif also should be the same.
+			 */
+
+		if (((rt->rt6i_dst.plen != 128 ||
+		      ipv6_addr_cmp(fl->fl6_dst, &rt->rt6i_dst.addr))
+		     && (np->daddr_cache == NULL ||
+			 ipv6_addr_cmp(fl->fl6_dst, np->daddr_cache)))
+		    || (fl->oif && fl->oif != (*dst)->dev->ifindex)) {
+			*dst = NULL;
+		} else
+			dst_hold(*dst);
+	}
+
+	if (*dst == NULL)
+		*dst = ip6_route_output(sk, fl);
+
+	if ((*dst)->error) {
+		IP6_INC_STATS(Ip6OutNoRoutes);
+		dst_release(*dst);
+		return -ENETUNREACH;
+	}
+
+	if (fl->fl6_src == NULL) {
+		*saddr = kmalloc(sizeof(struct in6_addr), GFP_ATOMIC);
+		err = ipv6_get_saddr(*dst, fl->fl6_dst, *saddr);
+
+		if (err) {
+#if IP6_DEBUG >= 2
+			printk(KERN_DEBUG "ip6_build_xmit: "
+			       "no availiable source address\n");
+#endif
+			return err;
+		}
+		fl->fl6_src = *saddr;
+	}
+
+        if (*dst) {
+		if ((err = xfrm_lookup(dst, fl, sk, 0)) < 0) {
+			dst_release(*dst);	
+			return -ENETUNREACH;
+		}
+        }
+
+	return 0;
+}
+
+int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb),
+		    void *from, int length, int transhdrlen,
+		    int hlimit, struct ipv6_txoptions *opt, struct flowi *fl, struct rt6_info *rt,
+		    unsigned int flags)
+{
+	struct inet_opt *inet = inet_sk(sk);
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct sk_buff *skb;
+	unsigned int maxfraglen, fragheaderlen;
+	int exthdrlen;
+	int hh_len;
+	int mtu;
+	int copy = 0;
+	int err;
+	int offset = 0;
+	int csummode = CHECKSUM_NONE;
+
+	if (flags&MSG_PROBE)
+		return 0;
+	if (skb_queue_empty(&sk->write_queue)) {
+		/*
+		 * setup for corking
+		 */
+		if (opt) {
+			if (np->cork.opt == NULL)
+				np->cork.opt = kmalloc(opt->tot_len, sk->allocation);
+			memcpy(np->cork.opt, opt, opt->tot_len);
+			inet->cork.flags |= IPCORK_OPT;
+			/* need source address above miyazawa*/
+			exthdrlen += opt->opt_flen ? opt->opt_flen : 0;
+		}
+		dst_hold(&rt->u.dst);
+		np->cork.rt = rt;
+		np->cork.fl = fl;
+		inet->cork.fragsize = mtu = dst_pmtu(&rt->u.dst);
+		inet->cork.length = 0;
+		inet->sndmsg_page = NULL;
+		inet->sndmsg_off = 0;
+		if ((exthdrlen = rt->u.dst.header_len) != 0) {
+			length += exthdrlen;
+			transhdrlen += exthdrlen;
+		}
+	} else {
+		rt = np->cork.rt;
+		if (inet->cork.flags & IPCORK_OPT)
+			opt = np->cork.opt;
+		transhdrlen = 0;
+		exthdrlen = 0;
+		mtu = inet->cork.fragsize;
+	}
+
+	hh_len = (rt->u.dst.dev->hard_header_len&~15) + 16;
+
+	fragheaderlen = sizeof(struct ipv6hdr) + (opt ? opt->opt_nflen : 0);
+	maxfraglen = ((mtu - fragheaderlen) & ~7) + fragheaderlen - sizeof(struct frag_hdr);
+
+	if (mtu < 65576) {
+		if (inet->cork.length + length > 0xFFFF - fragheaderlen) {
+			ipv6_local_error(sk, EMSGSIZE, fl, mtu-exthdrlen);
+			return -EMSGSIZE;
+		}
+	}
+
+	inet->cork.length += length;
+
+	if ((skb = skb_peek_tail(&sk->write_queue)) == NULL)
+		goto alloc_new_skb;
+
+	while (length > 0) {
+		if ((copy = maxfraglen - skb->len) <= 0) {
+			char *data;
+			unsigned int datalen;
+			unsigned int fraglen;
+			unsigned int alloclen;
+			BUG_TRAP(copy == 0);
+alloc_new_skb:
+			datalen = maxfraglen - fragheaderlen;
+			if (datalen > length)
+				datalen = length;
+			fraglen = datalen + fragheaderlen;
+			if ((flags & MSG_MORE) &&
+			    !(rt->u.dst.dev->features&NETIF_F_SG))
+				alloclen = maxfraglen;
+			else
+				alloclen = fraglen;
+			alloclen += sizeof(struct frag_hdr);
+			if (transhdrlen) {
+				skb = sock_alloc_send_skb(sk,
+						alloclen + hh_len + 15,
+						(flags & MSG_DONTWAIT), &err);
+			} else {
+				skb = NULL;
+				if (atomic_read(&sk->wmem_alloc) <= 2*sk->sndbuf)
+					skb = sock_wmalloc(sk,
+							   alloclen + hh_len + 15, 1,
+							   sk->allocation);
+				if (unlikely(skb == NULL))
+					err = -ENOBUFS;
+			}
+			if (skb == NULL)
+				goto error;
+			/*
+			 *	Fill in the control structures
+			 */
+			skb->ip_summed = csummode;
+			skb->csum = 0;
+			/* reserve 8 byte for fragmentation */
+			skb_reserve(skb, hh_len+sizeof(struct frag_hdr));
+
+			/*
+			 *	Find where to start putting bytes
+			 */
+			data = skb_put(skb, fraglen);
+			skb->nh.raw = data + exthdrlen;
+			data += fragheaderlen;
+			skb->h.raw = data + exthdrlen;
+			copy = datalen - transhdrlen;
+			if (copy > 0 && getfrag(from, data + transhdrlen, offset, copy, 0, skb) < 0) {
+				err = -EFAULT;
+				kfree_skb(skb);
+				goto error;
+			}
+
+			offset += copy;
+			length -= datalen;
+			transhdrlen = 0;
+			exthdrlen = 0;
+			csummode = CHECKSUM_NONE;
+
+			/*
+			 * Put the packet on the pending queue
+			 */
+			__skb_queue_tail(&sk->write_queue, skb);
+			continue;
+		}
+
+		if (copy > length)
+			copy = length;
+
+		if (!(rt->u.dst.dev->features&NETIF_F_SG)) {
+			unsigned int off;
+
+			off = skb->len;
+			if (getfrag(from, skb_put(skb, copy),
+						offset, copy, off, skb) < 0) {
+				__skb_trim(skb, off);
+				err = -EFAULT;
+				goto error;
+			}
+		} else {
+			int i = skb_shinfo(skb)->nr_frags;
+			skb_frag_t *frag = &skb_shinfo(skb)->frags[i-1];
+			struct page *page = inet->sndmsg_page;
+			int off = inet->sndmsg_off;
+			unsigned int left;
+
+			if (page && (left = PAGE_SIZE - off) > 0) {
+				if (copy >= left)
+					copy = left;
+				if (page != frag->page) {
+					if (i == MAX_SKB_FRAGS) {
+						err = -EMSGSIZE;
+						goto error;
+					}
+					get_page(page);
+					skb_fill_page_desc(skb, i, page, inet->sndmsg_off, 0);
+					frag = &skb_shinfo(skb)->frags[i];
+				}
+			} else if(i < MAX_SKB_FRAGS) {
+				if (copy > PAGE_SIZE)
+					copy = PAGE_SIZE;
+				page = alloc_pages(sk->allocation, 0);
+				if (page == NULL) {
+					err = -ENOMEM;
+					goto error;
+				}
+				inet->sndmsg_page = page;
+				inet->sndmsg_off = 0;
+
+				skb_fill_page_desc(skb, i, page, 0, 0);
+				frag = &skb_shinfo(skb)->frags[i];
+				skb->truesize += PAGE_SIZE;
+				atomic_add(PAGE_SIZE, &sk->wmem_alloc);
+			} else {
+				err = -EMSGSIZE;
+				goto error;
+			}
+			if (getfrag(from, page_address(frag->page)+frag->page_offset+frag->size, offset, copy, skb->len, skb) < 0) {
+				err = -EFAULT;
+				goto error;
+			}
+			inet->sndmsg_off += copy;
+			frag->size += copy;
+			skb->len += copy;
+			skb->data_len += copy;
+		}
+		offset += copy;
+		length -= copy;
+	}
+	return 0;
+error:
+	inet->cork.length -= length;
+	IP6_INC_STATS(Ip6OutDiscards);
+	return err;
+}
+
+int ip6_push_pending_frames(struct sock *sk)
+{
+	struct sk_buff *skb, *tmp_skb;
+	struct sk_buff **tail_skb;
+	struct in6_addr *final_dst = NULL;
+	struct inet_opt *inet = inet_sk(sk);
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct ipv6hdr *hdr;
+	struct ipv6_txoptions *opt = np->cork.opt;
+	struct rt6_info *rt = np->cork.rt;
+	struct flowi *fl = np->cork.fl;
+	unsigned char proto = fl->proto;
+	int err = 0;
+
+	if ((skb = __skb_dequeue(&sk->write_queue)) == NULL)
+		goto out;
+	tail_skb = &(skb_shinfo(skb)->frag_list);
+
+	/* move skb->data to ip header from ext header */
+	if (skb->data < skb->nh.raw)
+		__skb_pull(skb, skb->nh.raw - skb->data);
+	while ((tmp_skb = __skb_dequeue(&sk->write_queue)) != NULL) {
+		__skb_pull(tmp_skb, skb->h.raw - skb->nh.raw);
+		*tail_skb = tmp_skb;
+		tail_skb = &(tmp_skb->next);
+		skb->len += tmp_skb->len;
+		skb->data_len += tmp_skb->len;
+#if 0 /* Logically correct, but useless work, ip_fragment() will have to undo */
+		skb->truesize += tmp_skb->truesize;
+		__sock_put(tmp_skb->sk);
+		tmp_skb->destructor = NULL;
+		tmp_skb->sk = NULL;
+#endif
+	}
+
+	final_dst = fl->fl6_dst;
+	__skb_pull(skb, skb->h.raw - skb->nh.raw);
+	if (opt && opt->opt_flen)
+		ipv6_push_frag_opts(skb, opt, &proto);
+	if (opt && opt->opt_nflen)
+		ipv6_push_nfrag_opts(skb, opt, &proto, &final_dst);
+
+	skb->nh.ipv6h = hdr = (struct ipv6hdr*) skb_push(skb, sizeof(struct ipv6hdr));
+	
+	*(u32*)hdr = fl->fl6_flowlabel | htonl(0x60000000);
+
+	if (skb->len < 65536)
+		hdr->payload_len = htons(skb->len - sizeof(struct ipv6hdr));
+	else
+		hdr->payload_len = 0;
+	hdr->hop_limit = np->hop_limit;
+	hdr->nexthdr = proto;
+	memcpy(&hdr->saddr, fl->fl6_src, sizeof(struct in6_addr));
+	memcpy(&hdr->daddr, final_dst, sizeof(struct in6_addr));
+
+	skb->dst = dst_clone(&rt->u.dst);
+	err = NF_HOOK(PF_INET6, NF_IP6_LOCAL_OUT, skb, NULL, skb->dst->dev, dst_output);
+	if (err) {
+		if (err > 0)
+			err = inet->recverr ? net_xmit_errno(err) : 0;
+		if (err)
+			goto error;
+	}
+
+out:
+	inet->cork.flags &= ~IPCORK_OPT;
+	if (np->cork.opt) {
+		kfree(np->cork.opt);
+		np->cork.opt = NULL;
+	}
+	if (np->cork.rt) {
+		np->cork.rt = NULL;
+	}
+	if (np->cork.fl) {
+		np->cork.fl = NULL;
+	}
+	return err;
+error:
+	goto out;
+}
+
+void ip6_flush_pending_frames(struct sock *sk)
+{
+	struct inet_opt *inet = inet_sk(sk);
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct sk_buff *skb;
+
+	while ((skb = __skb_dequeue_tail(&sk->write_queue)) != NULL)
+		kfree_skb(skb);
+
+	inet->cork.flags &= ~IPCORK_OPT;
+
+	if (np->cork.opt) {
+		kfree(np->cork.opt);
+		np->cork.opt = NULL;
+	}
+	if (np->cork.rt) {
+		np->cork.rt = NULL;
+	}
+	if (np->cork.fl) {
+		np->cork.fl = NULL;
+	}
+}

net/ipv6/raw.c

@@ -12,6 +12,7 @@
  *	Fixes:
  *	Hideaki YOSHIFUJI	:	sin6_scope_id support
  *	YOSHIFUJI,H.@USAGI	:	raw checksum (RFC2292(bis) compliance) 
+ *	Kazunori MIYAZAWA @USAGI:	change process style to use ip6_append_data
  *
  *	This program is free software; you can redistribute it and/or
  *      modify it under the terms of the GNU General Public License
@@ -29,6 +30,8 @@
 #include <linux/netdevice.h>
 #include <linux/if_arp.h>
 #include <linux/icmpv6.h>
+#include <linux/netfilter.h>
+#include <linux/netfilter_ipv6.h>
 #include <asm/uaccess.h>
 #include <asm/ioctls.h>
 
@@ -438,87 +441,115 @@ static int rawv6_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 	goto out_free;
 }
 
-/*
- *	Sending...
- */
+static int rawv6_push_pending_frames(struct sock *sk, struct flowi *fl, struct raw6_opt *opt, int len)
+{
+	struct sk_buff *skb;
+	int err = 0;
+	u16 *csum;
 
-struct rawv6_fakehdr {
-	struct iovec	*iov;
-	struct sock	*sk;
-	__u32		len;
-	__u32		cksum;
-	__u32		proto;
-	struct in6_addr *daddr;
-};
+	if ((skb = skb_peek(&sk->write_queue)) == NULL)
+		goto out;
 
-static int rawv6_getfrag(const void *data, struct in6_addr *saddr, 
-			  char *buff, unsigned int offset, unsigned int len)
-{
-	struct iovec *iov = (struct iovec *) data;
+	if (opt->offset + 1 < len)
+		csum = (u16 *)(skb->h.raw + opt->offset);
+	else {
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (skb_queue_len(&sk->write_queue) == 1) {
+		/*
+		 * Only one fragment on the socket.
+		 */
+		/* should be check HW csum miyazawa */
+		*csum = csum_ipv6_magic(fl->fl6_src,
+					     fl->fl6_dst,
+					     len, fl->proto, skb->csum);
+	} else {
+		u32 tmp_csum = 0;
+
+		skb_queue_walk(&sk->write_queue, skb) {
+			tmp_csum = csum_add(tmp_csum, skb->csum);
+		}
 
-	return memcpy_fromiovecend(buff, iov, offset, len);
+		tmp_csum = csum_ipv6_magic(fl->fl6_src,
+					fl->fl6_dst,
+					len, fl->proto, tmp_csum);
+		*csum = tmp_csum;
+	}
+	if (*csum == 0)
+		*csum = -1;
+	ip6_push_pending_frames(sk);
+out:
+	return err;
 }
 
-static int rawv6_frag_cksum(const void *data, struct in6_addr *addr,
-			     char *buff, unsigned int offset, 
-			     unsigned int len)
+static int rawv6_send_hdrinc(struct sock *sk, void *from, int length,
+			struct flowi *fl, struct rt6_info *rt, 
+			unsigned int flags)
 {
-	struct rawv6_fakehdr *hdr = (struct rawv6_fakehdr *) data;
-	
-	if (csum_partial_copy_fromiovecend(buff, hdr->iov, offset, 
-						    len, &hdr->cksum))
-		return -EFAULT;
-	
-	if (offset == 0) {
-		struct sock *sk;
-		struct raw6_opt *opt;
-		struct in6_addr *daddr;
-		
-		sk = hdr->sk;
-		opt = raw6_sk(sk);
+	struct inet_opt *inet = inet_sk(sk);
+	struct ipv6hdr *iph;
+	struct sk_buff *skb;
+	unsigned int hh_len;
+	int err;
 
-		if (hdr->daddr)
-			daddr = hdr->daddr;
-		else
-			daddr = addr + 1;
-		
-		hdr->cksum = csum_ipv6_magic(addr, daddr, hdr->len,
-					     hdr->proto, hdr->cksum);
-		
-		if (opt->offset + 1 < len) {
-			__u16 *csum;
-
-			csum = (__u16 *) (buff + opt->offset);
-			if (*csum) {
-				/* in case cksum was not initialized */
-				__u32 sum = hdr->cksum;
-				sum += *csum;
-				*csum = hdr->cksum = (sum + (sum>>16));
-			} else {
-				*csum = hdr->cksum;
-			}
-		} else {
-			if (net_ratelimit())
-				printk(KERN_DEBUG "icmp: cksum offset too big\n");
-			return -EINVAL;
-		}
-	}	
-	return 0; 
-}
+	if (length > rt->u.dst.dev->mtu) {
+		ipv6_local_error(sk, EMSGSIZE, fl, rt->u.dst.dev->mtu);
+		return -EMSGSIZE;
+	}
+	if (flags&MSG_PROBE)
+		goto out;
+
+	hh_len = LL_RESERVED_SPACE(rt->u.dst.dev);
+
+	skb = sock_alloc_send_skb(sk, length+hh_len+15,
+				  flags&MSG_DONTWAIT, &err);
+	if (skb == NULL)
+		goto error; 
+	skb_reserve(skb, hh_len);
+
+	skb->priority = sk->priority;
+	skb->dst = dst_clone(&rt->u.dst);
+
+	skb->nh.ipv6h = iph = (struct ipv6hdr *)skb_put(skb, length);
+
+	skb->ip_summed = CHECKSUM_NONE;
+
+	skb->h.raw = skb->nh.raw;
+	err = memcpy_fromiovecend((void *)iph, from, 0, length);
+	if (err)
+		goto error_fault;
 
+	err = NF_HOOK(PF_INET6, NF_IP6_LOCAL_OUT, skb, NULL, rt->u.dst.dev,
+		      dst_output);
+	if (err > 0)
+		err = inet->recverr ? net_xmit_errno(err) : 0;
+	if (err)
+		goto error;
+out:
+	return 0;
 
+error_fault:
+	err = -EFAULT;
+	kfree_skb(skb);
+error:
+	IP6_INC_STATS(Ip6OutDiscards);
+	return err; 
+}
 static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, int len)
 {
 	struct ipv6_txoptions opt_space;
 	struct sockaddr_in6 * sin6 = (struct sockaddr_in6 *) msg->msg_name;
+	struct in6_addr *daddr, *saddr = NULL;
 	struct inet_opt *inet = inet_sk(sk);
 	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct raw6_opt *raw_opt = raw6_sk(sk);
 	struct ipv6_txoptions *opt = NULL;
 	struct ip6_flowlabel *flowlabel = NULL;
+	struct dst_entry *dst = NULL;
 	struct flowi fl;
 	int addr_len = msg->msg_namelen;
-	struct in6_addr *daddr;
-	struct raw6_opt *raw_opt;
 	int hlimit = -1;
 	u16 proto;
 	int err;
@@ -552,6 +583,8 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 
 		if (!proto)
 			proto = inet->num;
+		else if (proto != inet->num)
+			return(-EINVAL);
 
 		if (proto > 255)
 			return(-EINVAL);
@@ -590,6 +623,7 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 		 * unspecfied destination address 
 		 * treated as error... is this correct ?
 		 */
+		fl6_sock_release(flowlabel);
 		return(-EINVAL);
 	}
 
@@ -619,39 +653,71 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 	if (flowlabel)
 		opt = fl6_merge_options(&opt_space, flowlabel, opt);
 
-	raw_opt = raw6_sk(sk);
-
 	fl.proto = proto;
 	fl.fl6_dst = daddr;
 	if (fl.fl6_src == NULL && !ipv6_addr_any(&np->saddr))
 		fl.fl6_src = &np->saddr;
 	fl.fl_icmp_type = 0;
 	fl.fl_icmp_code = 0;
-	
-	if (raw_opt->checksum) {
-		struct rawv6_fakehdr hdr;
-		
-		hdr.iov = msg->msg_iov;
-		hdr.sk  = sk;
-		hdr.len = len;
-		hdr.cksum = 0;
-		hdr.proto = proto;
-
-		if (opt && opt->srcrt)
-			hdr.daddr = daddr;
+
+	/* merge ip6_build_xmit from ip6_output */
+	if (opt && opt->srcrt) {
+		struct rt0_hdr *rt0 = (struct rt0_hdr *) opt->srcrt;
+		fl.fl6_dst = rt0->addr;
+	}
+
+	if (!fl.oif && ipv6_addr_is_multicast(fl.nl_u.ip6_u.daddr))
+		fl.oif = np->mcast_oif;
+
+	err = ip6_dst_lookup(sk, &dst, &fl, &saddr);
+	if (err) goto out;
+
+	if (hlimit < 0) {
+		if (ipv6_addr_is_multicast(fl.fl6_dst))
+			hlimit = np->mcast_hops;
 		else
-			hdr.daddr = NULL;
+			hlimit = np->hop_limit;
+		if (hlimit < 0)
+			hlimit = ((struct rt6_info*)dst)->rt6i_hoplimit;
+	}
 
-		err = ip6_build_xmit(sk, rawv6_frag_cksum, &hdr, &fl, len,
-				     opt, hlimit, msg->msg_flags);
+	if (msg->msg_flags&MSG_CONFIRM)
+		goto do_confirm;
+
+back_from_confirm:
+	if (inet->hdrincl) {
+		err = rawv6_send_hdrinc(sk, msg->msg_iov, len, &fl, (struct rt6_info*)dst, msg->msg_flags);
 	} else {
-		err = ip6_build_xmit(sk, rawv6_getfrag, msg->msg_iov, &fl, len,
-				     opt, hlimit, msg->msg_flags);
+		lock_sock(sk);
+		err = ip6_append_data(sk, ip_generic_getfrag, msg->msg_iov, len, 0,
+					hlimit, opt, &fl, (struct rt6_info*)dst, msg->msg_flags);
+
+		if (err)
+			ip6_flush_pending_frames(sk);
+		else if (!(msg->msg_flags & MSG_MORE)) {
+			if (raw_opt->checksum) {
+				err = rawv6_push_pending_frames(sk, &fl, raw_opt, len);
+			} else {
+				err = ip6_push_pending_frames(sk);
+			}
+		}
 	}
+done:
+	ip6_dst_store(sk, dst, fl.nl_u.ip6_u.daddr == &np->daddr ? &np->daddr : NULL);
+	if (err > 0)
+		err = np->recverr ? net_xmit_errno(err) : 0;
 
+	release_sock(sk);
+out:	
 	fl6_sock_release(flowlabel);
-
+	if (saddr) kfree(saddr);
 	return err<0?err:len;
+do_confirm:
+	dst_confirm(dst);
+	if (!(msg->msg_flags & MSG_PROBE) || len)
+		goto back_from_confirm;
+	err = 0;
+	goto done;
 }
 
 static int rawv6_seticmpfilter(struct sock *sk, int level, int optname, 

net/ipv6/udp.c

@@ -14,6 +14,7 @@
  *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
  *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
  *					a single port at the same time.
+ *      Kazunori MIYAZAWA @USAGI:       change process style to use ip6_append_data
  *
  *	This program is free software; you can redistribute it and/or
  *      modify it under the terms of the GNU General Public License
@@ -738,96 +739,117 @@ static int udpv6_rcv(struct sk_buff **pskb, unsigned int *nhoffp)
 	kfree_skb(skb);
 	return(0);	
 }
-
 /*
- *	Sending
+ * Throw away all pending data and cancel the corking. Socket is locked.
  */
-
-struct udpv6fakehdr 
+static void udp_v6_flush_pending_frames(struct sock *sk)
 {
-	struct udphdr	uh;
-	struct iovec	*iov;
-	__u32		wcheck;
-	__u32		pl_len;
-	struct in6_addr *daddr;
-};
+	struct udp_opt *up = udp_sk(sk);
+
+	if (up->pending) {
+		up->pending = 0;
+		ip6_flush_pending_frames(sk);
+        }
+}
 
 /*
- *	with checksum
+ *	Sending
  */
 
-static int udpv6_getfrag(const void *data, struct in6_addr *addr,
-			 char *buff, unsigned int offset, unsigned int len)
+static int udp_v6_push_pending_frames(struct sock *sk, struct udp_opt *up)
 {
-	struct udpv6fakehdr *udh = (struct udpv6fakehdr *) data;
-	char *dst;
-	int final = 0;
-	int clen = len;
+	struct sk_buff *skb;
+	struct udphdr *uh;
+	struct ipv6_pinfo *np = inet6_sk(sk);
+	struct flowi *fl = np->cork.fl;
+	int err = 0;
 
-	dst = buff;
+	/* Grab the skbuff where UDP header space exists. */
+	if ((skb = skb_peek(&sk->write_queue)) == NULL)
+		goto out;
 
-	if (offset) {
-		offset -= sizeof(struct udphdr);
-	} else {
-		dst += sizeof(struct udphdr);
-		final = 1;
-		clen -= sizeof(struct udphdr);
+	/*
+	 * Create a UDP header
+	 */
+	uh = skb->h.uh;
+	uh->source = fl->fl_ip_sport;
+	uh->dest = fl->fl_ip_dport;
+	uh->len = htons(up->len);
+	uh->check = 0;
+
+	if (sk->no_check == UDP_CSUM_NOXMIT) {
+		skb->ip_summed = CHECKSUM_NONE;
+		goto send;
 	}
 
-	if (csum_partial_copy_fromiovecend(dst, udh->iov, offset,
-					   clen, &udh->wcheck))
-		return -EFAULT;
-
-	if (final) {
-		struct in6_addr *daddr;
-		
-		udh->wcheck = csum_partial((char *)udh, sizeof(struct udphdr),
-					   udh->wcheck);
+	if (skb_queue_len(&sk->write_queue) == 1) {
+		skb->csum = csum_partial((char *)uh,
+				sizeof(struct udphdr), skb->csum);
+		uh->check = csum_ipv6_magic(fl->fl6_src,
+					    fl->fl6_dst,
+					    up->len, fl->proto, skb->csum);
+	} else {
+		u32 tmp_csum = 0;
 
-		if (udh->daddr) {
-			daddr = udh->daddr;
-		} else {
-			/*
-			 *	use packet destination address
-			 *	this should improve cache locality
-			 */
-			daddr = addr + 1;
+		skb_queue_walk(&sk->write_queue, skb) {
+			tmp_csum = csum_add(tmp_csum, skb->csum);
 		}
-		udh->uh.check = csum_ipv6_magic(addr, daddr,
-						udh->pl_len, IPPROTO_UDP,
-						udh->wcheck);
-		if (udh->uh.check == 0)
-			udh->uh.check = -1;
+		tmp_csum = csum_partial((char *)uh,
+				sizeof(struct udphdr), tmp_csum);
+                tmp_csum = csum_ipv6_magic(fl->fl6_src,
+					   fl->fl6_dst,
+					   up->len, fl->proto, tmp_csum);
+                uh->check = tmp_csum;
 
-		memcpy(buff, udh, sizeof(struct udphdr));
 	}
-	return 0;
+	if (uh->check == 0)
+		uh->check = -1;
+
+send:
+	err = ip6_push_pending_frames(sk);
+out:
+	up->len = 0;
+	up->pending = 0;
+	return err;
 }
 
-static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, int ulen)
+static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, int len)
 {
 	struct ipv6_txoptions opt_space;
-	struct udpv6fakehdr udh;
+	struct udp_opt *up = udp_sk(sk);
 	struct inet_opt *inet = inet_sk(sk);
 	struct ipv6_pinfo *np = inet6_sk(sk);
 	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) msg->msg_name;
+	struct in6_addr *daddr, *saddr = NULL;
 	struct ipv6_txoptions *opt = NULL;
 	struct ip6_flowlabel *flowlabel = NULL;
 	struct flowi fl;
+	struct dst_entry *dst;
 	int addr_len = msg->msg_namelen;
-	struct in6_addr *daddr;
-	int len = ulen + sizeof(struct udphdr);
+	int ulen = len;
 	int addr_type;
 	int hlimit = -1;
-	
+	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
 	int err;
 	
 	/* Rough check on arithmetic overflow,
 	   better check is made in ip6_build_xmit
 	   */
-	if (ulen < 0 || ulen > INT_MAX - sizeof(struct udphdr))
+	if (len < 0 || len > INT_MAX - sizeof(struct udphdr))
 		return -EMSGSIZE;
 	
+	if (up->pending) {
+		/*
+		 * There are pending frames.
+		 * The socket lock must be held while it's corked.
+		 */
+		lock_sock(sk);
+		if (likely(up->pending))
+			goto do_append_data;
+		release_sock(sk);
+	}
+	ulen += sizeof(struct udphdr);
+
 	fl.fl6_flowlabel = 0;
 	fl.oif = 0;
 
@@ -835,7 +857,7 @@ static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 		if (sin6->sin6_family == AF_INET) {
 			if (__ipv6_only_sock(sk))
 				return -ENETUNREACH;
-			return udp_sendmsg(iocb, sk, msg, ulen);
+			return udp_sendmsg(iocb, sk, msg, len);
 		}
 
 		if (addr_len < SIN6_LEN_RFC2133)
@@ -847,7 +869,7 @@ static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 		if (sin6->sin6_port == 0)
 			return -EINVAL;
 
-		udh.uh.dest = sin6->sin6_port;
+		up->dport = sin6->sin6_port;
 		daddr = &sin6->sin6_addr;
 
 		if (np->sndflow) {
@@ -873,7 +895,7 @@ static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 		if (sk->state != TCP_ESTABLISHED)
 			return -ENOTCONN;
 
-		udh.uh.dest = inet->dport;
+		up->dport = inet->dport;
 		daddr = &np->daddr;
 		fl.fl6_flowlabel = np->flow_label;
 	}
@@ -888,15 +910,14 @@ static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 
 		sin.sin_family = AF_INET;
 		sin.sin_addr.s_addr = daddr->s6_addr32[3];
-		sin.sin_port = udh.uh.dest;
+		sin.sin_port = up->dport;
 		msg->msg_name = (struct sockaddr *)(&sin);
 		msg->msg_namelen = sizeof(sin);
 		fl6_sock_release(flowlabel);
 
-		return udp_sendmsg(iocb, sk, msg, ulen);
+		return udp_sendmsg(iocb, sk, msg, len);
 	}
 
-	udh.daddr = NULL;
 	if (!fl.oif)
 		fl.oif = sk->bound_dev_if;
 	fl.fl6_src = NULL;
@@ -922,33 +943,172 @@ static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg
 		opt = np->opt;
 	if (flowlabel)
 		opt = fl6_merge_options(&opt_space, flowlabel, opt);
-	if (opt && opt->srcrt)
-		udh.daddr = daddr;
-
-	udh.uh.source = inet->sport;
-	udh.uh.len = len < 0x10000 ? htons(len) : 0;
-	udh.uh.check = 0;
-	udh.iov = msg->msg_iov;
-	udh.wcheck = 0;
-	udh.pl_len = len;
 
 	fl.proto = IPPROTO_UDP;
 	fl.fl6_dst = daddr;
 	if (fl.fl6_src == NULL && !ipv6_addr_any(&np->saddr))
 		fl.fl6_src = &np->saddr;
-	fl.fl_ip_dport = udh.uh.dest;
-	fl.fl_ip_sport = udh.uh.source;
+	fl.fl_ip_dport = up->dport;
+	fl.fl_ip_sport = inet->sport;
+	
+	/* merge ip6_build_xmit from ip6_output */
+	if (opt && opt->srcrt) {
+		struct rt0_hdr *rt0 = (struct rt0_hdr *) opt->srcrt;
+		fl.fl6_dst = rt0->addr;
+	}
+
+	if (!fl.oif && ipv6_addr_is_multicast(fl.nl_u.ip6_u.daddr))
+		fl.oif = np->mcast_oif;
+
+	err = ip6_dst_lookup(sk, &dst, &fl, &saddr);
+	if (err) goto out;
+
+	if (hlimit < 0) {
+		if (ipv6_addr_is_multicast(fl.fl6_dst))
+			hlimit = np->mcast_hops;
+		else
+			hlimit = np->hop_limit;
+		if (hlimit < 0)
+			hlimit = ((struct rt6_info*)dst)->rt6i_hoplimit;
+	}
+
+	if (msg->msg_flags&MSG_CONFIRM)
+		goto do_confirm;
+back_from_confirm:
+
+	lock_sock(sk);
+	if (unlikely(up->pending)) {
+		/* The socket is already corked while preparing it. */
+		/* ... which is an evident application bug. --ANK */
+		release_sock(sk);
 
-	err = ip6_build_xmit(sk, udpv6_getfrag, &udh, &fl, len, opt, hlimit,
-			     msg->msg_flags);
+		NETDEBUG(if (net_ratelimit()) printk(KERN_DEBUG "udp cork app bug 2\n"));
+		err = -EINVAL;
+		goto out;
+	}
 
+	up->pending = 1;
+
+do_append_data:
+	up->len += ulen;
+	err = ip6_append_data(sk, ip_generic_getfrag, msg->msg_iov, ulen, sizeof(struct udphdr),
+			      hlimit, opt, &fl, (struct rt6_info*)dst,
+			      corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags);
+	if (err)
+		udp_v6_flush_pending_frames(sk);
+	else if (!corkreq)
+		err = udp_v6_push_pending_frames(sk, up);
+
+	ip6_dst_store(sk, dst, fl.nl_u.ip6_u.daddr == &np->daddr ? &np->daddr : NULL);
+	if (err > 0)
+		err = np->recverr ? net_xmit_errno(err) : 0;
+	release_sock(sk);
+out:
 	fl6_sock_release(flowlabel);
+	if (saddr) kfree(saddr);
+	if (!err) {
+		UDP6_INC_STATS_USER(UdpOutDatagrams);
+		return len;
+	}
+	return err;
 
-	if (err < 0)
-		return err;
+do_confirm:
+	dst_confirm(dst);
+	if (!(msg->msg_flags&MSG_PROBE) || len)
+		goto back_from_confirm;
+	err = 0;
+	goto out;
+}
+
+static int udpv6_destroy_sock(struct sock *sk)
+{
+	lock_sock(sk);
+	udp_v6_flush_pending_frames(sk);
+	release_sock(sk);
+
+	inet6_destroy_sock(sk);
+
+	return 0;
+}
+
+/*
+ *	Socket option code for UDP
+ */
+static int udpv6_setsockopt(struct sock *sk, int level, int optname, 
+			  char *optval, int optlen)
+{
+	struct udp_opt *up = udp_sk(sk);
+	int val;
+	int err = 0;
+
+	if (level != SOL_UDP)
+		return ipv6_setsockopt(sk, level, optname, optval, optlen);
+
+	if(optlen<sizeof(int))
+		return -EINVAL;
+
+	if (get_user(val, (int *)optval))
+		return -EFAULT;
+
+	switch(optname) {
+	case UDP_CORK:
+		if (val != 0) {
+			up->corkflag = 1;
+		} else {
+			up->corkflag = 0;
+			lock_sock(sk);
+			udp_v6_push_pending_frames(sk, up);
+			release_sock(sk);
+		}
+		break;
+		
+	case UDP_ENCAP:
+		up->encap_type = val;
+		break;
+
+	default:
+		err = -ENOPROTOOPT;
+		break;
+	};
+
+	return err;
+}
+
+static int udpv6_getsockopt(struct sock *sk, int level, int optname, 
+			  char *optval, int *optlen)
+{
+	struct udp_opt *up = udp_sk(sk);
+	int val, len;
+
+	if (level != SOL_UDP)
+		return ipv6_getsockopt(sk, level, optname, optval, optlen);
+
+	if(get_user(len,optlen))
+		return -EFAULT;
+
+	len = min_t(unsigned int, len, sizeof(int));
+	
+	if(len < 0)
+		return -EINVAL;
+
+	switch(optname) {
+	case UDP_CORK:
+		val = up->corkflag;
+		break;
+
+	case UDP_ENCAP:
+		val = up->encap_type;
+		break;
 
-	UDP6_INC_STATS_USER(UdpOutDatagrams);
-	return ulen;
+	default:
+		return -ENOPROTOOPT;
+	};
+
+  	if(put_user(len, optlen))
+  		return -EFAULT;
+	if(copy_to_user(optval, &val,len))
+		return -EFAULT;
+  	return 0;
 }
 
 static struct inet6_protocol udpv6_protocol = {
@@ -1038,9 +1198,9 @@ struct proto udpv6_prot = {
 	.connect =	udpv6_connect,
 	.disconnect =	udp_disconnect,
 	.ioctl =	udp_ioctl,
-	.destroy =	inet6_destroy_sock,
-	.setsockopt =	ipv6_setsockopt,
-	.getsockopt =	ipv6_getsockopt,
+	.destroy =	udpv6_destroy_sock,
+	.setsockopt =	udpv6_setsockopt,
+	.getsockopt =	udpv6_getsockopt,
 	.sendmsg =	udpv6_sendmsg,
 	.recvmsg =	udpv6_recvmsg,
 	.backlog_rcv =	udpv6_queue_rcv_skb,

net/key/af_key.c

@@ -1241,7 +1241,8 @@ static int pfkey_add(struct sock *sk, struct sk_buff *skb, struct sadb_msg *hdr,
 		}
 	}
 
-	if (x1 && x1->id.spi && hdr->sadb_msg_type == SADB_ADD) {
+	if (x1 && ((x1->id.spi && hdr->sadb_msg_type == SADB_ADD) ||
+	     (hdr->sadb_msg_type == SADB_UPDATE && xfrm_state_kern(x1)))) {
 		x->km.state = XFRM_STATE_DEAD;
 		xfrm_state_put(x);
 		xfrm_state_put(x1);
@@ -1286,6 +1287,11 @@ static int pfkey_delete(struct sock *sk, struct sk_buff *skb, struct sadb_msg *h
 	if (x == NULL)
 		return -ESRCH;
 
+	if (xfrm_state_kern(x)) {
+		xfrm_state_put(x);
+		return -EPERM;
+	}
+	
 	xfrm_state_delete(x);
 	xfrm_state_put(x);
 

net/netsyms.c

@@ -304,6 +304,7 @@ EXPORT_SYMBOL(xfrm_state_register_afinfo);
 EXPORT_SYMBOL(xfrm_state_unregister_afinfo);
 EXPORT_SYMBOL(xfrm_state_get_afinfo);
 EXPORT_SYMBOL(xfrm_state_put_afinfo);
+EXPORT_SYMBOL(xfrm_state_delete_tunnel);
 EXPORT_SYMBOL(xfrm_replay_check);
 EXPORT_SYMBOL(xfrm_replay_advance);
 EXPORT_SYMBOL(xfrm_check_selectors);
@@ -466,6 +467,8 @@ EXPORT_SYMBOL(sysctl_tcp_tw_recycle);
 EXPORT_SYMBOL(sysctl_max_syn_backlog);
 #endif
 
+EXPORT_SYMBOL(ip_generic_getfrag);
+
 #endif
 
 EXPORT_SYMBOL(tcp_read_sock);
@@ -570,6 +573,7 @@ EXPORT_SYMBOL(netif_rx);
 EXPORT_SYMBOL(netif_receive_skb);
 EXPORT_SYMBOL(dev_add_pack);
 EXPORT_SYMBOL(dev_remove_pack);
+EXPORT_SYMBOL(__dev_remove_pack);
 EXPORT_SYMBOL(dev_get);
 EXPORT_SYMBOL(dev_alloc);
 EXPORT_SYMBOL(dev_alloc_name);

net/packet/af_packet.c

@@ -774,6 +774,7 @@ static int packet_release(struct socket *sock)
 		 */
 		dev_remove_pack(&po->prot_hook);
 		po->running = 0;
+		po->num = 0;
 		__sock_put(sk);
 	}
 
@@ -819,9 +820,12 @@ static int packet_do_bind(struct sock *sk, struct net_device *dev, int protocol)
 
 	spin_lock(&po->bind_lock);
 	if (po->running) {
-		dev_remove_pack(&po->prot_hook);
 		__sock_put(sk);
 		po->running = 0;
+		po->num = 0;
+		spin_unlock(&po->bind_lock);
+		dev_remove_pack(&po->prot_hook);
+		spin_lock(&po->bind_lock);
 	}
 
 	po->num = protocol;
@@ -1374,7 +1378,7 @@ static int packet_notifier(struct notifier_block *this, unsigned long msg, void
 			if (dev->ifindex == po->ifindex) {
 				spin_lock(&po->bind_lock);
 				if (po->running) {
-					dev_remove_pack(&po->prot_hook);
+					__dev_remove_pack(&po->prot_hook);
 					__sock_put(sk);
 					po->running = 0;
 					sk->err = ENETDOWN;
@@ -1618,9 +1622,14 @@ static int packet_set_ring(struct sock *sk, struct tpacket_req *req, int closing
 
 	/* Detach socket from network */
 	spin_lock(&po->bind_lock);
-	if (po->running)
-		dev_remove_pack(&po->prot_hook);
+	if (po->running) {
+		__dev_remove_pack(&po->prot_hook);
+		po->num = 0;
+		po->running = 0;
+	}
 	spin_unlock(&po->bind_lock);
+		
+	synchronize_net();
 
 	err = -EBUSY;
 	if (closing || atomic_read(&po->mapped) == 0) {

net/sched/sch_ingress.c

@@ -222,11 +222,6 @@ used on the egress (might slow things for an iota)
 */
 
 	if (dev->qdisc_ingress) {
-		/* FIXME: Push down to ->enqueue functions --RR */
-		if (skb_is_nonlinear(*pskb)
-		    && skb_linearize(*pskb, GFP_ATOMIC) != 0)
-			return NF_DROP;
-
 		spin_lock(&dev->queue_lock);
 		if ((q = dev->qdisc_ingress) != NULL)
 			fwres = q->enqueue(skb, q);

net/xfrm/xfrm_state.c

@@ -172,6 +172,7 @@ struct xfrm_state *xfrm_state_alloc(void)
 	if (x) {
 		memset(x, 0, sizeof(struct xfrm_state));
 		atomic_set(&x->refcnt, 1);
+		atomic_set(&x->tunnel_users, 0);
 		INIT_LIST_HEAD(&x->bydst);
 		INIT_LIST_HEAD(&x->byspi);
 		init_timer(&x->timer);
@@ -234,6 +235,7 @@ static void __xfrm_state_delete(struct xfrm_state *x)
 
 void xfrm_state_delete(struct xfrm_state *x)
 {
+	xfrm_state_delete_tunnel(x);
 	spin_lock_bh(&x->lock);
 	__xfrm_state_delete(x);
 	spin_unlock_bh(&x->lock);
@@ -248,7 +250,8 @@ void xfrm_state_flush(u8 proto)
 	for (i = 0; i < XFRM_DST_HSIZE; i++) {
 restart:
 		list_for_each_entry(x, xfrm_state_bydst+i, bydst) {
-			if (proto == IPSEC_PROTO_ANY || x->id.proto == proto) {
+			if (!xfrm_state_kern(x) &&
+			    (proto == IPSEC_PROTO_ANY || x->id.proto == proto)) {
 				xfrm_state_hold(x);
 				spin_unlock_bh(&xfrm_state_lock);
 
@@ -790,6 +793,20 @@ void xfrm_state_put_afinfo(struct xfrm_state_afinfo *afinfo)
 	read_unlock(&afinfo->lock);
 }
 
+/* Temporarily located here until net/xfrm/xfrm_tunnel.c is created */
+void xfrm_state_delete_tunnel(struct xfrm_state *x)
+{
+	if (x->tunnel) {
+		struct xfrm_state *t = x->tunnel;
+
+		if (atomic_read(&t->tunnel_users) == 2)
+			xfrm_state_delete(t);
+		atomic_dec(&t->tunnel_users);
+		xfrm_state_put(t);
+		x->tunnel = NULL;
+	}
+}
+
 void __init xfrm_state_init(void)
 {
 	int i;

net/xfrm/xfrm_user.c

@@ -281,6 +281,11 @@ static int xfrm_del_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
 	if (x == NULL)
 		return -ESRCH;
 
+	if (xfrm_state_kern(x)) {
+		xfrm_state_put(x);
+		return -EPERM;
+	}
+
 	xfrm_state_delete(x);
 	xfrm_state_put(x);