| Andy Green | 08d1f21 | 2007-07-10 19:29:37 +0200 | [diff] [blame] | 1 | How to use radiotap headers | 
|  | 2 | =========================== | 
|  | 3 |  | 
|  | 4 | Pointer to the radiotap include file | 
|  | 5 | ------------------------------------ | 
|  | 6 |  | 
|  | 7 | Radiotap headers are variable-length and extensible, you can get most of the | 
|  | 8 | information you need to know on them from: | 
|  | 9 |  | 
|  | 10 | ./include/net/ieee80211_radiotap.h | 
|  | 11 |  | 
|  | 12 | This document gives an overview and warns on some corner cases. | 
|  | 13 |  | 
|  | 14 |  | 
|  | 15 | Structure of the header | 
|  | 16 | ----------------------- | 
|  | 17 |  | 
|  | 18 | There is a fixed portion at the start which contains a u32 bitmap that defines | 
|  | 19 | if the possible argument associated with that bit is present or not.  So if b0 | 
|  | 20 | of the it_present member of ieee80211_radiotap_header is set, it means that | 
|  | 21 | the header for argument index 0 (IEEE80211_RADIOTAP_TSFT) is present in the | 
|  | 22 | argument area. | 
|  | 23 |  | 
|  | 24 | < 8-byte ieee80211_radiotap_header > | 
|  | 25 | [ <possible argument bitmap extensions ... > ] | 
|  | 26 | [ <argument> ... ] | 
|  | 27 |  | 
|  | 28 | At the moment there are only 13 possible argument indexes defined, but in case | 
|  | 29 | we run out of space in the u32 it_present member, it is defined that b31 set | 
|  | 30 | indicates that there is another u32 bitmap following (shown as "possible | 
|  | 31 | argument bitmap extensions..." above), and the start of the arguments is moved | 
|  | 32 | forward 4 bytes each time. | 
|  | 33 |  | 
|  | 34 | Note also that the it_len member __le16 is set to the total number of bytes | 
|  | 35 | covered by the ieee80211_radiotap_header and any arguments following. | 
|  | 36 |  | 
|  | 37 |  | 
|  | 38 | Requirements for arguments | 
|  | 39 | -------------------------- | 
|  | 40 |  | 
|  | 41 | After the fixed part of the header, the arguments follow for each argument | 
|  | 42 | index whose matching bit is set in the it_present member of | 
|  | 43 | ieee80211_radiotap_header. | 
|  | 44 |  | 
|  | 45 | - the arguments are all stored little-endian! | 
|  | 46 |  | 
|  | 47 | - the argument payload for a given argument index has a fixed size.  So | 
|  | 48 | IEEE80211_RADIOTAP_TSFT being present always indicates an 8-byte argument is | 
|  | 49 | present.  See the comments in ./include/net/ieee80211_radiotap.h for a nice | 
|  | 50 | breakdown of all the argument sizes | 
|  | 51 |  | 
|  | 52 | - the arguments must be aligned to a boundary of the argument size using | 
|  | 53 | padding.  So a u16 argument must start on the next u16 boundary if it isn't | 
|  | 54 | already on one, a u32 must start on the next u32 boundary and so on. | 
|  | 55 |  | 
|  | 56 | - "alignment" is relative to the start of the ieee80211_radiotap_header, ie, | 
|  | 57 | the first byte of the radiotap header.  The absolute alignment of that first | 
|  | 58 | byte isn't defined.  So even if the whole radiotap header is starting at, eg, | 
|  | 59 | address 0x00000003, still the first byte of the radiotap header is treated as | 
|  | 60 | 0 for alignment purposes. | 
|  | 61 |  | 
|  | 62 | - the above point that there may be no absolute alignment for multibyte | 
|  | 63 | entities in the fixed radiotap header or the argument region means that you | 
|  | 64 | have to take special evasive action when trying to access these multibyte | 
|  | 65 | entities.  Some arches like Blackfin cannot deal with an attempt to | 
|  | 66 | dereference, eg, a u16 pointer that is pointing to an odd address.  Instead | 
|  | 67 | you have to use a kernel API get_unaligned() to dereference the pointer, | 
|  | 68 | which will do it bytewise on the arches that require that. | 
|  | 69 |  | 
|  | 70 | - The arguments for a given argument index can be a compound of multiple types | 
|  | 71 | together.  For example IEEE80211_RADIOTAP_CHANNEL has an argument payload | 
|  | 72 | consisting of two u16s of total length 4.  When this happens, the padding | 
|  | 73 | rule is applied dealing with a u16, NOT dealing with a 4-byte single entity. | 
|  | 74 |  | 
|  | 75 |  | 
|  | 76 | Example valid radiotap header | 
|  | 77 | ----------------------------- | 
|  | 78 |  | 
|  | 79 | 0x00, 0x00, // <-- radiotap version + pad byte | 
|  | 80 | 0x0b, 0x00, // <- radiotap header length | 
|  | 81 | 0x04, 0x0c, 0x00, 0x00, // <-- bitmap | 
|  | 82 | 0x6c, // <-- rate (in 500kHz units) | 
|  | 83 | 0x0c, //<-- tx power | 
|  | 84 | 0x01 //<-- antenna | 
|  | 85 |  | 
|  | 86 |  | 
| Andy Green | 179f831 | 2007-07-10 19:29:38 +0200 | [diff] [blame] | 87 | Using the Radiotap Parser | 
|  | 88 | ------------------------- | 
|  | 89 |  | 
|  | 90 | If you are having to parse a radiotap struct, you can radically simplify the | 
|  | 91 | job by using the radiotap parser that lives in net/wireless/radiotap.c and has | 
|  | 92 | its prototypes available in include/net/cfg80211.h.  You use it like this: | 
|  | 93 |  | 
|  | 94 | #include <net/cfg80211.h> | 
|  | 95 |  | 
|  | 96 | /* buf points to the start of the radiotap header part */ | 
|  | 97 |  | 
|  | 98 | int MyFunction(u8 * buf, int buflen) | 
|  | 99 | { | 
|  | 100 | int pkt_rate_100kHz = 0, antenna = 0, pwr = 0; | 
|  | 101 | struct ieee80211_radiotap_iterator iterator; | 
|  | 102 | int ret = ieee80211_radiotap_iterator_init(&iterator, buf, buflen); | 
|  | 103 |  | 
|  | 104 | while (!ret) { | 
|  | 105 |  | 
|  | 106 | ret = ieee80211_radiotap_iterator_next(&iterator); | 
|  | 107 |  | 
|  | 108 | if (ret) | 
|  | 109 | continue; | 
|  | 110 |  | 
|  | 111 | /* see if this argument is something we can use */ | 
|  | 112 |  | 
|  | 113 | switch (iterator.this_arg_index) { | 
|  | 114 | /* | 
|  | 115 | * You must take care when dereferencing iterator.this_arg | 
|  | 116 | * for multibyte types... the pointer is not aligned.  Use | 
|  | 117 | * get_unaligned((type *)iterator.this_arg) to dereference | 
|  | 118 | * iterator.this_arg for type "type" safely on all arches. | 
|  | 119 | */ | 
|  | 120 | case IEEE80211_RADIOTAP_RATE: | 
|  | 121 | /* radiotap "rate" u8 is in | 
|  | 122 | * 500kbps units, eg, 0x02=1Mbps | 
|  | 123 | */ | 
|  | 124 | pkt_rate_100kHz = (*iterator.this_arg) * 5; | 
|  | 125 | break; | 
|  | 126 |  | 
|  | 127 | case IEEE80211_RADIOTAP_ANTENNA: | 
|  | 128 | /* radiotap uses 0 for 1st ant */ | 
|  | 129 | antenna = *iterator.this_arg); | 
|  | 130 | break; | 
|  | 131 |  | 
|  | 132 | case IEEE80211_RADIOTAP_DBM_TX_POWER: | 
|  | 133 | pwr = *iterator.this_arg; | 
|  | 134 | break; | 
|  | 135 |  | 
|  | 136 | default: | 
|  | 137 | break; | 
|  | 138 | } | 
|  | 139 | }  /* while more rt headers */ | 
|  | 140 |  | 
|  | 141 | if (ret != -ENOENT) | 
|  | 142 | return TXRX_DROP; | 
|  | 143 |  | 
|  | 144 | /* discard the radiotap header part */ | 
|  | 145 | buf += iterator.max_length; | 
|  | 146 | buflen -= iterator.max_length; | 
|  | 147 |  | 
|  | 148 | ... | 
|  | 149 |  | 
|  | 150 | } | 
|  | 151 |  | 
| Andy Green | 08d1f21 | 2007-07-10 19:29:37 +0200 | [diff] [blame] | 152 | Andy Green <andy@warmcat.com> |