2 * Copyright (c) 1998-2002 Luigi Rizzo, Universita` di Pisa
3 * Portions Copyright (c) 2000 Akamba Corp.
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
15 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * $FreeBSD: src/sys/netinet/ip_dummynet.h,v 1.10.2.9 2003/05/13 09:31:06 maxim Exp $
28 * $DragonFly: src/sys/net/dummynet/ip_dummynet.h,v 1.11 2007/11/03 13:14:29 sephe Exp $
31 #ifndef _IP_DUMMYNET_H
32 #define _IP_DUMMYNET_H
35 * Definition of dummynet data structures. In the structures, I decided
36 * not to use the macros in <sys/queue.h> in the hope of making the code
37 * easier to port to other architectures. The type of lists and queue we
38 * use here is pretty simple anyways.
42 * We start with a heap, which is used in the scheduler to decide when
43 * to transmit packets etc.
45 * The key for the heap is used for two different values:
47 * 1. timer ticks- max 10K/second, so 32 bits are enough;
49 * 2. virtual times. These increase in steps of len/x, where len is the
50 * packet length, and x is either the weight of the flow, or the
52 * If we limit to max 1000 flows and a max weight of 100, then
53 * x needs 17 bits. The packet size is 16 bits, so we can easily
54 * overflow if we do not allow errors.
55 * So we use a key "dn_key" which is 64 bits. Some macros are used to
56 * compare key values and handle wraparounds.
57 * MAX64 returns the largest of two key values.
58 * MY_M is used as a shift count when doing fixed point arithmetic
59 * (a better name would be useful...).
61 typedef uint64_t dn_key; /* sorting key */
62 #define DN_KEY_LT(a,b) ((int64_t)((a)-(b)) < 0)
63 #define DN_KEY_LEQ(a,b) ((int64_t)((a)-(b)) <= 0)
64 #define DN_KEY_GT(a,b) ((int64_t)((a)-(b)) > 0)
65 #define DN_KEY_GEQ(a,b) ((int64_t)((a)-(b)) >= 0)
66 #define MAX64(x,y) (((int64_t)((y) - (x))) > 0) ? (y) : (x)
67 #define MY_M 16 /* number of left shift to obtain a larger precision */
70 * XXX With this scaling, max 1000 flows, max weight 100, 1Gbit/s, the
71 * virtual time wraps every 15 days.
75 * The maximum/minimum hash table size for queues.
76 * These values must be a power of 2.
78 #define DN_MIN_HASH_SIZE 4
79 #define DN_MAX_HASH_SIZE 65536
82 * A heap entry is made of a key and a pointer to the actual
83 * object stored in the heap.
84 * The heap is an array of dn_heap_entry entries, dynamically allocated.
85 * Current size is "size", with "elements" actually in use.
86 * The heap normally supports only ordered insert and extract from the top.
87 * If we want to extract an object from the middle of the heap, we
88 * have to know where the object itself is located in the heap (or we
89 * need to scan the whole array). To this purpose, an object has a
90 * field (int) which contains the index of the object itself into the
91 * heap. When the object is moved, the field must also be updated.
92 * The offset of the index in the object is stored in the 'offset'
93 * field in the heap descriptor. The assumption is that this offset
94 * is non-zero if we want to support extract from the middle.
96 struct dn_heap_entry {
97 dn_key key; /* sorting key. Topmost element is smallest one */
98 void *object; /* object pointer */
104 int offset; /* XXX if > 0 this is the offset of direct ptr to obj */
105 struct dn_heap_entry *p; /* really an array of "size" entries */
111 * struct dn_pkt identifies a packet in the dummynet queue, but
112 * is also used to tag packets passed back to the various destinations
113 * (ip_input(), ip_output() and so on).
114 * It is a tag(PACKET_TAG_DUMMYNET) associated with the actual mbuf.
118 struct dn_pkt *dn_next;
120 struct ip_fw *rule; /* matching rule */
121 int dn_dir; /* action when packet comes out. */
122 #define DN_TO_IP_OUT 1
123 #define DN_TO_IP_IN 2
124 #define DN_TO_ETH_DEMUX 4
125 #define DN_TO_ETH_OUT 5
127 dn_key output_time; /* when the pkt is due for delivery */
128 struct ifnet *ifp; /* interface, for ip_output */
129 struct sockaddr_in *dn_dst;
130 struct route ro; /* route, for ip_output. MUST COPY */
131 int flags; /* flags, for ip_output (IPv6 ?) */
135 * Overall structure of dummynet (with WF2Q+):
137 * In dummynet, packets are selected with the firewall rules, and passed
138 * to two different objects: PIPE or QUEUE.
140 * A QUEUE is just a queue with configurable size and queue management
141 * policy. It is also associated with a mask (to discriminate among
142 * different flows), a weight (used to give different shares of the
143 * bandwidth to different flows) and a "pipe", which essentially
144 * supplies the transmit clock for all queues associated with that
147 * A PIPE emulates a fixed-bandwidth link, whose bandwidth is
148 * configurable. The "clock" for a pipe can come from either an
149 * internal timer, or from the transmit interrupt of an interface.
150 * A pipe is also associated with one (or more, if masks are used)
151 * queue, where all packets for that pipe are stored.
153 * The bandwidth available on the pipe is shared by the queues
154 * associated with that pipe (only one in case the packet is sent
155 * to a PIPE) according to the WF2Q+ scheduling algorithm and the
156 * configured weights.
158 * In general, incoming packets are stored in the appropriate queue,
159 * which is then placed into one of a few heaps managed by a scheduler
160 * to decide when the packet should be extracted.
161 * The scheduler (a function called dummynet()) is run at every timer
162 * tick, and grabs queues from the head of the heaps when they are
163 * ready for processing.
165 * There are three data structures definining a pipe and associated queues:
167 * + dn_pipe, which contains the main configuration parameters related
168 * to delay and bandwidth;
169 * + dn_flow_set, which contains WF2Q+ configuration, flow
170 * masks, plr and RED configuration;
171 * + dn_flow_queue, which is the per-flow queue (containing the packets)
173 * Multiple dn_flow_set can be linked to the same pipe, and multiple
174 * dn_flow_queue can be linked to the same dn_flow_set.
175 * All data structures are linked in a linear list which is used for
176 * housekeeping purposes.
178 * During configuration, we create and initialize the dn_flow_set
179 * and dn_pipe structures (a dn_pipe also contains a dn_flow_set).
181 * At runtime: packets are sent to the appropriate dn_flow_set (either
182 * WFQ ones, or the one embedded in the dn_pipe for fixed-rate flows),
183 * which in turn dispatches them to the appropriate dn_flow_queue
184 * (created dynamically according to the masks).
186 * The transmit clock for fixed rate flows (ready_event()) selects the
187 * dn_flow_queue to be used to transmit the next packet. For WF2Q,
188 * wfq_ready_event() extract a pipe which in turn selects the right
189 * flow using a number of heaps defined into the pipe itself.
193 * per flow queue. This contains the flow identifier, the queue
194 * of packets, counters, and parameters used to support both RED and
197 * A dn_flow_queue is created and initialized whenever a packet for
198 * a new flow arrives.
200 struct dn_flow_queue {
201 struct dn_flow_queue *next;
202 struct ipfw_flow_id id;
204 struct dn_pkt *head, *tail; /* queue of packets */
207 u_long numbytes; /* credit for transmission (dynamic queues) */
209 uint64_t tot_pkts; /* statistics counters */
213 int hash_slot; /* debugging/diagnostic */
216 int avg; /* average queue length est. (scaled) */
217 int count; /* arrivals since last RED drop */
218 int random; /* random value (scaled) */
219 uint32_t q_time; /* start of queue idle time */
222 struct dn_flow_set *fs; /* parent flow set */
223 int heap_pos; /* position (index) of struct in heap */
224 dn_key sched_time; /* current time when queue enters ready_heap */
226 dn_key S, F; /* start time, finish time */
228 * Setting F < S means the timestamp is invalid. We only need
229 * to test this when the queue is empty.
234 * flow_set descriptor. Contains the "template" parameters for the
235 * queue configuration, and pointers to the hash table of dn_flow_queue's.
237 * The hash table is an array of lists -- we identify the slot by
238 * hashing the flow-id, then scan the list looking for a match.
239 * The size of the hash table (buckets) is configurable on a per-queue
242 * A dn_flow_set is created whenever a new queue or pipe is created (in the
243 * latter case, the structure is located inside the struct dn_pipe).
246 struct dn_flow_set *next; /* next flow set in all_flow_sets list */
248 u_short fs_nr; /* flow_set number */
249 u_short flags_fs; /* see 'Flow set flags' */
251 struct dn_pipe *pipe; /* pointer to parent pipe */
252 u_short parent_nr; /* parent pipe#, 0 if local to a pipe */
254 int weight; /* WFQ queue weight */
255 int qsize; /* queue size in slots or bytes */
256 int plr; /* pkt loss rate (2^31-1 means 100%) */
258 struct ipfw_flow_id flow_mask;
260 /* hash table of queues onto this flow_set */
261 int rq_size; /* number of slots */
262 int rq_elements; /* active elements */
263 struct dn_flow_queue **rq; /* array of rq_size entries */
265 uint32_t last_expired; /* do not expire too frequently */
266 int backlogged; /* #active queues for this flowset */
269 int w_q; /* queue weight (scaled) */
270 int max_th; /* maximum threshold for queue (scaled) */
271 int min_th; /* minimum threshold for queue (scaled) */
272 int max_p; /* maximum value for p_b (scaled) */
273 u_int c_1; /* max_p/(max_th-min_th) (scaled) */
274 u_int c_2; /* max_p*min_th/(max_th-min_th) (scaled) */
275 u_int c_3; /* for GRED, (1-max_p)/max_th (scaled) */
276 u_int c_4; /* for GRED, 1 - 2*max_p (scaled) */
277 u_int *w_q_lookup; /* lookup table for computing (1-w_q)^t */
278 u_int lookup_depth; /* depth of lookup table */
279 int lookup_step; /* granularity inside the lookup table */
280 int lookup_weight; /* equal to (1-w_q)^t / (1-w_q)^(t+1) */
281 int avg_pkt_size; /* medium packet size */
282 int max_pkt_size; /* max packet size */
286 * Pipe descriptor. Contains global parameters, delay-line queue,
287 * and the flow_set used for fixed-rate queues.
289 * For WF2Q+ support it also has 3 heaps holding dn_flow_queue:
290 * not_eligible_heap, for queues whose start time is higher
291 * than the virtual time. Sorted by start time.
292 * scheduler_heap, for queues eligible for scheduling. Sorted by
294 * idle_heap, all flows that are idle and can be removed. We
295 * do that on each tick so we do not slow down too much
296 * operations during forwarding.
299 struct dn_pipe { /* a pipe */
300 struct dn_pipe *next;
302 int pipe_nr; /* number */
303 int bandwidth; /* really, bytes/tick. */
304 int delay; /* really, ticks */
306 struct dn_pkt *head, *tail; /* packets in delay line */
309 struct dn_heap scheduler_heap; /* top extract - key Finish time*/
310 struct dn_heap not_eligible_heap; /* top extract- key Start time */
311 struct dn_heap idle_heap; /* random extract - key Start=Finish time */
313 dn_key V; /* virtual time */
314 int sum; /* sum of weights of all active sessions */
315 int numbytes; /* bits I can transmit (more or less). */
317 dn_key sched_time; /* time pipe was scheduled in ready_heap */
319 struct dn_flow_set fs; /* used with fixed-rate flows */
322 typedef int ip_dn_ctl_t(struct sockopt *); /* raw_ip.c */
323 typedef void ip_dn_ruledel_t(void *); /* ip_fw2.c */
324 typedef int ip_dn_io_t(struct mbuf *m, int pipe_nr, int dir,
325 struct ip_fw_args *fwa);
326 extern ip_dn_ctl_t *ip_dn_ctl_ptr;
327 extern ip_dn_ruledel_t *ip_dn_ruledel_ptr;
328 extern ip_dn_io_t *ip_dn_io_ptr;
329 #define DUMMYNET_LOADED (ip_dn_io_ptr != NULL)
333 struct dn_ioc_flowid {
334 uint16_t type; /* ETHERTYPE_ */
349 struct dn_ioc_flowqueue {
357 int hash_slot; /* debugging/diagnostic */
358 dn_key S; /* virtual start time */
359 dn_key F; /* virtual finish time */
361 struct dn_ioc_flowid id;
362 uint8_t reserved[16];
365 struct dn_ioc_flowset {
366 u_short fs_type; /* DN_IS_{QUEUE,PIPE}, MUST be first */
368 u_short fs_nr; /* flow_set number */
369 u_short flags_fs; /* see 'Flow set flags' */
370 u_short parent_nr; /* parent pipe#, 0 if local to a pipe */
372 int weight; /* WFQ queue weight */
373 int qsize; /* queue size in slots or bytes */
374 int plr; /* pkt loss rate (2^31-1 means 100%) */
376 /* Hash table information */
377 int rq_size; /* number of slots */
378 int rq_elements; /* active elements */
381 int w_q; /* queue weight (scaled) */
382 int max_th; /* maximum threshold for queue (scaled) */
383 int min_th; /* minimum threshold for queue (scaled) */
384 int max_p; /* maximum value for p_b (scaled) */
385 int lookup_step; /* granularity inside the lookup table */
386 int lookup_weight; /* equal to (1-w_q)^t / (1-w_q)^(t+1) */
388 struct dn_ioc_flowid flow_mask;
389 uint8_t reserved[16];
393 struct dn_ioc_flowset fs; /* MUST be first */
395 int pipe_nr; /* pipe number */
396 int bandwidth; /* bit/second */
397 int delay; /* milliseconds */
399 dn_key V; /* virtual time */
401 uint8_t reserved[16];
407 #define DN_HAVE_FLOW_MASK 0x0001
408 #define DN_IS_RED 0x0002
409 #define DN_IS_GENTLE_RED 0x0004
410 #define DN_QSIZE_IS_BYTES 0x0008 /* queue size is measured in bytes */
411 #define DN_NOERROR 0x0010 /* do not report ENOBUFS on drops */
412 #define DN_IS_PIPE 0x4000
413 #define DN_IS_QUEUE 0x8000
419 #define SCALE(x) ((x) << SCALE_RED)
420 #define SCALE_VAL(x) ((x) >> SCALE_RED)
421 #define SCALE_MUL(x, y) (((x) * (y)) >> SCALE_RED)
423 #endif /* !_IP_DUMMYNET_H */