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 1.19 2008/09/20 04:36:51 sephe Exp $
31 #ifndef _IP_DUMMYNET3_H_
32 #define _IP_DUMMYNET3_H_
34 #ifndef _IP_DUMMYNET_H
36 #define MODULE_DUMMYNET_ID 2
37 #define MODULE_DUMMYNET_NAME "dummynet"
41 //placeholder for kernel
44 enum ipfw_dummynet_opcodes {
50 * We start with a heap, which is used in the scheduler to decide when to
51 * transmit packets etc.
53 * The key for the heap is used for two different values:
55 * 1. Timer ticks- max 10K/second, so 32 bits are enough;
57 * 2. Virtual times. These increase in steps of len/x, where len is the
58 * packet length, and x is either the weight of the flow, or the sum
60 * If we limit to max 1000 flows and a max weight of 100, then x needs
61 * 17 bits. The packet size is 16 bits, so we can easily overflow if
62 * we do not allow errors.
64 * So we use a key "dn_key" which is 64 bits.
66 * MY_M is used as a shift count when doing fixed point arithmetic
67 * (a better name would be useful...).
69 typedef uint64_t dn_key; /* sorting key */
72 * Number of left shift to obtain a larger precision
74 * XXX With this scaling, max 1000 flows, max weight 100, 1Gbit/s, the
75 * virtual time wraps every 15 days.
82 * A heap entry is made of a key and a pointer to the actual object stored
85 * The heap is an array of dn_heap_entry entries, dynamically allocated.
86 * Current size is "size", with "elements" actually in use.
88 * The heap normally supports only ordered insert and extract from the top.
89 * If we want to extract an object from the middle of the heap, we have to
90 * know where the object itself is located in the heap (or we need to scan
91 * the whole array). To this purpose, an object has a field (int) which
92 * contains the index of the object itself into the heap. When the object
93 * is moved, the field must also be updated. The offset of the index in the
94 * object is stored in the 'offset' field in the heap descriptor. The
95 * assumption is that this offset is non-zero if we want to support extract
98 struct dn_heap_entry {
99 dn_key key; /* sorting key. Topmost element is smallest one */
100 void *object; /* object pointer */
106 int offset; /* XXX if > 0 this is the offset of direct ptr to obj */
107 struct dn_heap_entry *p; /* really an array of "size" entries */
111 uint16_t fid_type; /* ETHERTYPE_ */
123 #define fid_dst_ip fid_u.inet.dst_ip
124 #define fid_src_ip fid_u.inet.src_ip
125 #define fid_dst_port fid_u.inet.dst_port
126 #define fid_src_port fid_u.inet.src_port
127 #define fid_proto fid_u.inet.proto
128 #define fid_flags fid_u.inet.flags
131 typedef void (*ip_dn_unref_priv_t)(void *);
135 * struct dn_pkt identifies a packet in the dummynet queue, but is also used
136 * to tag packets passed back to the various destinations (ip_input(),
137 * ip_output() and so on).
139 * It is a tag (PACKET_TAG_DUMMYNET) associated with the actual mbuf.
143 TAILQ_ENTRY(dn_pkt) dn_next;
146 ip_dn_unref_priv_t dn_unref_priv;
148 uint32_t dn_flags; /* action when packet comes out. */
149 #define DN_FLAGS_IS_PIPE 0x10
150 #define DN_FLAGS_DIR_MASK 0x0f
151 #define DN_TO_IP_OUT 1
152 #define DN_TO_IP_IN 2
153 #define DN_TO_ETH_DEMUX 4
154 #define DN_TO_ETH_OUT 5
157 dn_key output_time; /* when the pkt is due for delivery */
158 struct ifnet *ifp; /* interface, for ip_output */
159 struct sockaddr_in *dn_dst;
160 struct route ro; /* route, for ip_output. MUST COPY */
161 int flags; /* flags, for ip_output (IPv6 ?) */
163 u_short pipe_nr; /* pipe/flow_set number */
166 struct dn_flow_id id; /* flow id */
167 int cpuid; /* target cpuid, for assertion */
168 struct lwkt_port *msgport; /* target msgport */
170 TAILQ_HEAD(dn_pkt_queue, dn_pkt);
173 * Overall structure of dummynet (with WF2Q+):
175 * In dummynet, packets are selected with the firewall rules, and passed to
176 * two different objects: PIPE or QUEUE.
178 * A QUEUE is just a queue with configurable size and queue management policy.
179 * It is also associated with a mask (to discriminate among different flows),
180 * a weight (used to give different shares of the bandwidth to different flows)
181 * and a "pipe", which essentially supplies the transmit clock for all queues
182 * associated with that pipe.
184 * A PIPE emulates a fixed-bandwidth link, whose bandwidth is configurable.
185 * The "clock" for a pipe comes from an internal timer. A pipe is also
186 * associated with one (or more, if masks are used) queue, where all packets
187 * for that pipe are stored.
189 * The bandwidth available on the pipe is shared by the queues associated with
190 * that pipe (only one in case the packet is sent to a PIPE) according to the
191 * WF2Q+ scheduling algorithm and the configured weights.
193 * In general, incoming packets are stored in the appropriate queue, which is
194 * then placed into one of a few heaps managed by a scheduler to decide when
195 * the packet should be extracted. The scheduler (a function called dummynet())
196 * is run at every timer tick, and grabs queues from the head of the heaps when
197 * they are ready for processing.
199 * There are three data structures definining a pipe and associated queues:
201 * + dn_pipe, which contains the main configuration parameters related to
202 * delay and bandwidth;
203 * + dn_flow_set, which contains WF2Q+ configuration, flow masks, plr and
205 * + dn_flow_queue, which is the per-flow queue (containing the packets)
207 * Multiple dn_flow_set can be linked to the same pipe, and multiple
208 * dn_flow_queue can be linked to the same dn_flow_set.
209 * All data structures are linked in a linear list which is used for
210 * housekeeping purposes.
212 * During configuration, we create and initialize the dn_flow_set and dn_pipe
213 * structures (a dn_pipe also contains a dn_flow_set).
215 * At runtime: packets are sent to the appropriate dn_flow_set (either WFQ
216 * ones, or the one embedded in the dn_pipe for fixed-rate flows), which in
217 * turn dispatches them to the appropriate dn_flow_queue (created dynamically
218 * according to the masks).
220 * The transmit clock for fixed rate flows (ready_event()) selects the
221 * dn_flow_queue to be used to transmit the next packet. For WF2Q,
222 * wfq_ready_event() extract a pipe which in turn selects the right flow using
223 * a number of heaps defined into the pipe itself.
227 * Per flow queue. This contains the flow identifier, the queue of packets,
228 * counters, and parameters used to support both RED and WF2Q+.
230 * A dn_flow_queue is created and initialized whenever a packet for a new
233 struct dn_flow_queue {
234 struct dn_flow_id id;
235 LIST_ENTRY(dn_flow_queue) q_link;
237 struct dn_pkt_queue queue; /* queue of packets */
240 u_long numbytes; /* credit for transmission (dynamic queues) */
242 uint64_t tot_pkts; /* statistics counters */
246 int hash_slot; /* debugging/diagnostic */
249 int avg; /* average queue length est. (scaled) */
250 int count; /* arrivals since last RED drop */
251 int random; /* random value (scaled) */
252 uint32_t q_time; /* start of queue idle time */
255 struct dn_flow_set *fs; /* parent flow set */
256 int heap_pos; /* position (index) of struct in heap */
257 dn_key sched_time; /* current time when queue enters ready_heap */
259 dn_key S, F; /* start time, finish time */
261 * Setting F < S means the timestamp is invalid. We only need
262 * to test this when the queue is empty.
265 LIST_HEAD(dn_flowqueue_head, dn_flow_queue);
268 * flow_set descriptor. Contains the "template" parameters for the queue
269 * configuration, and pointers to the hash table of dn_flow_queue's.
271 * The hash table is an array of lists -- we identify the slot by hashing
272 * the flow-id, then scan the list looking for a match.
273 * The size of the hash table (buckets) is configurable on a per-queue basis.
275 * A dn_flow_set is created whenever a new queue or pipe is created (in the
276 * latter case, the structure is located inside the struct dn_pipe).
279 u_short fs_nr; /* flow_set number */
280 u_short flags_fs; /* see 'Flow set flags' */
282 LIST_ENTRY(dn_flow_set) fs_link;
284 struct dn_pipe *pipe; /* pointer to parent pipe */
285 u_short parent_nr; /* parent pipe#, 0 if local to a pipe */
287 int weight; /* WFQ queue weight */
288 int qsize; /* queue size in slots or bytes */
289 int plr; /* pkt loss rate (2^31-1 means 100%) */
291 struct dn_flow_id flow_mask;
293 /* hash table of queues onto this flow_set */
294 int rq_size; /* number of slots */
295 int rq_elements; /* active elements */
296 struct dn_flowqueue_head *rq;/* array of rq_size entries */
298 uint32_t last_expired; /* do not expire too frequently */
299 int backlogged; /* #active queues for this flowset */
302 int w_q; /* queue weight (scaled) */
303 int max_th; /* maximum threshold for queue (scaled) */
304 int min_th; /* minimum threshold for queue (scaled) */
305 int max_p; /* maximum value for p_b (scaled) */
306 u_int c_1; /* max_p/(max_th-min_th) (scaled) */
307 u_int c_2; /* max_p*min_th/(max_th-min_th) (scaled) */
308 u_int c_3; /* for GRED, (1-max_p)/max_th (scaled) */
309 u_int c_4; /* for GRED, 1 - 2*max_p (scaled) */
310 u_int *w_q_lookup; /* lookup table for computing (1-w_q)^t */
311 u_int lookup_depth; /* depth of lookup table */
312 int lookup_step; /* granularity inside the lookup table */
313 int lookup_weight; /* equal to (1-w_q)^t / (1-w_q)^(t+1) */
314 int avg_pkt_size; /* medium packet size */
315 int max_pkt_size; /* max packet size */
317 LIST_HEAD(dn_flowset_head, dn_flow_set);
320 * Pipe descriptor. Contains global parameters, delay-line queue, and the
321 * flow_set used for fixed-rate queues.
323 * For WF2Q+ support it also has 3 heaps holding dn_flow_queue:
324 * + not_eligible_heap, for queues whose start time is higher than the
325 * virtual time. Sorted by start time.
326 * + scheduler_heap, for queues eligible for scheduling. Sorted by finish
328 * + idle_heap, all flows that are idle and can be removed. We do that on
329 * each tick so we do not slow down too much operations during forwarding.
331 struct dn_pipe { /* a pipe */
332 int pipe_nr; /* number */
333 int bandwidth; /* really, bytes/tick. */
334 int delay; /* really, ticks */
336 struct dn_pkt_queue p_queue;/* packets in delay line */
337 LIST_ENTRY(dn_pipe) p_link;
340 struct dn_heap scheduler_heap; /* top extract - key Finish time*/
341 struct dn_heap not_eligible_heap; /* top extract- key Start time */
342 struct dn_heap idle_heap; /* random extract - key Start=Finish time */
344 dn_key V; /* virtual time */
345 int sum; /* sum of weights of all active sessions */
346 int numbytes; /* bits I can transmit (more or less). */
348 dn_key sched_time; /* time pipe was scheduled in ready_heap */
350 struct dn_flow_set fs; /* used with fixed-rate flows */
352 LIST_HEAD(dn_pipe_head, dn_pipe);
357 size_t dn_sopt_arglen;
360 typedef int ip_dn_ctl_t(struct dn_sopt *);
361 typedef int ip_dn_io_t(struct mbuf *);
363 extern ip_dn_ctl_t *ip_dn_ctl_ptr;
364 extern ip_dn_io_t *ip_dn_io_ptr;
366 void ip_dn_queue(struct mbuf *);
367 void ip_dn_packet_free(struct dn_pkt *);
368 void ip_dn_packet_redispatch(struct dn_pkt *);
369 int ip_dn_sockopt(struct sockopt *);
371 #define DUMMYNET_LOADED (ip_dn_io_ptr != NULL)
375 struct dn_ioc_flowid {
376 uint16_t type; /* ETHERTYPE_ */
391 struct dn_ioc_flowqueue {
399 int hash_slot; /* debugging/diagnostic */
400 dn_key S; /* virtual start time */
401 dn_key F; /* virtual finish time */
403 struct dn_ioc_flowid id;
404 uint8_t reserved[16];
407 struct dn_ioc_flowset {
408 u_short fs_type; /* DN_IS_{QUEUE,PIPE}, MUST be first */
410 u_short fs_nr; /* flow_set number */
411 u_short flags_fs; /* see 'Flow set flags' */
412 u_short parent_nr; /* parent pipe#, 0 if local to a pipe */
414 int weight; /* WFQ queue weight */
415 int qsize; /* queue size in slots or bytes */
416 int plr; /* pkt loss rate (2^31-1 means 100%) */
418 /* Hash table information */
419 int rq_size; /* number of slots */
420 int rq_elements; /* active elements */
423 int w_q; /* queue weight (scaled) */
424 int max_th; /* maximum threshold for queue (scaled) */
425 int min_th; /* minimum threshold for queue (scaled) */
426 int max_p; /* maximum value for p_b (scaled) */
427 int lookup_step; /* granularity inside the lookup table */
428 int lookup_weight; /* equal to (1-w_q)^t / (1-w_q)^(t+1) */
430 struct dn_ioc_flowid flow_mask;
431 uint8_t reserved[16];
435 struct dn_ioc_flowset fs; /* MUST be first */
437 int pipe_nr; /* pipe number */
438 int bandwidth; /* bit/second */
439 int delay; /* milliseconds */
441 dn_key V; /* virtual time */
443 uint8_t reserved[16];
449 #define DN_HAVE_FLOW_MASK 0x0001
450 #define DN_IS_RED 0x0002
451 #define DN_IS_GENTLE_RED 0x0004
452 #define DN_QSIZE_IS_BYTES 0x0008 /* queue size is measured in bytes */
453 #define DN_NOERROR 0x0010 /* do not report ENOBUFS on drops */
454 #define DN_IS_PIPE 0x4000
455 #define DN_IS_QUEUE 0x8000
461 #define SCALE(x) ((x) << SCALE_RED)
462 #define SCALE_VAL(x) ((x) >> SCALE_RED)
463 #define SCALE_MUL(x, y) (((x) * (y)) >> SCALE_RED)
466 * Maximum pipe number
468 #define DN_PIPE_NR_MAX 65536
471 #endif /* !_IP_DUMMYNET_H */