blob: 230dd43fc9b32ef8614d9e441c8af72cc2cc3b40 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/*
2 * Definitions for the 'struct sk_buff' memory handlers.
3 *
4 * Authors:
5 * Alan Cox, <gw4pts@gw4pts.ampr.org>
6 * Florian La Roche, <rzsfl@rz.uni-sb.de>
7 *
8 * This program is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License
10 * as published by the Free Software Foundation; either version
11 * 2 of the License, or (at your option) any later version.
12 */
13
14#ifndef _LINUX_SKBUFF_H
15#define _LINUX_SKBUFF_H
16
Linus Torvalds1da177e2005-04-16 15:20:36 -070017#include <linux/kernel.h>
18#include <linux/compiler.h>
19#include <linux/time.h>
20#include <linux/cache.h>
21
22#include <asm/atomic.h>
23#include <asm/types.h>
24#include <linux/spinlock.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070025#include <linux/net.h>
Thomas Graf3fc7e8a2005-06-23 21:00:17 -070026#include <linux/textsearch.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070027#include <net/checksum.h>
Al Viroa80958f2006-12-04 20:41:19 +000028#include <linux/rcupdate.h>
Chris Leech97fc2f02006-05-23 17:55:33 -070029#include <linux/dmaengine.h>
Eric Dumazetb7aa0bf2007-04-19 16:16:32 -070030#include <linux/hrtimer.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070031
32#define HAVE_ALLOC_SKB /* For the drivers to know */
33#define HAVE_ALIGNABLE_SKB /* Ditto 8) */
Linus Torvalds1da177e2005-04-16 15:20:36 -070034
35#define CHECKSUM_NONE 0
Patrick McHardy84fa7932006-08-29 16:44:56 -070036#define CHECKSUM_PARTIAL 1
Linus Torvalds1da177e2005-04-16 15:20:36 -070037#define CHECKSUM_UNNECESSARY 2
Patrick McHardy84fa7932006-08-29 16:44:56 -070038#define CHECKSUM_COMPLETE 3
Linus Torvalds1da177e2005-04-16 15:20:36 -070039
40#define SKB_DATA_ALIGN(X) (((X) + (SMP_CACHE_BYTES - 1)) & \
41 ~(SMP_CACHE_BYTES - 1))
David S. Millerfc910a22007-03-25 20:27:59 -070042#define SKB_WITH_OVERHEAD(X) \
43 (((X) - sizeof(struct skb_shared_info)) & \
44 ~(SMP_CACHE_BYTES - 1))
45#define SKB_MAX_ORDER(X, ORDER) \
46 SKB_WITH_OVERHEAD((PAGE_SIZE << (ORDER)) - (X))
Linus Torvalds1da177e2005-04-16 15:20:36 -070047#define SKB_MAX_HEAD(X) (SKB_MAX_ORDER((X), 0))
48#define SKB_MAX_ALLOC (SKB_MAX_ORDER(0, 2))
49
50/* A. Checksumming of received packets by device.
51 *
52 * NONE: device failed to checksum this packet.
53 * skb->csum is undefined.
54 *
55 * UNNECESSARY: device parsed packet and wouldbe verified checksum.
56 * skb->csum is undefined.
57 * It is bad option, but, unfortunately, many of vendors do this.
58 * Apparently with secret goal to sell you new device, when you
59 * will add new protocol to your host. F.e. IPv6. 8)
60 *
Patrick McHardy84fa7932006-08-29 16:44:56 -070061 * COMPLETE: the most generic way. Device supplied checksum of _all_
Linus Torvalds1da177e2005-04-16 15:20:36 -070062 * the packet as seen by netif_rx in skb->csum.
63 * NOTE: Even if device supports only some protocols, but
Patrick McHardy84fa7932006-08-29 16:44:56 -070064 * is able to produce some skb->csum, it MUST use COMPLETE,
Linus Torvalds1da177e2005-04-16 15:20:36 -070065 * not UNNECESSARY.
66 *
67 * B. Checksumming on output.
68 *
69 * NONE: skb is checksummed by protocol or csum is not required.
70 *
Patrick McHardy84fa7932006-08-29 16:44:56 -070071 * PARTIAL: device is required to csum packet as seen by hard_start_xmit
Linus Torvalds1da177e2005-04-16 15:20:36 -070072 * from skb->h.raw to the end and to record the checksum
73 * at skb->h.raw+skb->csum.
74 *
75 * Device must show its capabilities in dev->features, set
76 * at device setup time.
77 * NETIF_F_HW_CSUM - it is clever device, it is able to checksum
78 * everything.
79 * NETIF_F_NO_CSUM - loopback or reliable single hop media.
80 * NETIF_F_IP_CSUM - device is dumb. It is able to csum only
81 * TCP/UDP over IPv4. Sigh. Vendors like this
82 * way by an unknown reason. Though, see comment above
83 * about CHECKSUM_UNNECESSARY. 8)
84 *
85 * Any questions? No questions, good. --ANK
86 */
87
Linus Torvalds1da177e2005-04-16 15:20:36 -070088struct net_device;
89
90#ifdef CONFIG_NETFILTER
91struct nf_conntrack {
92 atomic_t use;
93 void (*destroy)(struct nf_conntrack *);
94};
95
96#ifdef CONFIG_BRIDGE_NETFILTER
97struct nf_bridge_info {
98 atomic_t use;
99 struct net_device *physindev;
100 struct net_device *physoutdev;
101#if defined(CONFIG_VLAN_8021Q) || defined(CONFIG_VLAN_8021Q_MODULE)
102 struct net_device *netoutdev;
103#endif
104 unsigned int mask;
105 unsigned long data[32 / sizeof(unsigned long)];
106};
107#endif
108
109#endif
110
111struct sk_buff_head {
112 /* These two members must be first. */
113 struct sk_buff *next;
114 struct sk_buff *prev;
115
116 __u32 qlen;
117 spinlock_t lock;
118};
119
120struct sk_buff;
121
122/* To allow 64K frame to be packed as single skb without frag_list */
123#define MAX_SKB_FRAGS (65536/PAGE_SIZE + 2)
124
125typedef struct skb_frag_struct skb_frag_t;
126
127struct skb_frag_struct {
128 struct page *page;
129 __u16 page_offset;
130 __u16 size;
131};
132
133/* This data is invariant across clones and lives at
134 * the end of the header data, ie. at skb->end.
135 */
136struct skb_shared_info {
137 atomic_t dataref;
Benjamin LaHaise4947d3e2006-01-03 14:06:50 -0800138 unsigned short nr_frags;
Herbert Xu79671682006-06-22 02:40:14 -0700139 unsigned short gso_size;
140 /* Warning: this field is not always filled in (UFO)! */
141 unsigned short gso_segs;
142 unsigned short gso_type;
Al Viroae08e1f2006-11-08 00:27:11 -0800143 __be32 ip6_frag_id;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700144 struct sk_buff *frag_list;
145 skb_frag_t frags[MAX_SKB_FRAGS];
146};
147
148/* We divide dataref into two halves. The higher 16 bits hold references
149 * to the payload part of skb->data. The lower 16 bits hold references to
150 * the entire skb->data. It is up to the users of the skb to agree on
151 * where the payload starts.
152 *
153 * All users must obey the rule that the skb->data reference count must be
154 * greater than or equal to the payload reference count.
155 *
156 * Holding a reference to the payload part means that the user does not
157 * care about modifications to the header part of skb->data.
158 */
159#define SKB_DATAREF_SHIFT 16
160#define SKB_DATAREF_MASK ((1 << SKB_DATAREF_SHIFT) - 1)
161
David S. Millerd179cd12005-08-17 14:57:30 -0700162
163enum {
164 SKB_FCLONE_UNAVAILABLE,
165 SKB_FCLONE_ORIG,
166 SKB_FCLONE_CLONE,
167};
168
Herbert Xu79671682006-06-22 02:40:14 -0700169enum {
170 SKB_GSO_TCPV4 = 1 << 0,
Herbert Xuf83ef8c2006-06-30 13:37:03 -0700171 SKB_GSO_UDP = 1 << 1,
Herbert Xu576a30e2006-06-27 13:22:38 -0700172
173 /* This indicates the skb is from an untrusted source. */
174 SKB_GSO_DODGY = 1 << 2,
Michael Chanb0da85372006-06-29 12:30:00 -0700175
176 /* This indicates the tcp segment has CWR set. */
Herbert Xuf83ef8c2006-06-30 13:37:03 -0700177 SKB_GSO_TCP_ECN = 1 << 3,
178
179 SKB_GSO_TCPV6 = 1 << 4,
Herbert Xu79671682006-06-22 02:40:14 -0700180};
181
Linus Torvalds1da177e2005-04-16 15:20:36 -0700182/**
183 * struct sk_buff - socket buffer
184 * @next: Next buffer in list
185 * @prev: Previous buffer in list
Linus Torvalds1da177e2005-04-16 15:20:36 -0700186 * @sk: Socket we are owned by
Herbert Xu325ed822005-10-03 13:57:23 -0700187 * @tstamp: Time we arrived
Linus Torvalds1da177e2005-04-16 15:20:36 -0700188 * @dev: Device we arrived on/are leaving by
Patrick McHardyc01003c2007-03-29 11:46:52 -0700189 * @iif: ifindex of device we arrived on
Linus Torvalds1da177e2005-04-16 15:20:36 -0700190 * @h: Transport layer header
191 * @nh: Network layer header
192 * @mac: Link layer header
Martin Waitz67be2dd2005-05-01 08:59:26 -0700193 * @dst: destination entry
194 * @sp: the security path, used for xfrm
Linus Torvalds1da177e2005-04-16 15:20:36 -0700195 * @cb: Control buffer. Free for use by every layer. Put private vars here
196 * @len: Length of actual data
197 * @data_len: Data length
198 * @mac_len: Length of link layer header
199 * @csum: Checksum
Martin Waitz67be2dd2005-05-01 08:59:26 -0700200 * @local_df: allow local fragmentation
Linus Torvalds1da177e2005-04-16 15:20:36 -0700201 * @cloned: Head may be cloned (check refcnt to be sure)
202 * @nohdr: Payload reference only, must not modify header
203 * @pkt_type: Packet class
Randy Dunlapc83c2482005-10-18 22:07:41 -0700204 * @fclone: skbuff clone status
Linus Torvalds1da177e2005-04-16 15:20:36 -0700205 * @ip_summed: Driver fed us an IP checksum
206 * @priority: Packet queueing priority
207 * @users: User count - see {datagram,tcp}.c
208 * @protocol: Packet protocol from driver
Linus Torvalds1da177e2005-04-16 15:20:36 -0700209 * @truesize: Buffer size
210 * @head: Head of buffer
211 * @data: Data head pointer
212 * @tail: Tail pointer
213 * @end: End pointer
214 * @destructor: Destruct function
Thomas Graf82e91ff2006-11-09 15:19:14 -0800215 * @mark: Generic packet mark
Linus Torvalds1da177e2005-04-16 15:20:36 -0700216 * @nfct: Associated connection, if any
Randy Dunlapc83c2482005-10-18 22:07:41 -0700217 * @ipvs_property: skbuff is owned by ipvs
Linus Torvalds1da177e2005-04-16 15:20:36 -0700218 * @nfctinfo: Relationship of this skb to the connection
Randy Dunlap461ddf32005-11-20 21:25:15 -0800219 * @nfct_reasm: netfilter conntrack re-assembly pointer
Linus Torvalds1da177e2005-04-16 15:20:36 -0700220 * @nf_bridge: Saved data about a bridged frame - see br_netfilter.c
Linus Torvalds1da177e2005-04-16 15:20:36 -0700221 * @tc_index: Traffic control index
222 * @tc_verd: traffic control verdict
Randy Dunlapf4b8ea72006-06-22 16:00:11 -0700223 * @dma_cookie: a cookie to one of several possible DMA operations
224 * done by skb DMA functions
James Morris984bc162006-06-09 00:29:17 -0700225 * @secmark: security marking
Linus Torvalds1da177e2005-04-16 15:20:36 -0700226 */
227
228struct sk_buff {
229 /* These two members must be first. */
230 struct sk_buff *next;
231 struct sk_buff *prev;
232
Linus Torvalds1da177e2005-04-16 15:20:36 -0700233 struct sock *sk;
Eric Dumazetb7aa0bf2007-04-19 16:16:32 -0700234 ktime_t tstamp;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700235 struct net_device *dev;
Patrick McHardyc01003c2007-03-29 11:46:52 -0700236 int iif;
237 /* 4 byte hole on 64 bit*/
Linus Torvalds1da177e2005-04-16 15:20:36 -0700238
239 union {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700240 unsigned char *raw;
241 } h;
242
243 union {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700244 unsigned char *raw;
245 } nh;
246
247 union {
248 unsigned char *raw;
249 } mac;
250
251 struct dst_entry *dst;
252 struct sec_path *sp;
253
254 /*
255 * This is the control buffer. It is free to use for every
256 * layer. Please put your private variables there. If you
257 * want to keep them across layers you have to do a skb_clone()
258 * first. This is owned by whoever has the skb queued ATM.
259 */
Patrick McHardy3e3850e2006-01-06 23:04:54 -0800260 char cb[48];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700261
262 unsigned int len,
263 data_len,
Al Viro1f61ab52006-11-14 21:44:08 -0800264 mac_len;
Al Viroff1dcad2006-11-20 18:07:29 -0800265 union {
266 __wsum csum;
267 __u32 csum_offset;
268 };
Linus Torvalds1da177e2005-04-16 15:20:36 -0700269 __u32 priority;
Thomas Graf1cbb3382005-07-05 14:13:41 -0700270 __u8 local_df:1,
271 cloned:1,
272 ip_summed:2,
Harald Welte6869c4d2005-08-09 19:24:19 -0700273 nohdr:1,
274 nfctinfo:3;
David S. Millerd179cd12005-08-17 14:57:30 -0700275 __u8 pkt_type:3,
Patrick McHardyb84f4cc2005-11-20 21:19:21 -0800276 fclone:2,
277 ipvs_property:1;
Alexey Dobriyana0d3bea2005-08-11 16:05:50 -0700278 __be16 protocol;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700279
280 void (*destructor)(struct sk_buff *skb);
281#ifdef CONFIG_NETFILTER
Linus Torvalds1da177e2005-04-16 15:20:36 -0700282 struct nf_conntrack *nfct;
Yasuyuki Kozakai9fb9cbb2005-11-09 16:38:16 -0800283#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
284 struct sk_buff *nfct_reasm;
285#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700286#ifdef CONFIG_BRIDGE_NETFILTER
287 struct nf_bridge_info *nf_bridge;
288#endif
289#endif /* CONFIG_NETFILTER */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700290#ifdef CONFIG_NET_SCHED
Patrick McHardyb6b99eb2005-08-09 19:33:51 -0700291 __u16 tc_index; /* traffic control index */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700292#ifdef CONFIG_NET_CLS_ACT
Patrick McHardyb6b99eb2005-08-09 19:33:51 -0700293 __u16 tc_verd; /* traffic control verdict */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700294#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700295#endif
Chris Leech97fc2f02006-05-23 17:55:33 -0700296#ifdef CONFIG_NET_DMA
297 dma_cookie_t dma_cookie;
298#endif
James Morris984bc162006-06-09 00:29:17 -0700299#ifdef CONFIG_NETWORK_SECMARK
300 __u32 secmark;
301#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700302
Thomas Graf82e91ff2006-11-09 15:19:14 -0800303 __u32 mark;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700304
305 /* These elements must be at the end, see alloc_skb() for details. */
306 unsigned int truesize;
307 atomic_t users;
308 unsigned char *head,
309 *data,
310 *tail,
311 *end;
312};
313
314#ifdef __KERNEL__
315/*
316 * Handling routines are only of interest to the kernel
317 */
318#include <linux/slab.h>
319
320#include <asm/system.h>
321
Jörn Engel231d06a2006-03-20 21:28:35 -0800322extern void kfree_skb(struct sk_buff *skb);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700323extern void __kfree_skb(struct sk_buff *skb);
David S. Millerd179cd12005-08-17 14:57:30 -0700324extern struct sk_buff *__alloc_skb(unsigned int size,
Christoph Hellwigb30973f2006-12-06 20:32:36 -0800325 gfp_t priority, int fclone, int node);
David S. Millerd179cd12005-08-17 14:57:30 -0700326static inline struct sk_buff *alloc_skb(unsigned int size,
Al Virodd0fc662005-10-07 07:46:04 +0100327 gfp_t priority)
David S. Millerd179cd12005-08-17 14:57:30 -0700328{
Christoph Hellwigb30973f2006-12-06 20:32:36 -0800329 return __alloc_skb(size, priority, 0, -1);
David S. Millerd179cd12005-08-17 14:57:30 -0700330}
331
332static inline struct sk_buff *alloc_skb_fclone(unsigned int size,
Al Virodd0fc662005-10-07 07:46:04 +0100333 gfp_t priority)
David S. Millerd179cd12005-08-17 14:57:30 -0700334{
Christoph Hellwigb30973f2006-12-06 20:32:36 -0800335 return __alloc_skb(size, priority, 1, -1);
David S. Millerd179cd12005-08-17 14:57:30 -0700336}
337
Linus Torvalds1da177e2005-04-16 15:20:36 -0700338extern void kfree_skbmem(struct sk_buff *skb);
Victor Fusco86a76ca2005-07-08 14:57:47 -0700339extern struct sk_buff *skb_clone(struct sk_buff *skb,
Al Virodd0fc662005-10-07 07:46:04 +0100340 gfp_t priority);
Victor Fusco86a76ca2005-07-08 14:57:47 -0700341extern struct sk_buff *skb_copy(const struct sk_buff *skb,
Al Virodd0fc662005-10-07 07:46:04 +0100342 gfp_t priority);
Victor Fusco86a76ca2005-07-08 14:57:47 -0700343extern struct sk_buff *pskb_copy(struct sk_buff *skb,
Al Virodd0fc662005-10-07 07:46:04 +0100344 gfp_t gfp_mask);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700345extern int pskb_expand_head(struct sk_buff *skb,
Victor Fusco86a76ca2005-07-08 14:57:47 -0700346 int nhead, int ntail,
Al Virodd0fc662005-10-07 07:46:04 +0100347 gfp_t gfp_mask);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700348extern struct sk_buff *skb_realloc_headroom(struct sk_buff *skb,
349 unsigned int headroom);
350extern struct sk_buff *skb_copy_expand(const struct sk_buff *skb,
351 int newheadroom, int newtailroom,
Al Virodd0fc662005-10-07 07:46:04 +0100352 gfp_t priority);
Herbert Xu5b057c62006-06-23 02:06:41 -0700353extern int skb_pad(struct sk_buff *skb, int pad);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700354#define dev_kfree_skb(a) kfree_skb(a)
355extern void skb_over_panic(struct sk_buff *skb, int len,
356 void *here);
357extern void skb_under_panic(struct sk_buff *skb, int len,
358 void *here);
David S. Millerdc6de332006-04-20 00:10:50 -0700359extern void skb_truesize_bug(struct sk_buff *skb);
360
361static inline void skb_truesize_check(struct sk_buff *skb)
362{
363 if (unlikely((int)skb->truesize < sizeof(struct sk_buff) + skb->len))
364 skb_truesize_bug(skb);
365}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700366
Ananda Rajue89e9cf2005-10-18 15:46:41 -0700367extern int skb_append_datato_frags(struct sock *sk, struct sk_buff *skb,
368 int getfrag(void *from, char *to, int offset,
369 int len,int odd, struct sk_buff *skb),
370 void *from, int length);
371
Thomas Graf677e90e2005-06-23 20:59:51 -0700372struct skb_seq_state
373{
374 __u32 lower_offset;
375 __u32 upper_offset;
376 __u32 frag_idx;
377 __u32 stepped_offset;
378 struct sk_buff *root_skb;
379 struct sk_buff *cur_skb;
380 __u8 *frag_data;
381};
382
383extern void skb_prepare_seq_read(struct sk_buff *skb,
384 unsigned int from, unsigned int to,
385 struct skb_seq_state *st);
386extern unsigned int skb_seq_read(unsigned int consumed, const u8 **data,
387 struct skb_seq_state *st);
388extern void skb_abort_seq_read(struct skb_seq_state *st);
389
Thomas Graf3fc7e8a2005-06-23 21:00:17 -0700390extern unsigned int skb_find_text(struct sk_buff *skb, unsigned int from,
391 unsigned int to, struct ts_config *config,
392 struct ts_state *state);
393
Linus Torvalds1da177e2005-04-16 15:20:36 -0700394/* Internal */
395#define skb_shinfo(SKB) ((struct skb_shared_info *)((SKB)->end))
396
397/**
398 * skb_queue_empty - check if a queue is empty
399 * @list: queue head
400 *
401 * Returns true if the queue is empty, false otherwise.
402 */
403static inline int skb_queue_empty(const struct sk_buff_head *list)
404{
405 return list->next == (struct sk_buff *)list;
406}
407
408/**
409 * skb_get - reference buffer
410 * @skb: buffer to reference
411 *
412 * Makes another reference to a socket buffer and returns a pointer
413 * to the buffer.
414 */
415static inline struct sk_buff *skb_get(struct sk_buff *skb)
416{
417 atomic_inc(&skb->users);
418 return skb;
419}
420
421/*
422 * If users == 1, we are the only owner and are can avoid redundant
423 * atomic change.
424 */
425
426/**
Linus Torvalds1da177e2005-04-16 15:20:36 -0700427 * skb_cloned - is the buffer a clone
428 * @skb: buffer to check
429 *
430 * Returns true if the buffer was generated with skb_clone() and is
431 * one of multiple shared copies of the buffer. Cloned buffers are
432 * shared data so must not be written to under normal circumstances.
433 */
434static inline int skb_cloned(const struct sk_buff *skb)
435{
436 return skb->cloned &&
437 (atomic_read(&skb_shinfo(skb)->dataref) & SKB_DATAREF_MASK) != 1;
438}
439
440/**
441 * skb_header_cloned - is the header a clone
442 * @skb: buffer to check
443 *
444 * Returns true if modifying the header part of the buffer requires
445 * the data to be copied.
446 */
447static inline int skb_header_cloned(const struct sk_buff *skb)
448{
449 int dataref;
450
451 if (!skb->cloned)
452 return 0;
453
454 dataref = atomic_read(&skb_shinfo(skb)->dataref);
455 dataref = (dataref & SKB_DATAREF_MASK) - (dataref >> SKB_DATAREF_SHIFT);
456 return dataref != 1;
457}
458
459/**
460 * skb_header_release - release reference to header
461 * @skb: buffer to operate on
462 *
463 * Drop a reference to the header part of the buffer. This is done
464 * by acquiring a payload reference. You must not read from the header
465 * part of skb->data after this.
466 */
467static inline void skb_header_release(struct sk_buff *skb)
468{
469 BUG_ON(skb->nohdr);
470 skb->nohdr = 1;
471 atomic_add(1 << SKB_DATAREF_SHIFT, &skb_shinfo(skb)->dataref);
472}
473
474/**
475 * skb_shared - is the buffer shared
476 * @skb: buffer to check
477 *
478 * Returns true if more than one person has a reference to this
479 * buffer.
480 */
481static inline int skb_shared(const struct sk_buff *skb)
482{
483 return atomic_read(&skb->users) != 1;
484}
485
486/**
487 * skb_share_check - check if buffer is shared and if so clone it
488 * @skb: buffer to check
489 * @pri: priority for memory allocation
490 *
491 * If the buffer is shared the buffer is cloned and the old copy
492 * drops a reference. A new clone with a single reference is returned.
493 * If the buffer is not shared the original buffer is returned. When
494 * being called from interrupt status or with spinlocks held pri must
495 * be GFP_ATOMIC.
496 *
497 * NULL is returned on a memory allocation failure.
498 */
Victor Fusco86a76ca2005-07-08 14:57:47 -0700499static inline struct sk_buff *skb_share_check(struct sk_buff *skb,
Al Virodd0fc662005-10-07 07:46:04 +0100500 gfp_t pri)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700501{
502 might_sleep_if(pri & __GFP_WAIT);
503 if (skb_shared(skb)) {
504 struct sk_buff *nskb = skb_clone(skb, pri);
505 kfree_skb(skb);
506 skb = nskb;
507 }
508 return skb;
509}
510
511/*
512 * Copy shared buffers into a new sk_buff. We effectively do COW on
513 * packets to handle cases where we have a local reader and forward
514 * and a couple of other messy ones. The normal one is tcpdumping
515 * a packet thats being forwarded.
516 */
517
518/**
519 * skb_unshare - make a copy of a shared buffer
520 * @skb: buffer to check
521 * @pri: priority for memory allocation
522 *
523 * If the socket buffer is a clone then this function creates a new
524 * copy of the data, drops a reference count on the old copy and returns
525 * the new copy with the reference count at 1. If the buffer is not a clone
526 * the original buffer is returned. When called with a spinlock held or
527 * from interrupt state @pri must be %GFP_ATOMIC
528 *
529 * %NULL is returned on a memory allocation failure.
530 */
Victor Fuscoe2bf5212005-07-18 13:36:38 -0700531static inline struct sk_buff *skb_unshare(struct sk_buff *skb,
Al Virodd0fc662005-10-07 07:46:04 +0100532 gfp_t pri)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700533{
534 might_sleep_if(pri & __GFP_WAIT);
535 if (skb_cloned(skb)) {
536 struct sk_buff *nskb = skb_copy(skb, pri);
537 kfree_skb(skb); /* Free our shared copy */
538 skb = nskb;
539 }
540 return skb;
541}
542
543/**
544 * skb_peek
545 * @list_: list to peek at
546 *
547 * Peek an &sk_buff. Unlike most other operations you _MUST_
548 * be careful with this one. A peek leaves the buffer on the
549 * list and someone else may run off with it. You must hold
550 * the appropriate locks or have a private queue to do this.
551 *
552 * Returns %NULL for an empty list or a pointer to the head element.
553 * The reference count is not incremented and the reference is therefore
554 * volatile. Use with caution.
555 */
556static inline struct sk_buff *skb_peek(struct sk_buff_head *list_)
557{
558 struct sk_buff *list = ((struct sk_buff *)list_)->next;
559 if (list == (struct sk_buff *)list_)
560 list = NULL;
561 return list;
562}
563
564/**
565 * skb_peek_tail
566 * @list_: list to peek at
567 *
568 * Peek an &sk_buff. Unlike most other operations you _MUST_
569 * be careful with this one. A peek leaves the buffer on the
570 * list and someone else may run off with it. You must hold
571 * the appropriate locks or have a private queue to do this.
572 *
573 * Returns %NULL for an empty list or a pointer to the tail element.
574 * The reference count is not incremented and the reference is therefore
575 * volatile. Use with caution.
576 */
577static inline struct sk_buff *skb_peek_tail(struct sk_buff_head *list_)
578{
579 struct sk_buff *list = ((struct sk_buff *)list_)->prev;
580 if (list == (struct sk_buff *)list_)
581 list = NULL;
582 return list;
583}
584
585/**
586 * skb_queue_len - get queue length
587 * @list_: list to measure
588 *
589 * Return the length of an &sk_buff queue.
590 */
591static inline __u32 skb_queue_len(const struct sk_buff_head *list_)
592{
593 return list_->qlen;
594}
595
Arjan van de Ven76f10ad2006-08-02 14:06:55 -0700596/*
597 * This function creates a split out lock class for each invocation;
598 * this is needed for now since a whole lot of users of the skb-queue
599 * infrastructure in drivers have different locking usage (in hardirq)
600 * than the networking core (in softirq only). In the long run either the
601 * network layer or drivers should need annotation to consolidate the
602 * main types of usage into 3 classes.
603 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700604static inline void skb_queue_head_init(struct sk_buff_head *list)
605{
606 spin_lock_init(&list->lock);
607 list->prev = list->next = (struct sk_buff *)list;
608 list->qlen = 0;
609}
610
Pavel Emelianovc2ecba72007-04-17 12:45:31 -0700611static inline void skb_queue_head_init_class(struct sk_buff_head *list,
612 struct lock_class_key *class)
613{
614 skb_queue_head_init(list);
615 lockdep_set_class(&list->lock, class);
616}
617
Linus Torvalds1da177e2005-04-16 15:20:36 -0700618/*
619 * Insert an sk_buff at the start of a list.
620 *
621 * The "__skb_xxxx()" functions are the non-atomic ones that
622 * can only be called with interrupts disabled.
623 */
624
625/**
Stephen Hemminger300ce172005-10-30 13:47:34 -0800626 * __skb_queue_after - queue a buffer at the list head
627 * @list: list to use
628 * @prev: place after this buffer
629 * @newsk: buffer to queue
630 *
631 * Queue a buffer int the middle of a list. This function takes no locks
632 * and you must therefore hold required locks before calling it.
633 *
634 * A buffer cannot be placed on two lists at the same time.
635 */
636static inline void __skb_queue_after(struct sk_buff_head *list,
637 struct sk_buff *prev,
638 struct sk_buff *newsk)
639{
640 struct sk_buff *next;
641 list->qlen++;
642
643 next = prev->next;
644 newsk->next = next;
645 newsk->prev = prev;
646 next->prev = prev->next = newsk;
647}
648
649/**
Linus Torvalds1da177e2005-04-16 15:20:36 -0700650 * __skb_queue_head - queue a buffer at the list head
651 * @list: list to use
652 * @newsk: buffer to queue
653 *
654 * Queue a buffer at the start of a list. This function takes no locks
655 * and you must therefore hold required locks before calling it.
656 *
657 * A buffer cannot be placed on two lists at the same time.
658 */
659extern void skb_queue_head(struct sk_buff_head *list, struct sk_buff *newsk);
660static inline void __skb_queue_head(struct sk_buff_head *list,
661 struct sk_buff *newsk)
662{
Stephen Hemminger300ce172005-10-30 13:47:34 -0800663 __skb_queue_after(list, (struct sk_buff *)list, newsk);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700664}
665
666/**
667 * __skb_queue_tail - queue a buffer at the list tail
668 * @list: list to use
669 * @newsk: buffer to queue
670 *
671 * Queue a buffer at the end of a list. This function takes no locks
672 * and you must therefore hold required locks before calling it.
673 *
674 * A buffer cannot be placed on two lists at the same time.
675 */
676extern void skb_queue_tail(struct sk_buff_head *list, struct sk_buff *newsk);
677static inline void __skb_queue_tail(struct sk_buff_head *list,
678 struct sk_buff *newsk)
679{
680 struct sk_buff *prev, *next;
681
Linus Torvalds1da177e2005-04-16 15:20:36 -0700682 list->qlen++;
683 next = (struct sk_buff *)list;
684 prev = next->prev;
685 newsk->next = next;
686 newsk->prev = prev;
687 next->prev = prev->next = newsk;
688}
689
690
691/**
692 * __skb_dequeue - remove from the head of the queue
693 * @list: list to dequeue from
694 *
695 * Remove the head of the list. This function does not take any locks
696 * so must be used with appropriate locks held only. The head item is
697 * returned or %NULL if the list is empty.
698 */
699extern struct sk_buff *skb_dequeue(struct sk_buff_head *list);
700static inline struct sk_buff *__skb_dequeue(struct sk_buff_head *list)
701{
702 struct sk_buff *next, *prev, *result;
703
704 prev = (struct sk_buff *) list;
705 next = prev->next;
706 result = NULL;
707 if (next != prev) {
708 result = next;
709 next = next->next;
710 list->qlen--;
711 next->prev = prev;
712 prev->next = next;
713 result->next = result->prev = NULL;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700714 }
715 return result;
716}
717
718
719/*
720 * Insert a packet on a list.
721 */
David S. Miller8728b832005-08-09 19:25:21 -0700722extern void skb_insert(struct sk_buff *old, struct sk_buff *newsk, struct sk_buff_head *list);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700723static inline void __skb_insert(struct sk_buff *newsk,
724 struct sk_buff *prev, struct sk_buff *next,
725 struct sk_buff_head *list)
726{
727 newsk->next = next;
728 newsk->prev = prev;
729 next->prev = prev->next = newsk;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700730 list->qlen++;
731}
732
733/*
734 * Place a packet after a given packet in a list.
735 */
David S. Miller8728b832005-08-09 19:25:21 -0700736extern void skb_append(struct sk_buff *old, struct sk_buff *newsk, struct sk_buff_head *list);
737static inline void __skb_append(struct sk_buff *old, struct sk_buff *newsk, struct sk_buff_head *list)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700738{
David S. Miller8728b832005-08-09 19:25:21 -0700739 __skb_insert(newsk, old, old->next, list);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700740}
741
742/*
743 * remove sk_buff from list. _Must_ be called atomically, and with
744 * the list known..
745 */
David S. Miller8728b832005-08-09 19:25:21 -0700746extern void skb_unlink(struct sk_buff *skb, struct sk_buff_head *list);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700747static inline void __skb_unlink(struct sk_buff *skb, struct sk_buff_head *list)
748{
749 struct sk_buff *next, *prev;
750
751 list->qlen--;
752 next = skb->next;
753 prev = skb->prev;
754 skb->next = skb->prev = NULL;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700755 next->prev = prev;
756 prev->next = next;
757}
758
759
760/* XXX: more streamlined implementation */
761
762/**
763 * __skb_dequeue_tail - remove from the tail of the queue
764 * @list: list to dequeue from
765 *
766 * Remove the tail of the list. This function does not take any locks
767 * so must be used with appropriate locks held only. The tail item is
768 * returned or %NULL if the list is empty.
769 */
770extern struct sk_buff *skb_dequeue_tail(struct sk_buff_head *list);
771static inline struct sk_buff *__skb_dequeue_tail(struct sk_buff_head *list)
772{
773 struct sk_buff *skb = skb_peek_tail(list);
774 if (skb)
775 __skb_unlink(skb, list);
776 return skb;
777}
778
779
780static inline int skb_is_nonlinear(const struct sk_buff *skb)
781{
782 return skb->data_len;
783}
784
785static inline unsigned int skb_headlen(const struct sk_buff *skb)
786{
787 return skb->len - skb->data_len;
788}
789
790static inline int skb_pagelen(const struct sk_buff *skb)
791{
792 int i, len = 0;
793
794 for (i = (int)skb_shinfo(skb)->nr_frags - 1; i >= 0; i--)
795 len += skb_shinfo(skb)->frags[i].size;
796 return len + skb_headlen(skb);
797}
798
799static inline void skb_fill_page_desc(struct sk_buff *skb, int i,
800 struct page *page, int off, int size)
801{
802 skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
803
804 frag->page = page;
805 frag->page_offset = off;
806 frag->size = size;
807 skb_shinfo(skb)->nr_frags = i + 1;
808}
809
810#define SKB_PAGE_ASSERT(skb) BUG_ON(skb_shinfo(skb)->nr_frags)
811#define SKB_FRAG_ASSERT(skb) BUG_ON(skb_shinfo(skb)->frag_list)
812#define SKB_LINEAR_ASSERT(skb) BUG_ON(skb_is_nonlinear(skb))
813
814/*
815 * Add data to an sk_buff
816 */
817static inline unsigned char *__skb_put(struct sk_buff *skb, unsigned int len)
818{
819 unsigned char *tmp = skb->tail;
820 SKB_LINEAR_ASSERT(skb);
821 skb->tail += len;
822 skb->len += len;
823 return tmp;
824}
825
826/**
827 * skb_put - add data to a buffer
828 * @skb: buffer to use
829 * @len: amount of data to add
830 *
831 * This function extends the used data area of the buffer. If this would
832 * exceed the total buffer size the kernel will panic. A pointer to the
833 * first byte of the extra data is returned.
834 */
835static inline unsigned char *skb_put(struct sk_buff *skb, unsigned int len)
836{
837 unsigned char *tmp = skb->tail;
838 SKB_LINEAR_ASSERT(skb);
839 skb->tail += len;
840 skb->len += len;
841 if (unlikely(skb->tail>skb->end))
842 skb_over_panic(skb, len, current_text_addr());
843 return tmp;
844}
845
846static inline unsigned char *__skb_push(struct sk_buff *skb, unsigned int len)
847{
848 skb->data -= len;
849 skb->len += len;
850 return skb->data;
851}
852
853/**
854 * skb_push - add data to the start of a buffer
855 * @skb: buffer to use
856 * @len: amount of data to add
857 *
858 * This function extends the used data area of the buffer at the buffer
859 * start. If this would exceed the total buffer headroom the kernel will
860 * panic. A pointer to the first byte of the extra data is returned.
861 */
862static inline unsigned char *skb_push(struct sk_buff *skb, unsigned int len)
863{
864 skb->data -= len;
865 skb->len += len;
866 if (unlikely(skb->data<skb->head))
867 skb_under_panic(skb, len, current_text_addr());
868 return skb->data;
869}
870
871static inline unsigned char *__skb_pull(struct sk_buff *skb, unsigned int len)
872{
873 skb->len -= len;
874 BUG_ON(skb->len < skb->data_len);
875 return skb->data += len;
876}
877
878/**
879 * skb_pull - remove data from the start of a buffer
880 * @skb: buffer to use
881 * @len: amount of data to remove
882 *
883 * This function removes data from the start of a buffer, returning
884 * the memory to the headroom. A pointer to the next data in the buffer
885 * is returned. Once the data has been pulled future pushes will overwrite
886 * the old data.
887 */
888static inline unsigned char *skb_pull(struct sk_buff *skb, unsigned int len)
889{
890 return unlikely(len > skb->len) ? NULL : __skb_pull(skb, len);
891}
892
893extern unsigned char *__pskb_pull_tail(struct sk_buff *skb, int delta);
894
895static inline unsigned char *__pskb_pull(struct sk_buff *skb, unsigned int len)
896{
897 if (len > skb_headlen(skb) &&
898 !__pskb_pull_tail(skb, len-skb_headlen(skb)))
899 return NULL;
900 skb->len -= len;
901 return skb->data += len;
902}
903
904static inline unsigned char *pskb_pull(struct sk_buff *skb, unsigned int len)
905{
906 return unlikely(len > skb->len) ? NULL : __pskb_pull(skb, len);
907}
908
909static inline int pskb_may_pull(struct sk_buff *skb, unsigned int len)
910{
911 if (likely(len <= skb_headlen(skb)))
912 return 1;
913 if (unlikely(len > skb->len))
914 return 0;
915 return __pskb_pull_tail(skb, len-skb_headlen(skb)) != NULL;
916}
917
918/**
919 * skb_headroom - bytes at buffer head
920 * @skb: buffer to check
921 *
922 * Return the number of bytes of free space at the head of an &sk_buff.
923 */
924static inline int skb_headroom(const struct sk_buff *skb)
925{
926 return skb->data - skb->head;
927}
928
929/**
930 * skb_tailroom - bytes at buffer end
931 * @skb: buffer to check
932 *
933 * Return the number of bytes of free space at the tail of an sk_buff
934 */
935static inline int skb_tailroom(const struct sk_buff *skb)
936{
937 return skb_is_nonlinear(skb) ? 0 : skb->end - skb->tail;
938}
939
940/**
941 * skb_reserve - adjust headroom
942 * @skb: buffer to alter
943 * @len: bytes to move
944 *
945 * Increase the headroom of an empty &sk_buff by reducing the tail
946 * room. This is only allowed for an empty buffer.
947 */
David S. Miller8243126c2006-01-17 02:54:21 -0800948static inline void skb_reserve(struct sk_buff *skb, int len)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700949{
950 skb->data += len;
951 skb->tail += len;
952}
953
Arnaldo Carvalho de Melo9c702202007-04-25 18:04:18 -0700954static inline unsigned char *skb_transport_header(const struct sk_buff *skb)
955{
956 return skb->h.raw;
957}
958
Arnaldo Carvalho de Melobadff6d2007-03-13 13:06:52 -0300959static inline void skb_reset_transport_header(struct sk_buff *skb)
960{
961 skb->h.raw = skb->data;
962}
963
Arnaldo Carvalho de Melo967b05f2007-03-13 13:51:52 -0300964static inline void skb_set_transport_header(struct sk_buff *skb,
965 const int offset)
966{
967 skb->h.raw = skb->data + offset;
968}
969
Arnaldo Carvalho de Meloea2ae172007-04-25 17:55:53 -0700970static inline int skb_transport_offset(const struct sk_buff *skb)
971{
972 return skb->h.raw - skb->data;
973}
974
Arnaldo Carvalho de Melod56f90a2007-04-10 20:50:43 -0700975static inline unsigned char *skb_network_header(const struct sk_buff *skb)
976{
977 return skb->nh.raw;
978}
979
Arnaldo Carvalho de Meloc1d2bbe2007-04-10 20:45:18 -0700980static inline void skb_reset_network_header(struct sk_buff *skb)
981{
982 skb->nh.raw = skb->data;
983}
984
Arnaldo Carvalho de Meloc14d2452007-03-11 22:39:41 -0300985static inline void skb_set_network_header(struct sk_buff *skb, const int offset)
986{
987 skb->nh.raw = skb->data + offset;
988}
989
Arnaldo Carvalho de Melobbe735e2007-03-10 22:16:10 -0300990static inline int skb_network_offset(const struct sk_buff *skb)
991{
992 return skb->nh.raw - skb->data;
993}
994
Arnaldo Carvalho de Melocfe1fc72007-03-16 17:26:39 -0300995static inline u32 skb_network_header_len(const struct sk_buff *skb)
996{
997 return skb->h.raw - skb->nh.raw;
998}
999
Arnaldo Carvalho de Melo98e399f2007-03-19 15:33:04 -07001000static inline unsigned char *skb_mac_header(const struct sk_buff *skb)
1001{
1002 return skb->mac.raw;
1003}
1004
1005static inline int skb_mac_header_was_set(const struct sk_buff *skb)
1006{
1007 return skb->mac.raw != NULL;
1008}
1009
Arnaldo Carvalho de Melo459a98e2007-03-19 15:30:44 -07001010static inline void skb_reset_mac_header(struct sk_buff *skb)
1011{
1012 skb->mac.raw = skb->data;
1013}
1014
Arnaldo Carvalho de Melo48d49d0c2007-03-10 12:30:58 -03001015static inline void skb_set_mac_header(struct sk_buff *skb, const int offset)
1016{
1017 skb->mac.raw = skb->data + offset;
1018}
1019
Linus Torvalds1da177e2005-04-16 15:20:36 -07001020/*
1021 * CPUs often take a performance hit when accessing unaligned memory
1022 * locations. The actual performance hit varies, it can be small if the
1023 * hardware handles it or large if we have to take an exception and fix it
1024 * in software.
1025 *
1026 * Since an ethernet header is 14 bytes network drivers often end up with
1027 * the IP header at an unaligned offset. The IP header can be aligned by
1028 * shifting the start of the packet by 2 bytes. Drivers should do this
1029 * with:
1030 *
1031 * skb_reserve(NET_IP_ALIGN);
1032 *
1033 * The downside to this alignment of the IP header is that the DMA is now
1034 * unaligned. On some architectures the cost of an unaligned DMA is high
1035 * and this cost outweighs the gains made by aligning the IP header.
1036 *
1037 * Since this trade off varies between architectures, we allow NET_IP_ALIGN
1038 * to be overridden.
1039 */
1040#ifndef NET_IP_ALIGN
1041#define NET_IP_ALIGN 2
1042#endif
1043
Anton Blanchard025be812006-03-31 02:27:06 -08001044/*
1045 * The networking layer reserves some headroom in skb data (via
1046 * dev_alloc_skb). This is used to avoid having to reallocate skb data when
1047 * the header has to grow. In the default case, if the header has to grow
1048 * 16 bytes or less we avoid the reallocation.
1049 *
1050 * Unfortunately this headroom changes the DMA alignment of the resulting
1051 * network packet. As for NET_IP_ALIGN, this unaligned DMA is expensive
1052 * on some architectures. An architecture can override this value,
1053 * perhaps setting it to a cacheline in size (since that will maintain
1054 * cacheline alignment of the DMA). It must be a power of 2.
1055 *
1056 * Various parts of the networking layer expect at least 16 bytes of
1057 * headroom, you should not reduce this.
1058 */
1059#ifndef NET_SKB_PAD
1060#define NET_SKB_PAD 16
1061#endif
1062
Herbert Xu3cc0e872006-06-09 16:13:38 -07001063extern int ___pskb_trim(struct sk_buff *skb, unsigned int len);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001064
1065static inline void __skb_trim(struct sk_buff *skb, unsigned int len)
1066{
Herbert Xu3cc0e872006-06-09 16:13:38 -07001067 if (unlikely(skb->data_len)) {
1068 WARN_ON(1);
1069 return;
1070 }
1071 skb->len = len;
1072 skb->tail = skb->data + len;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001073}
1074
1075/**
1076 * skb_trim - remove end from a buffer
1077 * @skb: buffer to alter
1078 * @len: new length
1079 *
1080 * Cut the length of a buffer down by removing data from the tail. If
1081 * the buffer is already under the length specified it is not modified.
Herbert Xu3cc0e872006-06-09 16:13:38 -07001082 * The skb must be linear.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001083 */
1084static inline void skb_trim(struct sk_buff *skb, unsigned int len)
1085{
1086 if (skb->len > len)
1087 __skb_trim(skb, len);
1088}
1089
1090
1091static inline int __pskb_trim(struct sk_buff *skb, unsigned int len)
1092{
Herbert Xu3cc0e872006-06-09 16:13:38 -07001093 if (skb->data_len)
1094 return ___pskb_trim(skb, len);
1095 __skb_trim(skb, len);
1096 return 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001097}
1098
1099static inline int pskb_trim(struct sk_buff *skb, unsigned int len)
1100{
1101 return (len < skb->len) ? __pskb_trim(skb, len) : 0;
1102}
1103
1104/**
Herbert Xue9fa4f72006-08-13 20:12:58 -07001105 * pskb_trim_unique - remove end from a paged unique (not cloned) buffer
1106 * @skb: buffer to alter
1107 * @len: new length
1108 *
1109 * This is identical to pskb_trim except that the caller knows that
1110 * the skb is not cloned so we should never get an error due to out-
1111 * of-memory.
1112 */
1113static inline void pskb_trim_unique(struct sk_buff *skb, unsigned int len)
1114{
1115 int err = pskb_trim(skb, len);
1116 BUG_ON(err);
1117}
1118
1119/**
Linus Torvalds1da177e2005-04-16 15:20:36 -07001120 * skb_orphan - orphan a buffer
1121 * @skb: buffer to orphan
1122 *
1123 * If a buffer currently has an owner then we call the owner's
1124 * destructor function and make the @skb unowned. The buffer continues
1125 * to exist but is no longer charged to its former owner.
1126 */
1127static inline void skb_orphan(struct sk_buff *skb)
1128{
1129 if (skb->destructor)
1130 skb->destructor(skb);
1131 skb->destructor = NULL;
1132 skb->sk = NULL;
1133}
1134
1135/**
1136 * __skb_queue_purge - empty a list
1137 * @list: list to empty
1138 *
1139 * Delete all buffers on an &sk_buff list. Each buffer is removed from
1140 * the list and one reference dropped. This function does not take the
1141 * list lock and the caller must hold the relevant locks to use it.
1142 */
1143extern void skb_queue_purge(struct sk_buff_head *list);
1144static inline void __skb_queue_purge(struct sk_buff_head *list)
1145{
1146 struct sk_buff *skb;
1147 while ((skb = __skb_dequeue(list)) != NULL)
1148 kfree_skb(skb);
1149}
1150
1151/**
Christoph Hellwigb4e54de2006-07-24 15:31:14 -07001152 * __dev_alloc_skb - allocate an skbuff for receiving
Linus Torvalds1da177e2005-04-16 15:20:36 -07001153 * @length: length to allocate
1154 * @gfp_mask: get_free_pages mask, passed to alloc_skb
1155 *
1156 * Allocate a new &sk_buff and assign it a usage count of one. The
1157 * buffer has unspecified headroom built in. Users should allocate
1158 * the headroom they think they need without accounting for the
1159 * built in space. The built in space is used for optimisations.
1160 *
Christoph Hellwig766ea8c2006-08-07 15:49:53 -07001161 * %NULL is returned if there is no free memory.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001162 */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001163static inline struct sk_buff *__dev_alloc_skb(unsigned int length,
Al Virodd0fc662005-10-07 07:46:04 +01001164 gfp_t gfp_mask)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001165{
Anton Blanchard025be812006-03-31 02:27:06 -08001166 struct sk_buff *skb = alloc_skb(length + NET_SKB_PAD, gfp_mask);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001167 if (likely(skb))
Anton Blanchard025be812006-03-31 02:27:06 -08001168 skb_reserve(skb, NET_SKB_PAD);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001169 return skb;
1170}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001171
1172/**
Christoph Hellwigb4e54de2006-07-24 15:31:14 -07001173 * dev_alloc_skb - allocate an skbuff for receiving
Linus Torvalds1da177e2005-04-16 15:20:36 -07001174 * @length: length to allocate
1175 *
1176 * Allocate a new &sk_buff and assign it a usage count of one. The
1177 * buffer has unspecified headroom built in. Users should allocate
1178 * the headroom they think they need without accounting for the
1179 * built in space. The built in space is used for optimisations.
1180 *
Christoph Hellwig766ea8c2006-08-07 15:49:53 -07001181 * %NULL is returned if there is no free memory. Although this function
Linus Torvalds1da177e2005-04-16 15:20:36 -07001182 * allocates memory it can be called from an interrupt.
1183 */
1184static inline struct sk_buff *dev_alloc_skb(unsigned int length)
1185{
1186 return __dev_alloc_skb(length, GFP_ATOMIC);
1187}
1188
Christoph Hellwig8af27452006-07-31 22:35:23 -07001189extern struct sk_buff *__netdev_alloc_skb(struct net_device *dev,
1190 unsigned int length, gfp_t gfp_mask);
1191
1192/**
1193 * netdev_alloc_skb - allocate an skbuff for rx on a specific device
1194 * @dev: network device to receive on
1195 * @length: length to allocate
1196 *
1197 * Allocate a new &sk_buff and assign it a usage count of one. The
1198 * buffer has unspecified headroom built in. Users should allocate
1199 * the headroom they think they need without accounting for the
1200 * built in space. The built in space is used for optimisations.
1201 *
1202 * %NULL is returned if there is no free memory. Although this function
1203 * allocates memory it can be called from an interrupt.
1204 */
1205static inline struct sk_buff *netdev_alloc_skb(struct net_device *dev,
1206 unsigned int length)
1207{
1208 return __netdev_alloc_skb(dev, length, GFP_ATOMIC);
1209}
1210
Linus Torvalds1da177e2005-04-16 15:20:36 -07001211/**
1212 * skb_cow - copy header of skb when it is required
1213 * @skb: buffer to cow
1214 * @headroom: needed headroom
1215 *
1216 * If the skb passed lacks sufficient headroom or its data part
1217 * is shared, data is reallocated. If reallocation fails, an error
1218 * is returned and original skb is not changed.
1219 *
1220 * The result is skb with writable area skb->head...skb->tail
1221 * and at least @headroom of space at head.
1222 */
1223static inline int skb_cow(struct sk_buff *skb, unsigned int headroom)
1224{
Anton Blanchard025be812006-03-31 02:27:06 -08001225 int delta = (headroom > NET_SKB_PAD ? headroom : NET_SKB_PAD) -
1226 skb_headroom(skb);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001227
1228 if (delta < 0)
1229 delta = 0;
1230
1231 if (delta || skb_cloned(skb))
Anton Blanchard025be812006-03-31 02:27:06 -08001232 return pskb_expand_head(skb, (delta + (NET_SKB_PAD-1)) &
1233 ~(NET_SKB_PAD-1), 0, GFP_ATOMIC);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001234 return 0;
1235}
1236
1237/**
1238 * skb_padto - pad an skbuff up to a minimal size
1239 * @skb: buffer to pad
1240 * @len: minimal length
1241 *
1242 * Pads up a buffer to ensure the trailing bytes exist and are
1243 * blanked. If the buffer already contains sufficient data it
Herbert Xu5b057c62006-06-23 02:06:41 -07001244 * is untouched. Otherwise it is extended. Returns zero on
1245 * success. The skb is freed on error.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001246 */
1247
Herbert Xu5b057c62006-06-23 02:06:41 -07001248static inline int skb_padto(struct sk_buff *skb, unsigned int len)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001249{
1250 unsigned int size = skb->len;
1251 if (likely(size >= len))
Herbert Xu5b057c62006-06-23 02:06:41 -07001252 return 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001253 return skb_pad(skb, len-size);
1254}
1255
1256static inline int skb_add_data(struct sk_buff *skb,
1257 char __user *from, int copy)
1258{
1259 const int off = skb->len;
1260
1261 if (skb->ip_summed == CHECKSUM_NONE) {
1262 int err = 0;
Al Viro50842052006-11-14 21:36:34 -08001263 __wsum csum = csum_and_copy_from_user(from, skb_put(skb, copy),
Linus Torvalds1da177e2005-04-16 15:20:36 -07001264 copy, 0, &err);
1265 if (!err) {
1266 skb->csum = csum_block_add(skb->csum, csum, off);
1267 return 0;
1268 }
1269 } else if (!copy_from_user(skb_put(skb, copy), from, copy))
1270 return 0;
1271
1272 __skb_trim(skb, off);
1273 return -EFAULT;
1274}
1275
1276static inline int skb_can_coalesce(struct sk_buff *skb, int i,
1277 struct page *page, int off)
1278{
1279 if (i) {
1280 struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i - 1];
1281
1282 return page == frag->page &&
1283 off == frag->page_offset + frag->size;
1284 }
1285 return 0;
1286}
1287
Herbert Xu364c6ba2006-06-09 16:10:40 -07001288static inline int __skb_linearize(struct sk_buff *skb)
1289{
1290 return __pskb_pull_tail(skb, skb->data_len) ? 0 : -ENOMEM;
1291}
1292
Linus Torvalds1da177e2005-04-16 15:20:36 -07001293/**
1294 * skb_linearize - convert paged skb to linear one
1295 * @skb: buffer to linarize
Linus Torvalds1da177e2005-04-16 15:20:36 -07001296 *
1297 * If there is no free memory -ENOMEM is returned, otherwise zero
1298 * is returned and the old skb data released.
1299 */
Herbert Xu364c6ba2006-06-09 16:10:40 -07001300static inline int skb_linearize(struct sk_buff *skb)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001301{
Herbert Xu364c6ba2006-06-09 16:10:40 -07001302 return skb_is_nonlinear(skb) ? __skb_linearize(skb) : 0;
1303}
1304
1305/**
1306 * skb_linearize_cow - make sure skb is linear and writable
1307 * @skb: buffer to process
1308 *
1309 * If there is no free memory -ENOMEM is returned, otherwise zero
1310 * is returned and the old skb data released.
1311 */
1312static inline int skb_linearize_cow(struct sk_buff *skb)
1313{
1314 return skb_is_nonlinear(skb) || skb_cloned(skb) ?
1315 __skb_linearize(skb) : 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001316}
1317
1318/**
1319 * skb_postpull_rcsum - update checksum for received skb after pull
1320 * @skb: buffer to update
1321 * @start: start of data before pull
1322 * @len: length of data pulled
1323 *
1324 * After doing a pull on a received packet, you need to call this to
Patrick McHardy84fa7932006-08-29 16:44:56 -07001325 * update the CHECKSUM_COMPLETE checksum, or set ip_summed to
1326 * CHECKSUM_NONE so that it can be recomputed from scratch.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001327 */
1328
1329static inline void skb_postpull_rcsum(struct sk_buff *skb,
Herbert Xucbb042f2006-03-20 22:43:56 -08001330 const void *start, unsigned int len)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001331{
Patrick McHardy84fa7932006-08-29 16:44:56 -07001332 if (skb->ip_summed == CHECKSUM_COMPLETE)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001333 skb->csum = csum_sub(skb->csum, csum_partial(start, len, 0));
1334}
1335
Herbert Xucbb042f2006-03-20 22:43:56 -08001336unsigned char *skb_pull_rcsum(struct sk_buff *skb, unsigned int len);
1337
Linus Torvalds1da177e2005-04-16 15:20:36 -07001338/**
1339 * pskb_trim_rcsum - trim received skb and update checksum
1340 * @skb: buffer to trim
1341 * @len: new length
1342 *
1343 * This is exactly the same as pskb_trim except that it ensures the
1344 * checksum of received packets are still valid after the operation.
1345 */
1346
1347static inline int pskb_trim_rcsum(struct sk_buff *skb, unsigned int len)
1348{
Stephen Hemminger0e4e4222005-09-08 12:32:03 -07001349 if (likely(len >= skb->len))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001350 return 0;
Patrick McHardy84fa7932006-08-29 16:44:56 -07001351 if (skb->ip_summed == CHECKSUM_COMPLETE)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001352 skb->ip_summed = CHECKSUM_NONE;
1353 return __pskb_trim(skb, len);
1354}
1355
Linus Torvalds1da177e2005-04-16 15:20:36 -07001356#define skb_queue_walk(queue, skb) \
1357 for (skb = (queue)->next; \
1358 prefetch(skb->next), (skb != (struct sk_buff *)(queue)); \
1359 skb = skb->next)
1360
Stephen Hemminger300ce172005-10-30 13:47:34 -08001361#define skb_queue_reverse_walk(queue, skb) \
1362 for (skb = (queue)->prev; \
1363 prefetch(skb->prev), (skb != (struct sk_buff *)(queue)); \
1364 skb = skb->prev)
1365
Linus Torvalds1da177e2005-04-16 15:20:36 -07001366
1367extern struct sk_buff *skb_recv_datagram(struct sock *sk, unsigned flags,
1368 int noblock, int *err);
1369extern unsigned int datagram_poll(struct file *file, struct socket *sock,
1370 struct poll_table_struct *wait);
1371extern int skb_copy_datagram_iovec(const struct sk_buff *from,
1372 int offset, struct iovec *to,
1373 int size);
Herbert Xufb286bb2005-11-10 13:01:24 -08001374extern int skb_copy_and_csum_datagram_iovec(struct sk_buff *skb,
Linus Torvalds1da177e2005-04-16 15:20:36 -07001375 int hlen,
1376 struct iovec *iov);
1377extern void skb_free_datagram(struct sock *sk, struct sk_buff *skb);
Herbert Xu3305b802005-12-13 23:16:37 -08001378extern void skb_kill_datagram(struct sock *sk, struct sk_buff *skb,
1379 unsigned int flags);
Al Viro2bbbc862006-11-14 21:37:14 -08001380extern __wsum skb_checksum(const struct sk_buff *skb, int offset,
1381 int len, __wsum csum);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001382extern int skb_copy_bits(const struct sk_buff *skb, int offset,
1383 void *to, int len);
Herbert Xu357b40a2005-04-19 22:30:14 -07001384extern int skb_store_bits(const struct sk_buff *skb, int offset,
1385 void *from, int len);
Al Viro81d77662006-11-14 21:37:33 -08001386extern __wsum skb_copy_and_csum_bits(const struct sk_buff *skb,
Linus Torvalds1da177e2005-04-16 15:20:36 -07001387 int offset, u8 *to, int len,
Al Viro81d77662006-11-14 21:37:33 -08001388 __wsum csum);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001389extern void skb_copy_and_csum_dev(const struct sk_buff *skb, u8 *to);
1390extern void skb_split(struct sk_buff *skb,
1391 struct sk_buff *skb1, const u32 len);
1392
Herbert Xu576a30e2006-06-27 13:22:38 -07001393extern struct sk_buff *skb_segment(struct sk_buff *skb, int features);
Arnaldo Carvalho de Melo20380732005-08-16 02:18:02 -03001394
Linus Torvalds1da177e2005-04-16 15:20:36 -07001395static inline void *skb_header_pointer(const struct sk_buff *skb, int offset,
1396 int len, void *buffer)
1397{
1398 int hlen = skb_headlen(skb);
1399
Patrick McHardy55820ee2005-07-05 14:08:10 -07001400 if (hlen - offset >= len)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001401 return skb->data + offset;
1402
1403 if (skb_copy_bits(skb, offset, buffer, len) < 0)
1404 return NULL;
1405
1406 return buffer;
1407}
1408
1409extern void skb_init(void);
1410extern void skb_add_mtu(int mtu);
1411
Patrick McHardya61bbcf2005-08-14 17:24:31 -07001412/**
1413 * skb_get_timestamp - get timestamp from a skb
1414 * @skb: skb to get stamp from
1415 * @stamp: pointer to struct timeval to store stamp in
1416 *
1417 * Timestamps are stored in the skb as offsets to a base timestamp.
1418 * This function converts the offset back to a struct timeval and stores
1419 * it in stamp.
1420 */
Stephen Hemmingerf2c38392005-09-06 15:48:03 -07001421static inline void skb_get_timestamp(const struct sk_buff *skb, struct timeval *stamp)
Patrick McHardya61bbcf2005-08-14 17:24:31 -07001422{
Eric Dumazetb7aa0bf2007-04-19 16:16:32 -07001423 *stamp = ktime_to_timeval(skb->tstamp);
Patrick McHardya61bbcf2005-08-14 17:24:31 -07001424}
1425
Eric Dumazetb7aa0bf2007-04-19 16:16:32 -07001426static inline void __net_timestamp(struct sk_buff *skb)
Patrick McHardya61bbcf2005-08-14 17:24:31 -07001427{
Eric Dumazetb7aa0bf2007-04-19 16:16:32 -07001428 skb->tstamp = ktime_get_real();
Patrick McHardya61bbcf2005-08-14 17:24:31 -07001429}
1430
Patrick McHardya61bbcf2005-08-14 17:24:31 -07001431
Herbert Xu759e5d02007-03-25 20:10:56 -07001432extern __sum16 __skb_checksum_complete_head(struct sk_buff *skb, int len);
Al Virob51655b2006-11-14 21:40:42 -08001433extern __sum16 __skb_checksum_complete(struct sk_buff *skb);
Herbert Xufb286bb2005-11-10 13:01:24 -08001434
1435/**
1436 * skb_checksum_complete - Calculate checksum of an entire packet
1437 * @skb: packet to process
1438 *
1439 * This function calculates the checksum over the entire packet plus
1440 * the value of skb->csum. The latter can be used to supply the
1441 * checksum of a pseudo header as used by TCP/UDP. It returns the
1442 * checksum.
1443 *
1444 * For protocols that contain complete checksums such as ICMP/TCP/UDP,
1445 * this function can be used to verify that checksum on received
1446 * packets. In that case the function should return zero if the
1447 * checksum is correct. In particular, this function will return zero
1448 * if skb->ip_summed is CHECKSUM_UNNECESSARY which indicates that the
1449 * hardware has already verified the correctness of the checksum.
1450 */
1451static inline unsigned int skb_checksum_complete(struct sk_buff *skb)
1452{
1453 return skb->ip_summed != CHECKSUM_UNNECESSARY &&
1454 __skb_checksum_complete(skb);
1455}
1456
Linus Torvalds1da177e2005-04-16 15:20:36 -07001457#ifdef CONFIG_NETFILTER
1458static inline void nf_conntrack_put(struct nf_conntrack *nfct)
1459{
1460 if (nfct && atomic_dec_and_test(&nfct->use))
1461 nfct->destroy(nfct);
1462}
1463static inline void nf_conntrack_get(struct nf_conntrack *nfct)
1464{
1465 if (nfct)
1466 atomic_inc(&nfct->use);
1467}
Yasuyuki Kozakai9fb9cbb2005-11-09 16:38:16 -08001468#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
1469static inline void nf_conntrack_get_reasm(struct sk_buff *skb)
1470{
1471 if (skb)
1472 atomic_inc(&skb->users);
1473}
1474static inline void nf_conntrack_put_reasm(struct sk_buff *skb)
1475{
1476 if (skb)
1477 kfree_skb(skb);
1478}
1479#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07001480#ifdef CONFIG_BRIDGE_NETFILTER
1481static inline void nf_bridge_put(struct nf_bridge_info *nf_bridge)
1482{
1483 if (nf_bridge && atomic_dec_and_test(&nf_bridge->use))
1484 kfree(nf_bridge);
1485}
1486static inline void nf_bridge_get(struct nf_bridge_info *nf_bridge)
1487{
1488 if (nf_bridge)
1489 atomic_inc(&nf_bridge->use);
1490}
1491#endif /* CONFIG_BRIDGE_NETFILTER */
Patrick McHardya193a4a2006-03-20 19:23:05 -08001492static inline void nf_reset(struct sk_buff *skb)
1493{
1494 nf_conntrack_put(skb->nfct);
1495 skb->nfct = NULL;
1496#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
1497 nf_conntrack_put_reasm(skb->nfct_reasm);
1498 skb->nfct_reasm = NULL;
1499#endif
1500#ifdef CONFIG_BRIDGE_NETFILTER
1501 nf_bridge_put(skb->nf_bridge);
1502 skb->nf_bridge = NULL;
1503#endif
1504}
1505
Yasuyuki Kozakaiedda5532007-03-14 16:43:37 -07001506/* Note: This doesn't put any conntrack and bridge info in dst. */
1507static inline void __nf_copy(struct sk_buff *dst, const struct sk_buff *src)
1508{
1509 dst->nfct = src->nfct;
1510 nf_conntrack_get(src->nfct);
1511 dst->nfctinfo = src->nfctinfo;
1512#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
1513 dst->nfct_reasm = src->nfct_reasm;
1514 nf_conntrack_get_reasm(src->nfct_reasm);
1515#endif
1516#ifdef CONFIG_BRIDGE_NETFILTER
1517 dst->nf_bridge = src->nf_bridge;
1518 nf_bridge_get(src->nf_bridge);
1519#endif
1520}
1521
Yasuyuki Kozakaie7ac05f2007-03-14 16:44:01 -07001522static inline void nf_copy(struct sk_buff *dst, const struct sk_buff *src)
1523{
1524 nf_conntrack_put(dst->nfct);
1525#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
1526 nf_conntrack_put_reasm(dst->nfct_reasm);
1527#endif
1528#ifdef CONFIG_BRIDGE_NETFILTER
1529 nf_bridge_put(dst->nf_bridge);
1530#endif
1531 __nf_copy(dst, src);
1532}
1533
Linus Torvalds1da177e2005-04-16 15:20:36 -07001534#else /* CONFIG_NETFILTER */
1535static inline void nf_reset(struct sk_buff *skb) {}
Yasuyuki Kozakaiedda5532007-03-14 16:43:37 -07001536static inline void __nf_copy(struct sk_buff *dst, const struct sk_buff *src) {}
Yasuyuki Kozakaie7ac05f2007-03-14 16:44:01 -07001537static inline void nf_copy(struct sk_buff *dst, const struct sk_buff *src) {}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001538#endif /* CONFIG_NETFILTER */
1539
James Morris984bc162006-06-09 00:29:17 -07001540#ifdef CONFIG_NETWORK_SECMARK
1541static inline void skb_copy_secmark(struct sk_buff *to, const struct sk_buff *from)
1542{
1543 to->secmark = from->secmark;
1544}
1545
1546static inline void skb_init_secmark(struct sk_buff *skb)
1547{
1548 skb->secmark = 0;
1549}
1550#else
1551static inline void skb_copy_secmark(struct sk_buff *to, const struct sk_buff *from)
1552{ }
1553
1554static inline void skb_init_secmark(struct sk_buff *skb)
1555{ }
1556#endif
1557
Herbert Xu89114af2006-07-08 13:34:32 -07001558static inline int skb_is_gso(const struct sk_buff *skb)
1559{
1560 return skb_shinfo(skb)->gso_size;
1561}
1562
Linus Torvalds1da177e2005-04-16 15:20:36 -07001563#endif /* __KERNEL__ */
1564#endif /* _LINUX_SKBUFF_H */