1 /* 2 * Copyright (C) 2011 Matteo Landi, Luigi Rizzo. All rights reserved. 3 * 4 * Redistribution and use in source and binary forms, with or without 5 * modification, are permitted provided that the following conditions are 6 * met: 7 * 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 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 14 * distribution. 15 * 16 * 3. Neither the name of the authors nor the names of their contributors 17 * may be used to endorse or promote products derived from this 18 * software without specific prior written permission. 19 * 20 * THIS SOFTWARE IS PROVIDED BY MATTEO LANDI AND CONTRIBUTORS "AS IS" AND 21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL MATTEO LANDI OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF 30 * THE POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * $FreeBSD$ 35 * $Id: netmap.h 10601 2012-02-21 16:40:14Z luigi $ 36 * 37 * Definitions of constants and the structures used by the netmap 38 * framework, for the part visible to both kernel and userspace. 39 * Detailed info on netmap is available with "man netmap" or at 40 * 41 * http://info.iet.unipi.it/~luigi/netmap/ 42 */ 43 44 #ifndef _NET_NETMAP_H_ 45 #define _NET_NETMAP_H_ 46 47 /* 48 * --- Netmap data structures --- 49 * 50 * The data structures used by netmap are shown below. Those in 51 * capital letters are in an mmapp()ed area shared with userspace, 52 * while others are private to the kernel. 53 * Shared structures do not contain pointers but only memory 54 * offsets, so that addressing is portable between kernel and userspace. 55 56 57 softc 58 +----------------+ 59 | standard fields| 60 | if_pspare[0] ----------+ 61 +----------------+ | 62 | 63 +----------------+<------+ 64 |(netmap_adapter)| 65 | | netmap_kring 66 | tx_rings *--------------------------------->+---------------+ 67 | | netmap_kring | ring *---------. 68 | rx_rings *--------->+---------------+ | nr_hwcur | | 69 +----------------+ | ring *--------. | nr_hwavail | V 70 | nr_hwcur | | | selinfo | | 71 | nr_hwavail | | +---------------+ . 72 | selinfo | | | ... | . 73 +---------------+ | |(ntx+1 entries)| 74 | .... | | | | 75 |(nrx+1 entries)| | +---------------+ 76 | | | 77 KERNEL +---------------+ | 78 | 79 ==================================================================== 80 | 81 USERSPACE | NETMAP_RING 82 +---->+-------------+ 83 / | cur | 84 NETMAP_IF (nifp, one per file desc.) / | avail | 85 +---------------+ / | buf_ofs | 86 | ni_tx_rings | / +=============+ 87 | ni_rx_rings | / | buf_idx | slot[0] 88 | | / | len, flags | 89 | | / +-------------+ 90 +===============+ / | buf_idx | slot[1] 91 | txring_ofs[0] | (rel.to nifp)--' | len, flags | 92 | txring_ofs[1] | +-------------+ 93 (num_rings+1 entries) (nr_num_slots entries) 94 | txring_ofs[n] | | buf_idx | slot[n-1] 95 +---------------+ | len, flags | 96 | rxring_ofs[0] | +-------------+ 97 | rxring_ofs[1] | 98 (num_rings+1 entries) 99 | txring_ofs[n] | 100 +---------------+ 101 102 * The private descriptor ('softc' or 'adapter') of each interface 103 * is extended with a "struct netmap_adapter" containing netmap-related 104 * info (see description in dev/netmap/netmap_kernel.h. 105 * Among other things, tx_rings and rx_rings point to the arrays of 106 * "struct netmap_kring" which in turn reache the various 107 * "struct netmap_ring", shared with userspace. 108 109 * The NETMAP_RING is the userspace-visible replica of the NIC ring. 110 * Each slot has the index of a buffer, its length and some flags. 111 * In user space, the buffer address is computed as 112 * (char *)ring + buf_ofs + index*NETMAP_BUF_SIZE 113 * In the kernel, buffers do not necessarily need to be contiguous, 114 * and the virtual and physical addresses are derived through 115 * a lookup table. 116 * To associate a different buffer to a slot, applications must 117 * write the new index in buf_idx, and set NS_BUF_CHANGED flag to 118 * make sure that the kernel updates the hardware ring as needed. 119 * 120 * Normally the driver is not requested to report the result of 121 * transmissions (this can dramatically speed up operation). 122 * However the user may request to report completion by setting 123 * NS_REPORT. 124 */ 125 struct netmap_slot { 126 uint32_t buf_idx; /* buffer index */ 127 uint16_t len; /* packet length, to be copied to/from the hw ring */ 128 uint16_t flags; /* buf changed, etc. */ 129 #define NS_BUF_CHANGED 0x0001 /* must resync the map, buffer changed */ 130 #define NS_REPORT 0x0002 /* ask the hardware to report results 131 * e.g. by generating an interrupt 132 */ 133 }; 134 135 /* 136 * Netmap representation of a TX or RX ring (also known as "queue"). 137 * This is a queue implemented as a fixed-size circular array. 138 * At the software level, two fields are important: avail and cur. 139 * 140 * In TX rings: 141 * avail indicates the number of slots available for transmission. 142 * It is updated by the kernel after every netmap system call. 143 * It MUST BE decremented by the application when it appends a 144 * packet. 145 * cur indicates the slot to use for the next packet 146 * to send (i.e. the "tail" of the queue). 147 * It MUST BE incremented by the application before 148 * netmap system calls to reflect the number of newly 149 * sent packets. 150 * It is checked by the kernel on netmap system calls 151 * (normally unmodified by the kernel unless invalid). 152 * 153 * The kernel side of netmap uses two additional fields in its own 154 * private ring structure, netmap_kring: 155 * nr_hwcur is a copy of nr_cur on an NIOCTXSYNC. 156 * nr_hwavail is the number of slots known as available by the 157 * hardware. It is updated on an INTR (inc by the 158 * number of packets sent) and on a NIOCTXSYNC 159 * (decrease by nr_cur - nr_hwcur) 160 * A special case, nr_hwavail is -1 if the transmit 161 * side is idle (no pending transmits). 162 * 163 * In RX rings: 164 * avail is the number of packets available (possibly 0). 165 * It MUST BE decremented by the application when it consumes 166 * a packet, and it is updated to nr_hwavail on a NIOCRXSYNC 167 * cur indicates the first slot that contains a packet not 168 * processed yet (the "head" of the queue). 169 * It MUST BE incremented by the software when it consumes 170 * a packet. 171 * reserved indicates the number of buffers before 'cur' 172 * that the application has still in use. Normally 0, 173 * it MUST BE incremented by the application when it 174 * does not return the buffer immediately, and decremented 175 * when the buffer is finally freed. 176 * 177 * The kernel side of netmap uses two additional fields in the kring: 178 * nr_hwcur is a copy of nr_cur on an NIOCRXSYNC 179 * nr_hwavail is the number of packets available. It is updated 180 * on INTR (inc by the number of new packets arrived) 181 * and on NIOCRXSYNC (decreased by nr_cur - nr_hwcur). 182 * 183 * DATA OWNERSHIP/LOCKING: 184 * The netmap_ring is owned by the user program and it is only 185 * accessed or modified in the upper half of the kernel during 186 * a system call. 187 * 188 * The netmap_kring is only modified by the upper half of the kernel. 189 */ 190 struct netmap_ring { 191 /* 192 * nr_buf_base_ofs is meant to be used through macros. 193 * It contains the offset of the buffer region from this 194 * descriptor. 195 */ 196 const ssize_t buf_ofs; 197 const uint32_t num_slots; /* number of slots in the ring. */ 198 uint32_t avail; /* number of usable slots */ 199 uint32_t cur; /* 'current' r/w position */ 200 uint32_t reserved; /* not refilled before current */ 201 202 const uint16_t nr_buf_size; 203 uint16_t flags; 204 #define NR_TIMESTAMP 0x0002 /* set timestamp on *sync() */ 205 206 struct timeval ts; /* time of last *sync() */ 207 208 /* the slots follow. This struct has variable size */ 209 struct netmap_slot slot[0]; /* array of slots. */ 210 }; 211 212 213 /* 214 * Netmap representation of an interface and its queue(s). 215 * There is one netmap_if for each file descriptor on which we want 216 * to select/poll. We assume that on each interface has the same number 217 * of receive and transmit queues. 218 * select/poll operates on one or all pairs depending on the value of 219 * nmr_queueid passed on the ioctl. 220 */ 221 struct netmap_if { 222 char ni_name[IFNAMSIZ]; /* name of the interface. */ 223 const u_int ni_version; /* API version, currently unused */ 224 const u_int ni_rx_rings; /* number of rx rings */ 225 const u_int ni_tx_rings; /* if zero, same as ni_rx_rings */ 226 /* 227 * The following array contains the offset of each netmap ring 228 * from this structure. The first ni_tx_queues+1 entries refer 229 * to the tx rings, the next ni_rx_queues+1 refer to the rx rings 230 * (the last entry in each block refers to the host stack rings). 231 * The area is filled up by the kernel on NIOCREG, 232 * and then only read by userspace code. 233 */ 234 const ssize_t ring_ofs[0]; 235 }; 236 237 #ifndef NIOCREGIF 238 /* 239 * ioctl names and related fields 240 * 241 * NIOCGINFO takes a struct ifreq, the interface name is the input, 242 * the outputs are number of queues and number of descriptor 243 * for each queue (useful to set number of threads etc.). 244 * 245 * NIOCREGIF takes an interface name within a struct ifreq, 246 * and activates netmap mode on the interface (if possible). 247 * 248 * NIOCUNREGIF unregisters the interface associated to the fd. 249 * 250 * NIOCTXSYNC, NIOCRXSYNC synchronize tx or rx queues, 251 * whose identity is set in NIOCREGIF through nr_ringid 252 */ 253 254 /* 255 * struct nmreq overlays a struct ifreq 256 */ 257 struct nmreq { 258 char nr_name[IFNAMSIZ]; 259 uint32_t nr_version; /* API version */ 260 #define NETMAP_API 3 /* current version */ 261 uint32_t nr_offset; /* nifp offset in the shared region */ 262 uint32_t nr_memsize; /* size of the shared region */ 263 uint32_t nr_tx_slots; /* slots in tx rings */ 264 uint32_t nr_rx_slots; /* slots in rx rings */ 265 uint16_t nr_tx_rings; /* number of tx rings */ 266 uint16_t nr_rx_rings; /* number of rx rings */ 267 uint16_t nr_ringid; /* ring(s) we care about */ 268 #define NETMAP_HW_RING 0x4000 /* low bits indicate one hw ring */ 269 #define NETMAP_SW_RING 0x2000 /* process the sw ring */ 270 #define NETMAP_NO_TX_POLL 0x1000 /* no automatic txsync on poll */ 271 #define NETMAP_RING_MASK 0xfff /* the ring number */ 272 uint16_t spare1; 273 uint32_t spare2[4]; 274 }; 275 276 /* 277 * FreeBSD uses the size value embedded in the _IOWR to determine 278 * how much to copy in/out. So we need it to match the actual 279 * data structure we pass. We put some spares in the structure 280 * to ease compatibility with other versions 281 */ 282 #define NIOCGINFO _IOWR('i', 145, struct nmreq) /* return IF info */ 283 #define NIOCREGIF _IOWR('i', 146, struct nmreq) /* interface register */ 284 #define NIOCUNREGIF _IO('i', 147) /* interface unregister */ 285 #define NIOCTXSYNC _IO('i', 148) /* sync tx queues */ 286 #define NIOCRXSYNC _IO('i', 149) /* sync rx queues */ 287 #endif /* !NIOCREGIF */ 288 289 #endif /* _NET_NETMAP_H_ */ 290