1 ===================== 1 =====================
2 BPF Type Format (BTF) 2 BPF Type Format (BTF)
3 ===================== 3 =====================
4 4
5 1. Introduction 5 1. Introduction
6 =============== 6 ===============
7 7
8 BTF (BPF Type Format) is the metadata format w 8 BTF (BPF Type Format) is the metadata format which encodes the debug info
9 related to BPF program/map. The name BTF was u 9 related to BPF program/map. The name BTF was used initially to describe data
10 types. The BTF was later extended to include f 10 types. The BTF was later extended to include function info for defined
11 subroutines, and line info for source/line inf 11 subroutines, and line info for source/line information.
12 12
13 The debug info is used for map pretty print, f 13 The debug info is used for map pretty print, function signature, etc. The
14 function signature enables better bpf program/ 14 function signature enables better bpf program/function kernel symbol. The line
15 info helps generate source annotated translate 15 info helps generate source annotated translated byte code, jited code and
16 verifier log. 16 verifier log.
17 17
18 The BTF specification contains two parts, 18 The BTF specification contains two parts,
19 * BTF kernel API 19 * BTF kernel API
20 * BTF ELF file format 20 * BTF ELF file format
21 21
22 The kernel API is the contract between user sp 22 The kernel API is the contract between user space and kernel. The kernel
23 verifies the BTF info before using it. The ELF 23 verifies the BTF info before using it. The ELF file format is a user space
24 contract between ELF file and libbpf loader. 24 contract between ELF file and libbpf loader.
25 25
26 The type and string sections are part of the B 26 The type and string sections are part of the BTF kernel API, describing the
27 debug info (mostly types related) referenced b 27 debug info (mostly types related) referenced by the bpf program. These two
28 sections are discussed in details in :ref:`BTF 28 sections are discussed in details in :ref:`BTF_Type_String`.
29 29
30 .. _BTF_Type_String: 30 .. _BTF_Type_String:
31 31
32 2. BTF Type and String Encoding 32 2. BTF Type and String Encoding
33 =============================== 33 ===============================
34 34
35 The file ``include/uapi/linux/btf.h`` provides 35 The file ``include/uapi/linux/btf.h`` provides high-level definition of how
36 types/strings are encoded. 36 types/strings are encoded.
37 37
38 The beginning of data blob must be:: 38 The beginning of data blob must be::
39 39
40 struct btf_header { 40 struct btf_header {
41 __u16 magic; 41 __u16 magic;
42 __u8 version; 42 __u8 version;
43 __u8 flags; 43 __u8 flags;
44 __u32 hdr_len; 44 __u32 hdr_len;
45 45
46 /* All offsets are in bytes relative t 46 /* All offsets are in bytes relative to the end of this header */
47 __u32 type_off; /* offset of t 47 __u32 type_off; /* offset of type section */
48 __u32 type_len; /* length of t 48 __u32 type_len; /* length of type section */
49 __u32 str_off; /* offset of s 49 __u32 str_off; /* offset of string section */
50 __u32 str_len; /* length of s 50 __u32 str_len; /* length of string section */
51 }; 51 };
52 52
53 The magic is ``0xeB9F``, which has different e 53 The magic is ``0xeB9F``, which has different encoding for big and little
54 endian systems, and can be used to test whethe 54 endian systems, and can be used to test whether BTF is generated for big- or
55 little-endian target. The ``btf_header`` is de 55 little-endian target. The ``btf_header`` is designed to be extensible with
56 ``hdr_len`` equal to ``sizeof(struct btf_heade 56 ``hdr_len`` equal to ``sizeof(struct btf_header)`` when a data blob is
57 generated. 57 generated.
58 58
59 2.1 String Encoding 59 2.1 String Encoding
60 ------------------- 60 -------------------
61 61
62 The first string in the string section must be 62 The first string in the string section must be a null string. The rest of
63 string table is a concatenation of other null- 63 string table is a concatenation of other null-terminated strings.
64 64
65 2.2 Type Encoding 65 2.2 Type Encoding
66 ----------------- 66 -----------------
67 67
68 The type id ``0`` is reserved for ``void`` typ 68 The type id ``0`` is reserved for ``void`` type. The type section is parsed
69 sequentially and type id is assigned to each r 69 sequentially and type id is assigned to each recognized type starting from id
70 ``1``. Currently, the following types are supp 70 ``1``. Currently, the following types are supported::
71 71
72 #define BTF_KIND_INT 1 /* 72 #define BTF_KIND_INT 1 /* Integer */
73 #define BTF_KIND_PTR 2 /* 73 #define BTF_KIND_PTR 2 /* Pointer */
74 #define BTF_KIND_ARRAY 3 /* 74 #define BTF_KIND_ARRAY 3 /* Array */
75 #define BTF_KIND_STRUCT 4 /* 75 #define BTF_KIND_STRUCT 4 /* Struct */
76 #define BTF_KIND_UNION 5 /* 76 #define BTF_KIND_UNION 5 /* Union */
77 #define BTF_KIND_ENUM 6 /* 77 #define BTF_KIND_ENUM 6 /* Enumeration up to 32-bit values */
78 #define BTF_KIND_FWD 7 /* 78 #define BTF_KIND_FWD 7 /* Forward */
79 #define BTF_KIND_TYPEDEF 8 /* 79 #define BTF_KIND_TYPEDEF 8 /* Typedef */
80 #define BTF_KIND_VOLATILE 9 /* 80 #define BTF_KIND_VOLATILE 9 /* Volatile */
81 #define BTF_KIND_CONST 10 /* 81 #define BTF_KIND_CONST 10 /* Const */
82 #define BTF_KIND_RESTRICT 11 /* 82 #define BTF_KIND_RESTRICT 11 /* Restrict */
83 #define BTF_KIND_FUNC 12 /* 83 #define BTF_KIND_FUNC 12 /* Function */
84 #define BTF_KIND_FUNC_PROTO 13 /* 84 #define BTF_KIND_FUNC_PROTO 13 /* Function Proto */
85 #define BTF_KIND_VAR 14 /* 85 #define BTF_KIND_VAR 14 /* Variable */
86 #define BTF_KIND_DATASEC 15 /* 86 #define BTF_KIND_DATASEC 15 /* Section */
87 #define BTF_KIND_FLOAT 16 /* 87 #define BTF_KIND_FLOAT 16 /* Floating point */
88 #define BTF_KIND_DECL_TAG 17 /* 88 #define BTF_KIND_DECL_TAG 17 /* Decl Tag */
89 #define BTF_KIND_TYPE_TAG 18 /* 89 #define BTF_KIND_TYPE_TAG 18 /* Type Tag */
90 #define BTF_KIND_ENUM64 19 /* 90 #define BTF_KIND_ENUM64 19 /* Enumeration up to 64-bit values */
91 91
92 Note that the type section encodes debug info, 92 Note that the type section encodes debug info, not just pure types.
93 ``BTF_KIND_FUNC`` is not a type, and it repres 93 ``BTF_KIND_FUNC`` is not a type, and it represents a defined subprogram.
94 94
95 Each type contains the following common data:: 95 Each type contains the following common data::
96 96
97 struct btf_type { 97 struct btf_type {
98 __u32 name_off; 98 __u32 name_off;
99 /* "info" bits arrangement 99 /* "info" bits arrangement
100 * bits 0-15: vlen (e.g. # of struct' 100 * bits 0-15: vlen (e.g. # of struct's members)
101 * bits 16-23: unused 101 * bits 16-23: unused
102 * bits 24-28: kind (e.g. int, ptr, ar 102 * bits 24-28: kind (e.g. int, ptr, array...etc)
103 * bits 29-30: unused 103 * bits 29-30: unused
104 * bit 31: kind_flag, currently us 104 * bit 31: kind_flag, currently used by
105 * struct, union, fwd, enu 105 * struct, union, fwd, enum and enum64.
106 */ 106 */
107 __u32 info; 107 __u32 info;
108 /* "size" is used by INT, ENUM, STRUCT 108 /* "size" is used by INT, ENUM, STRUCT, UNION and ENUM64.
109 * "size" tells the size of the type i 109 * "size" tells the size of the type it is describing.
110 * 110 *
111 * "type" is used by PTR, TYPEDEF, VOL 111 * "type" is used by PTR, TYPEDEF, VOLATILE, CONST, RESTRICT,
112 * FUNC, FUNC_PROTO, DECL_TAG and TYPE 112 * FUNC, FUNC_PROTO, DECL_TAG and TYPE_TAG.
113 * "type" is a type_id referring to an 113 * "type" is a type_id referring to another type.
114 */ 114 */
115 union { 115 union {
116 __u32 size; 116 __u32 size;
117 __u32 type; 117 __u32 type;
118 }; 118 };
119 }; 119 };
120 120
121 For certain kinds, the common data are followe 121 For certain kinds, the common data are followed by kind-specific data. The
122 ``name_off`` in ``struct btf_type`` specifies 122 ``name_off`` in ``struct btf_type`` specifies the offset in the string table.
123 The following sections detail encoding of each 123 The following sections detail encoding of each kind.
124 124
125 2.2.1 BTF_KIND_INT 125 2.2.1 BTF_KIND_INT
126 ~~~~~~~~~~~~~~~~~~ 126 ~~~~~~~~~~~~~~~~~~
127 127
128 ``struct btf_type`` encoding requirement: 128 ``struct btf_type`` encoding requirement:
129 * ``name_off``: any valid offset 129 * ``name_off``: any valid offset
130 * ``info.kind_flag``: 0 130 * ``info.kind_flag``: 0
131 * ``info.kind``: BTF_KIND_INT 131 * ``info.kind``: BTF_KIND_INT
132 * ``info.vlen``: 0 132 * ``info.vlen``: 0
133 * ``size``: the size of the int type in bytes 133 * ``size``: the size of the int type in bytes.
134 134
135 ``btf_type`` is followed by a ``u32`` with the 135 ``btf_type`` is followed by a ``u32`` with the following bits arrangement::
136 136
137 #define BTF_INT_ENCODING(VAL) (((VAL) & 0x 137 #define BTF_INT_ENCODING(VAL) (((VAL) & 0x0f000000) >> 24)
138 #define BTF_INT_OFFSET(VAL) (((VAL) & 0x 138 #define BTF_INT_OFFSET(VAL) (((VAL) & 0x00ff0000) >> 16)
139 #define BTF_INT_BITS(VAL) ((VAL) & 0x 139 #define BTF_INT_BITS(VAL) ((VAL) & 0x000000ff)
140 140
141 The ``BTF_INT_ENCODING`` has the following att 141 The ``BTF_INT_ENCODING`` has the following attributes::
142 142
143 #define BTF_INT_SIGNED (1 << 0) 143 #define BTF_INT_SIGNED (1 << 0)
144 #define BTF_INT_CHAR (1 << 1) 144 #define BTF_INT_CHAR (1 << 1)
145 #define BTF_INT_BOOL (1 << 2) 145 #define BTF_INT_BOOL (1 << 2)
146 146
147 The ``BTF_INT_ENCODING()`` provides extra info 147 The ``BTF_INT_ENCODING()`` provides extra information: signedness, char, or
148 bool, for the int type. The char and bool enco 148 bool, for the int type. The char and bool encoding are mostly useful for
149 pretty print. At most one encoding can be spec 149 pretty print. At most one encoding can be specified for the int type.
150 150
151 The ``BTF_INT_BITS()`` specifies the number of 151 The ``BTF_INT_BITS()`` specifies the number of actual bits held by this int
152 type. For example, a 4-bit bitfield encodes `` 152 type. For example, a 4-bit bitfield encodes ``BTF_INT_BITS()`` equals to 4.
153 The ``btf_type.size * 8`` must be equal to or 153 The ``btf_type.size * 8`` must be equal to or greater than ``BTF_INT_BITS()``
154 for the type. The maximum value of ``BTF_INT_B 154 for the type. The maximum value of ``BTF_INT_BITS()`` is 128.
155 155
156 The ``BTF_INT_OFFSET()`` specifies the startin 156 The ``BTF_INT_OFFSET()`` specifies the starting bit offset to calculate values
157 for this int. For example, a bitfield struct m 157 for this int. For example, a bitfield struct member has:
158 158
159 * btf member bit offset 100 from the start of 159 * btf member bit offset 100 from the start of the structure,
160 * btf member pointing to an int type, 160 * btf member pointing to an int type,
161 * the int type has ``BTF_INT_OFFSET() = 2`` a 161 * the int type has ``BTF_INT_OFFSET() = 2`` and ``BTF_INT_BITS() = 4``
162 162
163 Then in the struct memory layout, this member 163 Then in the struct memory layout, this member will occupy ``4`` bits starting
164 from bits ``100 + 2 = 102``. 164 from bits ``100 + 2 = 102``.
165 165
166 Alternatively, the bitfield struct member can 166 Alternatively, the bitfield struct member can be the following to access the
167 same bits as the above: 167 same bits as the above:
168 168
169 * btf member bit offset 102, 169 * btf member bit offset 102,
170 * btf member pointing to an int type, 170 * btf member pointing to an int type,
171 * the int type has ``BTF_INT_OFFSET() = 0`` a 171 * the int type has ``BTF_INT_OFFSET() = 0`` and ``BTF_INT_BITS() = 4``
172 172
173 The original intention of ``BTF_INT_OFFSET()`` 173 The original intention of ``BTF_INT_OFFSET()`` is to provide flexibility of
174 bitfield encoding. Currently, both llvm and pa 174 bitfield encoding. Currently, both llvm and pahole generate
175 ``BTF_INT_OFFSET() = 0`` for all int types. 175 ``BTF_INT_OFFSET() = 0`` for all int types.
176 176
177 2.2.2 BTF_KIND_PTR 177 2.2.2 BTF_KIND_PTR
178 ~~~~~~~~~~~~~~~~~~ 178 ~~~~~~~~~~~~~~~~~~
179 179
180 ``struct btf_type`` encoding requirement: 180 ``struct btf_type`` encoding requirement:
181 * ``name_off``: 0 181 * ``name_off``: 0
182 * ``info.kind_flag``: 0 182 * ``info.kind_flag``: 0
183 * ``info.kind``: BTF_KIND_PTR 183 * ``info.kind``: BTF_KIND_PTR
184 * ``info.vlen``: 0 184 * ``info.vlen``: 0
185 * ``type``: the pointee type of the pointer 185 * ``type``: the pointee type of the pointer
186 186
187 No additional type data follow ``btf_type``. 187 No additional type data follow ``btf_type``.
188 188
189 2.2.3 BTF_KIND_ARRAY 189 2.2.3 BTF_KIND_ARRAY
190 ~~~~~~~~~~~~~~~~~~~~ 190 ~~~~~~~~~~~~~~~~~~~~
191 191
192 ``struct btf_type`` encoding requirement: 192 ``struct btf_type`` encoding requirement:
193 * ``name_off``: 0 193 * ``name_off``: 0
194 * ``info.kind_flag``: 0 194 * ``info.kind_flag``: 0
195 * ``info.kind``: BTF_KIND_ARRAY 195 * ``info.kind``: BTF_KIND_ARRAY
196 * ``info.vlen``: 0 196 * ``info.vlen``: 0
197 * ``size/type``: 0, not used 197 * ``size/type``: 0, not used
198 198
199 ``btf_type`` is followed by one ``struct btf_a 199 ``btf_type`` is followed by one ``struct btf_array``::
200 200
201 struct btf_array { 201 struct btf_array {
202 __u32 type; 202 __u32 type;
203 __u32 index_type; 203 __u32 index_type;
204 __u32 nelems; 204 __u32 nelems;
205 }; 205 };
206 206
207 The ``struct btf_array`` encoding: 207 The ``struct btf_array`` encoding:
208 * ``type``: the element type 208 * ``type``: the element type
209 * ``index_type``: the index type 209 * ``index_type``: the index type
210 * ``nelems``: the number of elements for thi 210 * ``nelems``: the number of elements for this array (``0`` is also allowed).
211 211
212 The ``index_type`` can be any regular int type 212 The ``index_type`` can be any regular int type (``u8``, ``u16``, ``u32``,
213 ``u64``, ``unsigned __int128``). The original 213 ``u64``, ``unsigned __int128``). The original design of including
214 ``index_type`` follows DWARF, which has an ``i 214 ``index_type`` follows DWARF, which has an ``index_type`` for its array type.
215 Currently in BTF, beyond type verification, th 215 Currently in BTF, beyond type verification, the ``index_type`` is not used.
216 216
217 The ``struct btf_array`` allows chaining throu 217 The ``struct btf_array`` allows chaining through element type to represent
218 multidimensional arrays. For example, for ``in 218 multidimensional arrays. For example, for ``int a[5][6]``, the following type
219 information illustrates the chaining: 219 information illustrates the chaining:
220 220
221 * [1]: int 221 * [1]: int
222 * [2]: array, ``btf_array.type = [1]``, ``bt 222 * [2]: array, ``btf_array.type = [1]``, ``btf_array.nelems = 6``
223 * [3]: array, ``btf_array.type = [2]``, ``bt 223 * [3]: array, ``btf_array.type = [2]``, ``btf_array.nelems = 5``
224 224
225 Currently, both pahole and llvm collapse multi 225 Currently, both pahole and llvm collapse multidimensional array into
226 one-dimensional array, e.g., for ``a[5][6]``, 226 one-dimensional array, e.g., for ``a[5][6]``, the ``btf_array.nelems`` is
227 equal to ``30``. This is because the original 227 equal to ``30``. This is because the original use case is map pretty print
228 where the whole array is dumped out so one-dim 228 where the whole array is dumped out so one-dimensional array is enough. As
229 more BTF usage is explored, pahole and llvm ca 229 more BTF usage is explored, pahole and llvm can be changed to generate proper
230 chained representation for multidimensional ar 230 chained representation for multidimensional arrays.
231 231
232 2.2.4 BTF_KIND_STRUCT 232 2.2.4 BTF_KIND_STRUCT
233 ~~~~~~~~~~~~~~~~~~~~~ 233 ~~~~~~~~~~~~~~~~~~~~~
234 2.2.5 BTF_KIND_UNION 234 2.2.5 BTF_KIND_UNION
235 ~~~~~~~~~~~~~~~~~~~~ 235 ~~~~~~~~~~~~~~~~~~~~
236 236
237 ``struct btf_type`` encoding requirement: 237 ``struct btf_type`` encoding requirement:
238 * ``name_off``: 0 or offset to a valid C ide 238 * ``name_off``: 0 or offset to a valid C identifier
239 * ``info.kind_flag``: 0 or 1 239 * ``info.kind_flag``: 0 or 1
240 * ``info.kind``: BTF_KIND_STRUCT or BTF_KIND 240 * ``info.kind``: BTF_KIND_STRUCT or BTF_KIND_UNION
241 * ``info.vlen``: the number of struct/union 241 * ``info.vlen``: the number of struct/union members
242 * ``info.size``: the size of the struct/unio 242 * ``info.size``: the size of the struct/union in bytes
243 243
244 ``btf_type`` is followed by ``info.vlen`` numb 244 ``btf_type`` is followed by ``info.vlen`` number of ``struct btf_member``.::
245 245
246 struct btf_member { 246 struct btf_member {
247 __u32 name_off; 247 __u32 name_off;
248 __u32 type; 248 __u32 type;
249 __u32 offset; 249 __u32 offset;
250 }; 250 };
251 251
252 ``struct btf_member`` encoding: 252 ``struct btf_member`` encoding:
253 * ``name_off``: offset to a valid C identifi 253 * ``name_off``: offset to a valid C identifier
254 * ``type``: the member type 254 * ``type``: the member type
255 * ``offset``: <see below> 255 * ``offset``: <see below>
256 256
257 If the type info ``kind_flag`` is not set, the 257 If the type info ``kind_flag`` is not set, the offset contains only bit offset
258 of the member. Note that the base type of the 258 of the member. Note that the base type of the bitfield can only be int or enum
259 type. If the bitfield size is 32, the base typ 259 type. If the bitfield size is 32, the base type can be either int or enum
260 type. If the bitfield size is not 32, the base 260 type. If the bitfield size is not 32, the base type must be int, and int type
261 ``BTF_INT_BITS()`` encodes the bitfield size. 261 ``BTF_INT_BITS()`` encodes the bitfield size.
262 262
263 If the ``kind_flag`` is set, the ``btf_member. 263 If the ``kind_flag`` is set, the ``btf_member.offset`` contains both member
264 bitfield size and bit offset. The bitfield siz 264 bitfield size and bit offset. The bitfield size and bit offset are calculated
265 as below.:: 265 as below.::
266 266
267 #define BTF_MEMBER_BITFIELD_SIZE(val) ((va 267 #define BTF_MEMBER_BITFIELD_SIZE(val) ((val) >> 24)
268 #define BTF_MEMBER_BIT_OFFSET(val) ((va 268 #define BTF_MEMBER_BIT_OFFSET(val) ((val) & 0xffffff)
269 269
270 In this case, if the base type is an int type, 270 In this case, if the base type is an int type, it must be a regular int type:
271 271
272 * ``BTF_INT_OFFSET()`` must be 0. 272 * ``BTF_INT_OFFSET()`` must be 0.
273 * ``BTF_INT_BITS()`` must be equal to ``{1,2 273 * ``BTF_INT_BITS()`` must be equal to ``{1,2,4,8,16} * 8``.
274 274
275 Commit 9d5f9f701b18 introduced ``kind_flag`` a 275 Commit 9d5f9f701b18 introduced ``kind_flag`` and explains why both modes
276 exist. 276 exist.
277 277
278 2.2.6 BTF_KIND_ENUM 278 2.2.6 BTF_KIND_ENUM
279 ~~~~~~~~~~~~~~~~~~~ 279 ~~~~~~~~~~~~~~~~~~~
280 280
281 ``struct btf_type`` encoding requirement: 281 ``struct btf_type`` encoding requirement:
282 * ``name_off``: 0 or offset to a valid C ide 282 * ``name_off``: 0 or offset to a valid C identifier
283 * ``info.kind_flag``: 0 for unsigned, 1 for 283 * ``info.kind_flag``: 0 for unsigned, 1 for signed
284 * ``info.kind``: BTF_KIND_ENUM 284 * ``info.kind``: BTF_KIND_ENUM
285 * ``info.vlen``: number of enum values 285 * ``info.vlen``: number of enum values
286 * ``size``: 1/2/4/8 286 * ``size``: 1/2/4/8
287 287
288 ``btf_type`` is followed by ``info.vlen`` numb 288 ``btf_type`` is followed by ``info.vlen`` number of ``struct btf_enum``.::
289 289
290 struct btf_enum { 290 struct btf_enum {
291 __u32 name_off; 291 __u32 name_off;
292 __s32 val; 292 __s32 val;
293 }; 293 };
294 294
295 The ``btf_enum`` encoding: 295 The ``btf_enum`` encoding:
296 * ``name_off``: offset to a valid C identifi 296 * ``name_off``: offset to a valid C identifier
297 * ``val``: any value 297 * ``val``: any value
298 298
299 If the original enum value is signed and the s 299 If the original enum value is signed and the size is less than 4,
300 that value will be sign extended into 4 bytes. 300 that value will be sign extended into 4 bytes. If the size is 8,
301 the value will be truncated into 4 bytes. 301 the value will be truncated into 4 bytes.
302 302
303 2.2.7 BTF_KIND_FWD 303 2.2.7 BTF_KIND_FWD
304 ~~~~~~~~~~~~~~~~~~ 304 ~~~~~~~~~~~~~~~~~~
305 305
306 ``struct btf_type`` encoding requirement: 306 ``struct btf_type`` encoding requirement:
307 * ``name_off``: offset to a valid C identifi 307 * ``name_off``: offset to a valid C identifier
308 * ``info.kind_flag``: 0 for struct, 1 for un 308 * ``info.kind_flag``: 0 for struct, 1 for union
309 * ``info.kind``: BTF_KIND_FWD 309 * ``info.kind``: BTF_KIND_FWD
310 * ``info.vlen``: 0 310 * ``info.vlen``: 0
311 * ``type``: 0 311 * ``type``: 0
312 312
313 No additional type data follow ``btf_type``. 313 No additional type data follow ``btf_type``.
314 314
315 2.2.8 BTF_KIND_TYPEDEF 315 2.2.8 BTF_KIND_TYPEDEF
316 ~~~~~~~~~~~~~~~~~~~~~~ 316 ~~~~~~~~~~~~~~~~~~~~~~
317 317
318 ``struct btf_type`` encoding requirement: 318 ``struct btf_type`` encoding requirement:
319 * ``name_off``: offset to a valid C identifi 319 * ``name_off``: offset to a valid C identifier
320 * ``info.kind_flag``: 0 320 * ``info.kind_flag``: 0
321 * ``info.kind``: BTF_KIND_TYPEDEF 321 * ``info.kind``: BTF_KIND_TYPEDEF
322 * ``info.vlen``: 0 322 * ``info.vlen``: 0
323 * ``type``: the type which can be referred b 323 * ``type``: the type which can be referred by name at ``name_off``
324 324
325 No additional type data follow ``btf_type``. 325 No additional type data follow ``btf_type``.
326 326
327 2.2.9 BTF_KIND_VOLATILE 327 2.2.9 BTF_KIND_VOLATILE
328 ~~~~~~~~~~~~~~~~~~~~~~~ 328 ~~~~~~~~~~~~~~~~~~~~~~~
329 329
330 ``struct btf_type`` encoding requirement: 330 ``struct btf_type`` encoding requirement:
331 * ``name_off``: 0 331 * ``name_off``: 0
332 * ``info.kind_flag``: 0 332 * ``info.kind_flag``: 0
333 * ``info.kind``: BTF_KIND_VOLATILE 333 * ``info.kind``: BTF_KIND_VOLATILE
334 * ``info.vlen``: 0 334 * ``info.vlen``: 0
335 * ``type``: the type with ``volatile`` quali 335 * ``type``: the type with ``volatile`` qualifier
336 336
337 No additional type data follow ``btf_type``. 337 No additional type data follow ``btf_type``.
338 338
339 2.2.10 BTF_KIND_CONST 339 2.2.10 BTF_KIND_CONST
340 ~~~~~~~~~~~~~~~~~~~~~ 340 ~~~~~~~~~~~~~~~~~~~~~
341 341
342 ``struct btf_type`` encoding requirement: 342 ``struct btf_type`` encoding requirement:
343 * ``name_off``: 0 343 * ``name_off``: 0
344 * ``info.kind_flag``: 0 344 * ``info.kind_flag``: 0
345 * ``info.kind``: BTF_KIND_CONST 345 * ``info.kind``: BTF_KIND_CONST
346 * ``info.vlen``: 0 346 * ``info.vlen``: 0
347 * ``type``: the type with ``const`` qualifie 347 * ``type``: the type with ``const`` qualifier
348 348
349 No additional type data follow ``btf_type``. 349 No additional type data follow ``btf_type``.
350 350
351 2.2.11 BTF_KIND_RESTRICT 351 2.2.11 BTF_KIND_RESTRICT
352 ~~~~~~~~~~~~~~~~~~~~~~~~ 352 ~~~~~~~~~~~~~~~~~~~~~~~~
353 353
354 ``struct btf_type`` encoding requirement: 354 ``struct btf_type`` encoding requirement:
355 * ``name_off``: 0 355 * ``name_off``: 0
356 * ``info.kind_flag``: 0 356 * ``info.kind_flag``: 0
357 * ``info.kind``: BTF_KIND_RESTRICT 357 * ``info.kind``: BTF_KIND_RESTRICT
358 * ``info.vlen``: 0 358 * ``info.vlen``: 0
359 * ``type``: the type with ``restrict`` quali 359 * ``type``: the type with ``restrict`` qualifier
360 360
361 No additional type data follow ``btf_type``. 361 No additional type data follow ``btf_type``.
362 362
363 2.2.12 BTF_KIND_FUNC 363 2.2.12 BTF_KIND_FUNC
364 ~~~~~~~~~~~~~~~~~~~~ 364 ~~~~~~~~~~~~~~~~~~~~
365 365
366 ``struct btf_type`` encoding requirement: 366 ``struct btf_type`` encoding requirement:
367 * ``name_off``: offset to a valid C identifi 367 * ``name_off``: offset to a valid C identifier
368 * ``info.kind_flag``: 0 368 * ``info.kind_flag``: 0
369 * ``info.kind``: BTF_KIND_FUNC 369 * ``info.kind``: BTF_KIND_FUNC
370 * ``info.vlen``: linkage information (BTF_FU 370 * ``info.vlen``: linkage information (BTF_FUNC_STATIC, BTF_FUNC_GLOBAL
371 or BTF_FUNC_EXTERN - see :r 371 or BTF_FUNC_EXTERN - see :ref:`BTF_Function_Linkage_Constants`)
372 * ``type``: a BTF_KIND_FUNC_PROTO type 372 * ``type``: a BTF_KIND_FUNC_PROTO type
373 373
374 No additional type data follow ``btf_type``. 374 No additional type data follow ``btf_type``.
375 375
376 A BTF_KIND_FUNC defines not a type, but a subp 376 A BTF_KIND_FUNC defines not a type, but a subprogram (function) whose
377 signature is defined by ``type``. The subprogr 377 signature is defined by ``type``. The subprogram is thus an instance of that
378 type. The BTF_KIND_FUNC may in turn be referen 378 type. The BTF_KIND_FUNC may in turn be referenced by a func_info in the
379 :ref:`BTF_Ext_Section` (ELF) or in the argumen 379 :ref:`BTF_Ext_Section` (ELF) or in the arguments to :ref:`BPF_Prog_Load`
380 (ABI). 380 (ABI).
381 381
382 Currently, only linkage values of BTF_FUNC_STA 382 Currently, only linkage values of BTF_FUNC_STATIC and BTF_FUNC_GLOBAL are
383 supported in the kernel. 383 supported in the kernel.
384 384
385 2.2.13 BTF_KIND_FUNC_PROTO 385 2.2.13 BTF_KIND_FUNC_PROTO
386 ~~~~~~~~~~~~~~~~~~~~~~~~~~ 386 ~~~~~~~~~~~~~~~~~~~~~~~~~~
387 387
388 ``struct btf_type`` encoding requirement: 388 ``struct btf_type`` encoding requirement:
389 * ``name_off``: 0 389 * ``name_off``: 0
390 * ``info.kind_flag``: 0 390 * ``info.kind_flag``: 0
391 * ``info.kind``: BTF_KIND_FUNC_PROTO 391 * ``info.kind``: BTF_KIND_FUNC_PROTO
392 * ``info.vlen``: # of parameters 392 * ``info.vlen``: # of parameters
393 * ``type``: the return type 393 * ``type``: the return type
394 394
395 ``btf_type`` is followed by ``info.vlen`` numb 395 ``btf_type`` is followed by ``info.vlen`` number of ``struct btf_param``.::
396 396
397 struct btf_param { 397 struct btf_param {
398 __u32 name_off; 398 __u32 name_off;
399 __u32 type; 399 __u32 type;
400 }; 400 };
401 401
402 If a BTF_KIND_FUNC_PROTO type is referred by a 402 If a BTF_KIND_FUNC_PROTO type is referred by a BTF_KIND_FUNC type, then
403 ``btf_param.name_off`` must point to a valid C 403 ``btf_param.name_off`` must point to a valid C identifier except for the
404 possible last argument representing the variab 404 possible last argument representing the variable argument. The btf_param.type
405 refers to parameter type. 405 refers to parameter type.
406 406
407 If the function has variable arguments, the la 407 If the function has variable arguments, the last parameter is encoded with
408 ``name_off = 0`` and ``type = 0``. 408 ``name_off = 0`` and ``type = 0``.
409 409
410 2.2.14 BTF_KIND_VAR 410 2.2.14 BTF_KIND_VAR
411 ~~~~~~~~~~~~~~~~~~~ 411 ~~~~~~~~~~~~~~~~~~~
412 412
413 ``struct btf_type`` encoding requirement: 413 ``struct btf_type`` encoding requirement:
414 * ``name_off``: offset to a valid C identifi 414 * ``name_off``: offset to a valid C identifier
415 * ``info.kind_flag``: 0 415 * ``info.kind_flag``: 0
416 * ``info.kind``: BTF_KIND_VAR 416 * ``info.kind``: BTF_KIND_VAR
417 * ``info.vlen``: 0 417 * ``info.vlen``: 0
418 * ``type``: the type of the variable 418 * ``type``: the type of the variable
419 419
420 ``btf_type`` is followed by a single ``struct 420 ``btf_type`` is followed by a single ``struct btf_variable`` with the
421 following data:: 421 following data::
422 422
423 struct btf_var { 423 struct btf_var {
424 __u32 linkage; 424 __u32 linkage;
425 }; 425 };
426 426
427 ``btf_var.linkage`` may take the values: BTF_V 427 ``btf_var.linkage`` may take the values: BTF_VAR_STATIC, BTF_VAR_GLOBAL_ALLOCATED or BTF_VAR_GLOBAL_EXTERN -
428 see :ref:`BTF_Var_Linkage_Constants`. 428 see :ref:`BTF_Var_Linkage_Constants`.
429 429
430 Not all type of global variables are supported 430 Not all type of global variables are supported by LLVM at this point.
431 The following is currently available: 431 The following is currently available:
432 432
433 * static variables with or without section a 433 * static variables with or without section attributes
434 * global variables with section attributes 434 * global variables with section attributes
435 435
436 The latter is for future extraction of map key 436 The latter is for future extraction of map key/value type id's from a
437 map definition. 437 map definition.
438 438
439 2.2.15 BTF_KIND_DATASEC 439 2.2.15 BTF_KIND_DATASEC
440 ~~~~~~~~~~~~~~~~~~~~~~~ 440 ~~~~~~~~~~~~~~~~~~~~~~~
441 441
442 ``struct btf_type`` encoding requirement: 442 ``struct btf_type`` encoding requirement:
443 * ``name_off``: offset to a valid name assoc 443 * ``name_off``: offset to a valid name associated with a variable or
444 one of .data/.bss/.rodata 444 one of .data/.bss/.rodata
445 * ``info.kind_flag``: 0 445 * ``info.kind_flag``: 0
446 * ``info.kind``: BTF_KIND_DATASEC 446 * ``info.kind``: BTF_KIND_DATASEC
447 * ``info.vlen``: # of variables 447 * ``info.vlen``: # of variables
448 * ``size``: total section size in bytes (0 a 448 * ``size``: total section size in bytes (0 at compilation time, patched
449 to actual size by BPF loaders su 449 to actual size by BPF loaders such as libbpf)
450 450
451 ``btf_type`` is followed by ``info.vlen`` numb 451 ``btf_type`` is followed by ``info.vlen`` number of ``struct btf_var_secinfo``.::
452 452
453 struct btf_var_secinfo { 453 struct btf_var_secinfo {
454 __u32 type; 454 __u32 type;
455 __u32 offset; 455 __u32 offset;
456 __u32 size; 456 __u32 size;
457 }; 457 };
458 458
459 ``struct btf_var_secinfo`` encoding: 459 ``struct btf_var_secinfo`` encoding:
460 * ``type``: the type of the BTF_KIND_VAR var 460 * ``type``: the type of the BTF_KIND_VAR variable
461 * ``offset``: the in-section offset of the v 461 * ``offset``: the in-section offset of the variable
462 * ``size``: the size of the variable in byte 462 * ``size``: the size of the variable in bytes
463 463
464 2.2.16 BTF_KIND_FLOAT 464 2.2.16 BTF_KIND_FLOAT
465 ~~~~~~~~~~~~~~~~~~~~~ 465 ~~~~~~~~~~~~~~~~~~~~~
466 466
467 ``struct btf_type`` encoding requirement: 467 ``struct btf_type`` encoding requirement:
468 * ``name_off``: any valid offset 468 * ``name_off``: any valid offset
469 * ``info.kind_flag``: 0 469 * ``info.kind_flag``: 0
470 * ``info.kind``: BTF_KIND_FLOAT 470 * ``info.kind``: BTF_KIND_FLOAT
471 * ``info.vlen``: 0 471 * ``info.vlen``: 0
472 * ``size``: the size of the float type in byt 472 * ``size``: the size of the float type in bytes: 2, 4, 8, 12 or 16.
473 473
474 No additional type data follow ``btf_type``. 474 No additional type data follow ``btf_type``.
475 475
476 2.2.17 BTF_KIND_DECL_TAG 476 2.2.17 BTF_KIND_DECL_TAG
477 ~~~~~~~~~~~~~~~~~~~~~~~~ 477 ~~~~~~~~~~~~~~~~~~~~~~~~
478 478
479 ``struct btf_type`` encoding requirement: 479 ``struct btf_type`` encoding requirement:
480 * ``name_off``: offset to a non-empty string 480 * ``name_off``: offset to a non-empty string
481 * ``info.kind_flag``: 0 481 * ``info.kind_flag``: 0
482 * ``info.kind``: BTF_KIND_DECL_TAG 482 * ``info.kind``: BTF_KIND_DECL_TAG
483 * ``info.vlen``: 0 483 * ``info.vlen``: 0
484 * ``type``: ``struct``, ``union``, ``func``, 484 * ``type``: ``struct``, ``union``, ``func``, ``var`` or ``typedef``
485 485
486 ``btf_type`` is followed by ``struct btf_decl_ 486 ``btf_type`` is followed by ``struct btf_decl_tag``.::
487 487
488 struct btf_decl_tag { 488 struct btf_decl_tag {
489 __u32 component_idx; 489 __u32 component_idx;
490 }; 490 };
491 491
492 The ``name_off`` encodes btf_decl_tag attribut 492 The ``name_off`` encodes btf_decl_tag attribute string.
493 The ``type`` should be ``struct``, ``union``, 493 The ``type`` should be ``struct``, ``union``, ``func``, ``var`` or ``typedef``.
494 For ``var`` or ``typedef`` type, ``btf_decl_ta 494 For ``var`` or ``typedef`` type, ``btf_decl_tag.component_idx`` must be ``-1``.
495 For the other three types, if the btf_decl_tag 495 For the other three types, if the btf_decl_tag attribute is
496 applied to the ``struct``, ``union`` or ``func 496 applied to the ``struct``, ``union`` or ``func`` itself,
497 ``btf_decl_tag.component_idx`` must be ``-1``. 497 ``btf_decl_tag.component_idx`` must be ``-1``. Otherwise,
498 the attribute is applied to a ``struct``/``uni 498 the attribute is applied to a ``struct``/``union`` member or
499 a ``func`` argument, and ``btf_decl_tag.compon 499 a ``func`` argument, and ``btf_decl_tag.component_idx`` should be a
500 valid index (starting from 0) pointing to a me 500 valid index (starting from 0) pointing to a member or an argument.
501 501
502 2.2.18 BTF_KIND_TYPE_TAG 502 2.2.18 BTF_KIND_TYPE_TAG
503 ~~~~~~~~~~~~~~~~~~~~~~~~ 503 ~~~~~~~~~~~~~~~~~~~~~~~~
504 504
505 ``struct btf_type`` encoding requirement: 505 ``struct btf_type`` encoding requirement:
506 * ``name_off``: offset to a non-empty string 506 * ``name_off``: offset to a non-empty string
507 * ``info.kind_flag``: 0 507 * ``info.kind_flag``: 0
508 * ``info.kind``: BTF_KIND_TYPE_TAG 508 * ``info.kind``: BTF_KIND_TYPE_TAG
509 * ``info.vlen``: 0 509 * ``info.vlen``: 0
510 * ``type``: the type with ``btf_type_tag`` at 510 * ``type``: the type with ``btf_type_tag`` attribute
511 511
512 Currently, ``BTF_KIND_TYPE_TAG`` is only emitt 512 Currently, ``BTF_KIND_TYPE_TAG`` is only emitted for pointer types.
513 It has the following btf type chain: 513 It has the following btf type chain:
514 :: 514 ::
515 515
516 ptr -> [type_tag]* 516 ptr -> [type_tag]*
517 -> [const | volatile | restrict | typede 517 -> [const | volatile | restrict | typedef]*
518 -> base_type 518 -> base_type
519 519
520 Basically, a pointer type points to zero or mo 520 Basically, a pointer type points to zero or more
521 type_tag, then zero or more const/volatile/res 521 type_tag, then zero or more const/volatile/restrict/typedef
522 and finally the base type. The base type is on 522 and finally the base type. The base type is one of
523 int, ptr, array, struct, union, enum, func_pro 523 int, ptr, array, struct, union, enum, func_proto and float types.
524 524
525 2.2.19 BTF_KIND_ENUM64 525 2.2.19 BTF_KIND_ENUM64
526 ~~~~~~~~~~~~~~~~~~~~~~ 526 ~~~~~~~~~~~~~~~~~~~~~~
527 527
528 ``struct btf_type`` encoding requirement: 528 ``struct btf_type`` encoding requirement:
529 * ``name_off``: 0 or offset to a valid C ide 529 * ``name_off``: 0 or offset to a valid C identifier
530 * ``info.kind_flag``: 0 for unsigned, 1 for 530 * ``info.kind_flag``: 0 for unsigned, 1 for signed
531 * ``info.kind``: BTF_KIND_ENUM64 531 * ``info.kind``: BTF_KIND_ENUM64
532 * ``info.vlen``: number of enum values 532 * ``info.vlen``: number of enum values
533 * ``size``: 1/2/4/8 533 * ``size``: 1/2/4/8
534 534
535 ``btf_type`` is followed by ``info.vlen`` numb 535 ``btf_type`` is followed by ``info.vlen`` number of ``struct btf_enum64``.::
536 536
537 struct btf_enum64 { 537 struct btf_enum64 {
538 __u32 name_off; 538 __u32 name_off;
539 __u32 val_lo32; 539 __u32 val_lo32;
540 __u32 val_hi32; 540 __u32 val_hi32;
541 }; 541 };
542 542
543 The ``btf_enum64`` encoding: 543 The ``btf_enum64`` encoding:
544 * ``name_off``: offset to a valid C identifi 544 * ``name_off``: offset to a valid C identifier
545 * ``val_lo32``: lower 32-bit value for a 64- 545 * ``val_lo32``: lower 32-bit value for a 64-bit value
546 * ``val_hi32``: high 32-bit value for a 64-b 546 * ``val_hi32``: high 32-bit value for a 64-bit value
547 547
548 If the original enum value is signed and the s 548 If the original enum value is signed and the size is less than 8,
549 that value will be sign extended into 8 bytes. 549 that value will be sign extended into 8 bytes.
550 550
551 2.3 Constant Values 551 2.3 Constant Values
552 ------------------- 552 -------------------
553 553
554 .. _BTF_Function_Linkage_Constants: 554 .. _BTF_Function_Linkage_Constants:
555 555
556 2.3.1 Function Linkage Constant Values 556 2.3.1 Function Linkage Constant Values
557 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 557 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
558 .. table:: Function Linkage Values and Meaning 558 .. table:: Function Linkage Values and Meanings
559 559
560 =================== ===== =========== 560 =================== ===== ===========
561 kind value description 561 kind value description
562 =================== ===== =========== 562 =================== ===== ===========
563 ``BTF_FUNC_STATIC`` 0x0 definition of su 563 ``BTF_FUNC_STATIC`` 0x0 definition of subprogram not visible outside containing compilation unit
564 ``BTF_FUNC_GLOBAL`` 0x1 definition of su 564 ``BTF_FUNC_GLOBAL`` 0x1 definition of subprogram visible outside containing compilation unit
565 ``BTF_FUNC_EXTERN`` 0x2 declaration of a 565 ``BTF_FUNC_EXTERN`` 0x2 declaration of a subprogram whose definition is outside the containing compilation unit
566 =================== ===== =========== 566 =================== ===== ===========
567 567
568 568
569 .. _BTF_Var_Linkage_Constants: 569 .. _BTF_Var_Linkage_Constants:
570 570
571 2.3.2 Variable Linkage Constant Values 571 2.3.2 Variable Linkage Constant Values
572 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 572 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
573 .. table:: Variable Linkage Values and Meaning 573 .. table:: Variable Linkage Values and Meanings
574 574
575 ============================ ===== ======= 575 ============================ ===== ===========
576 kind value descrip 576 kind value description
577 ============================ ===== ======= 577 ============================ ===== ===========
578 ``BTF_VAR_STATIC`` 0x0 definit 578 ``BTF_VAR_STATIC`` 0x0 definition of global variable not visible outside containing compilation unit
579 ``BTF_VAR_GLOBAL_ALLOCATED`` 0x1 definit 579 ``BTF_VAR_GLOBAL_ALLOCATED`` 0x1 definition of global variable visible outside containing compilation unit
580 ``BTF_VAR_GLOBAL_EXTERN`` 0x2 declara 580 ``BTF_VAR_GLOBAL_EXTERN`` 0x2 declaration of global variable whose definition is outside the containing compilation unit
581 ============================ ===== ======= 581 ============================ ===== ===========
582 582
583 3. BTF Kernel API 583 3. BTF Kernel API
584 ================= 584 =================
585 585
586 The following bpf syscall command involves BTF 586 The following bpf syscall command involves BTF:
587 * BPF_BTF_LOAD: load a blob of BTF data int 587 * BPF_BTF_LOAD: load a blob of BTF data into kernel
588 * BPF_MAP_CREATE: map creation with btf key 588 * BPF_MAP_CREATE: map creation with btf key and value type info.
589 * BPF_PROG_LOAD: prog load with btf functio 589 * BPF_PROG_LOAD: prog load with btf function and line info.
590 * BPF_BTF_GET_FD_BY_ID: get a btf fd 590 * BPF_BTF_GET_FD_BY_ID: get a btf fd
591 * BPF_OBJ_GET_INFO_BY_FD: btf, func_info, l 591 * BPF_OBJ_GET_INFO_BY_FD: btf, func_info, line_info
592 and other btf related info are returned. 592 and other btf related info are returned.
593 593
594 The workflow typically looks like: 594 The workflow typically looks like:
595 :: 595 ::
596 596
597 Application: 597 Application:
598 BPF_BTF_LOAD 598 BPF_BTF_LOAD
599 | 599 |
600 v 600 v
601 BPF_MAP_CREATE and BPF_PROG_LOAD 601 BPF_MAP_CREATE and BPF_PROG_LOAD
602 | 602 |
603 V 603 V
604 ...... 604 ......
605 605
606 Introspection tool: 606 Introspection tool:
607 ...... 607 ......
608 BPF_{PROG,MAP}_GET_NEXT_ID (get prog/map 608 BPF_{PROG,MAP}_GET_NEXT_ID (get prog/map id's)
609 | 609 |
610 V 610 V
611 BPF_{PROG,MAP}_GET_FD_BY_ID (get a prog/ 611 BPF_{PROG,MAP}_GET_FD_BY_ID (get a prog/map fd)
612 | 612 |
613 V 613 V
614 BPF_OBJ_GET_INFO_BY_FD (get bpf_prog_inf 614 BPF_OBJ_GET_INFO_BY_FD (get bpf_prog_info/bpf_map_info with btf_id)
615 | 615 | |
616 V 616 V |
617 BPF_BTF_GET_FD_BY_ID (get btf_fd) 617 BPF_BTF_GET_FD_BY_ID (get btf_fd) |
618 | 618 | |
619 V 619 V |
620 BPF_OBJ_GET_INFO_BY_FD (get btf) 620 BPF_OBJ_GET_INFO_BY_FD (get btf) |
621 | 621 | |
622 V 622 V V
623 pretty print types, dump func signatures 623 pretty print types, dump func signatures and line info, etc.
624 624
625 625
626 3.1 BPF_BTF_LOAD 626 3.1 BPF_BTF_LOAD
627 ---------------- 627 ----------------
628 628
629 Load a blob of BTF data into kernel. A blob of 629 Load a blob of BTF data into kernel. A blob of data, described in
630 :ref:`BTF_Type_String`, can be directly loaded 630 :ref:`BTF_Type_String`, can be directly loaded into the kernel. A ``btf_fd``
631 is returned to a userspace. 631 is returned to a userspace.
632 632
633 3.2 BPF_MAP_CREATE 633 3.2 BPF_MAP_CREATE
634 ------------------ 634 ------------------
635 635
636 A map can be created with ``btf_fd`` and speci 636 A map can be created with ``btf_fd`` and specified key/value type id.::
637 637
638 __u32 btf_fd; /* fd pointing to 638 __u32 btf_fd; /* fd pointing to a BTF type data */
639 __u32 btf_key_type_id; /* BTF typ 639 __u32 btf_key_type_id; /* BTF type_id of the key */
640 __u32 btf_value_type_id; /* BTF typ 640 __u32 btf_value_type_id; /* BTF type_id of the value */
641 641
642 In libbpf, the map can be defined with extra a 642 In libbpf, the map can be defined with extra annotation like below:
643 :: 643 ::
644 644
645 struct { 645 struct {
646 __uint(type, BPF_MAP_TYPE_ARRAY); 646 __uint(type, BPF_MAP_TYPE_ARRAY);
647 __type(key, int); 647 __type(key, int);
648 __type(value, struct ipv_counts); 648 __type(value, struct ipv_counts);
649 __uint(max_entries, 4); 649 __uint(max_entries, 4);
650 } btf_map SEC(".maps"); 650 } btf_map SEC(".maps");
651 651
652 During ELF parsing, libbpf is able to extract 652 During ELF parsing, libbpf is able to extract key/value type_id's and assign
653 them to BPF_MAP_CREATE attributes automaticall 653 them to BPF_MAP_CREATE attributes automatically.
654 654
655 .. _BPF_Prog_Load: 655 .. _BPF_Prog_Load:
656 656
657 3.3 BPF_PROG_LOAD 657 3.3 BPF_PROG_LOAD
658 ----------------- 658 -----------------
659 659
660 During prog_load, func_info and line_info can 660 During prog_load, func_info and line_info can be passed to kernel with proper
661 values for the following attributes: 661 values for the following attributes:
662 :: 662 ::
663 663
664 __u32 insn_cnt; 664 __u32 insn_cnt;
665 __aligned_u64 insns; 665 __aligned_u64 insns;
666 ...... 666 ......
667 __u32 prog_btf_fd; /* fd poin 667 __u32 prog_btf_fd; /* fd pointing to BTF type data */
668 __u32 func_info_rec_size; /* 668 __u32 func_info_rec_size; /* userspace bpf_func_info size */
669 __aligned_u64 func_info; /* func in 669 __aligned_u64 func_info; /* func info */
670 __u32 func_info_cnt; /* number 670 __u32 func_info_cnt; /* number of bpf_func_info records */
671 __u32 line_info_rec_size; /* 671 __u32 line_info_rec_size; /* userspace bpf_line_info size */
672 __aligned_u64 line_info; /* line in 672 __aligned_u64 line_info; /* line info */
673 __u32 line_info_cnt; /* number 673 __u32 line_info_cnt; /* number of bpf_line_info records */
674 674
675 The func_info and line_info are an array of be 675 The func_info and line_info are an array of below, respectively.::
676 676
677 struct bpf_func_info { 677 struct bpf_func_info {
678 __u32 insn_off; /* [0, insn_cnt - 1] 678 __u32 insn_off; /* [0, insn_cnt - 1] */
679 __u32 type_id; /* pointing to a BTF 679 __u32 type_id; /* pointing to a BTF_KIND_FUNC type */
680 }; 680 };
681 struct bpf_line_info { 681 struct bpf_line_info {
682 __u32 insn_off; /* [0, insn_cnt - 1] 682 __u32 insn_off; /* [0, insn_cnt - 1] */
683 __u32 file_name_off; /* offset to st 683 __u32 file_name_off; /* offset to string table for the filename */
684 __u32 line_off; /* offset to string 684 __u32 line_off; /* offset to string table for the source line */
685 __u32 line_col; /* line number and c 685 __u32 line_col; /* line number and column number */
686 }; 686 };
687 687
688 func_info_rec_size is the size of each func_in 688 func_info_rec_size is the size of each func_info record, and
689 line_info_rec_size is the size of each line_in 689 line_info_rec_size is the size of each line_info record. Passing the record
690 size to kernel make it possible to extend the 690 size to kernel make it possible to extend the record itself in the future.
691 691
692 Below are requirements for func_info: 692 Below are requirements for func_info:
693 * func_info[0].insn_off must be 0. 693 * func_info[0].insn_off must be 0.
694 * the func_info insn_off is in strictly incr 694 * the func_info insn_off is in strictly increasing order and matches
695 bpf func boundaries. 695 bpf func boundaries.
696 696
697 Below are requirements for line_info: 697 Below are requirements for line_info:
698 * the first insn in each func must have a li 698 * the first insn in each func must have a line_info record pointing to it.
699 * the line_info insn_off is in strictly incr 699 * the line_info insn_off is in strictly increasing order.
700 700
701 For line_info, the line number and column numb 701 For line_info, the line number and column number are defined as below:
702 :: 702 ::
703 703
704 #define BPF_LINE_INFO_LINE_NUM(line_col) 704 #define BPF_LINE_INFO_LINE_NUM(line_col) ((line_col) >> 10)
705 #define BPF_LINE_INFO_LINE_COL(line_col) 705 #define BPF_LINE_INFO_LINE_COL(line_col) ((line_col) & 0x3ff)
706 706
707 3.4 BPF_{PROG,MAP}_GET_NEXT_ID 707 3.4 BPF_{PROG,MAP}_GET_NEXT_ID
708 ------------------------------ 708 ------------------------------
709 709
710 In kernel, every loaded program, map or btf ha 710 In kernel, every loaded program, map or btf has a unique id. The id won't
711 change during the lifetime of a program, map, 711 change during the lifetime of a program, map, or btf.
712 712
713 The bpf syscall command BPF_{PROG,MAP}_GET_NEX 713 The bpf syscall command BPF_{PROG,MAP}_GET_NEXT_ID returns all id's, one for
714 each command, to user space, for bpf program o 714 each command, to user space, for bpf program or maps, respectively, so an
715 inspection tool can inspect all programs and m 715 inspection tool can inspect all programs and maps.
716 716
717 3.5 BPF_{PROG,MAP}_GET_FD_BY_ID 717 3.5 BPF_{PROG,MAP}_GET_FD_BY_ID
718 ------------------------------- 718 -------------------------------
719 719
720 An introspection tool cannot use id to get det 720 An introspection tool cannot use id to get details about program or maps.
721 A file descriptor needs to be obtained first f 721 A file descriptor needs to be obtained first for reference-counting purpose.
722 722
723 3.6 BPF_OBJ_GET_INFO_BY_FD 723 3.6 BPF_OBJ_GET_INFO_BY_FD
724 -------------------------- 724 --------------------------
725 725
726 Once a program/map fd is acquired, an introspe 726 Once a program/map fd is acquired, an introspection tool can get the detailed
727 information from kernel about this fd, some of 727 information from kernel about this fd, some of which are BTF-related. For
728 example, ``bpf_map_info`` returns ``btf_id`` a 728 example, ``bpf_map_info`` returns ``btf_id`` and key/value type ids.
729 ``bpf_prog_info`` returns ``btf_id``, func_inf 729 ``bpf_prog_info`` returns ``btf_id``, func_info, and line info for translated
730 bpf byte codes, and jited_line_info. 730 bpf byte codes, and jited_line_info.
731 731
732 3.7 BPF_BTF_GET_FD_BY_ID 732 3.7 BPF_BTF_GET_FD_BY_ID
733 ------------------------ 733 ------------------------
734 734
735 With ``btf_id`` obtained in ``bpf_map_info`` a 735 With ``btf_id`` obtained in ``bpf_map_info`` and ``bpf_prog_info``, bpf
736 syscall command BPF_BTF_GET_FD_BY_ID can retri 736 syscall command BPF_BTF_GET_FD_BY_ID can retrieve a btf fd. Then, with
737 command BPF_OBJ_GET_INFO_BY_FD, the btf blob, 737 command BPF_OBJ_GET_INFO_BY_FD, the btf blob, originally loaded into the
738 kernel with BPF_BTF_LOAD, can be retrieved. 738 kernel with BPF_BTF_LOAD, can be retrieved.
739 739
740 With the btf blob, ``bpf_map_info``, and ``bpf 740 With the btf blob, ``bpf_map_info``, and ``bpf_prog_info``, an introspection
741 tool has full btf knowledge and is able to pre 741 tool has full btf knowledge and is able to pretty print map key/values, dump
742 func signatures and line info, along with byte 742 func signatures and line info, along with byte/jit codes.
743 743
744 4. ELF File Format Interface 744 4. ELF File Format Interface
745 ============================ 745 ============================
746 746
747 4.1 .BTF section 747 4.1 .BTF section
748 ---------------- 748 ----------------
749 749
750 The .BTF section contains type and string data 750 The .BTF section contains type and string data. The format of this section is
751 same as the one describe in :ref:`BTF_Type_Str 751 same as the one describe in :ref:`BTF_Type_String`.
752 752
753 .. _BTF_Ext_Section: 753 .. _BTF_Ext_Section:
754 754
755 4.2 .BTF.ext section 755 4.2 .BTF.ext section
756 -------------------- 756 --------------------
757 757
758 The .BTF.ext section encodes func_info, line_i 758 The .BTF.ext section encodes func_info, line_info and CO-RE relocations
759 which needs loader manipulation before loading 759 which needs loader manipulation before loading into the kernel.
760 760
761 The specification for .BTF.ext section is defi 761 The specification for .BTF.ext section is defined at ``tools/lib/bpf/btf.h``
762 and ``tools/lib/bpf/btf.c``. 762 and ``tools/lib/bpf/btf.c``.
763 763
764 The current header of .BTF.ext section:: 764 The current header of .BTF.ext section::
765 765
766 struct btf_ext_header { 766 struct btf_ext_header {
767 __u16 magic; 767 __u16 magic;
768 __u8 version; 768 __u8 version;
769 __u8 flags; 769 __u8 flags;
770 __u32 hdr_len; 770 __u32 hdr_len;
771 771
772 /* All offsets are in bytes relative t 772 /* All offsets are in bytes relative to the end of this header */
773 __u32 func_info_off; 773 __u32 func_info_off;
774 __u32 func_info_len; 774 __u32 func_info_len;
775 __u32 line_info_off; 775 __u32 line_info_off;
776 __u32 line_info_len; 776 __u32 line_info_len;
777 777
778 /* optional part of .BTF.ext header */ 778 /* optional part of .BTF.ext header */
779 __u32 core_relo_off; 779 __u32 core_relo_off;
780 __u32 core_relo_len; 780 __u32 core_relo_len;
781 }; 781 };
782 782
783 It is very similar to .BTF section. Instead of 783 It is very similar to .BTF section. Instead of type/string section, it
784 contains func_info, line_info and core_relo su 784 contains func_info, line_info and core_relo sub-sections.
785 See :ref:`BPF_Prog_Load` for details about fun 785 See :ref:`BPF_Prog_Load` for details about func_info and line_info
786 record format. 786 record format.
787 787
788 The func_info is organized as below.:: 788 The func_info is organized as below.::
789 789
790 func_info_rec_size /* __u32 790 func_info_rec_size /* __u32 value */
791 btf_ext_info_sec for section #1 /* func_i 791 btf_ext_info_sec for section #1 /* func_info for section #1 */
792 btf_ext_info_sec for section #2 /* func_i 792 btf_ext_info_sec for section #2 /* func_info for section #2 */
793 ... 793 ...
794 794
795 ``func_info_rec_size`` specifies the size of ` 795 ``func_info_rec_size`` specifies the size of ``bpf_func_info`` structure when
796 .BTF.ext is generated. ``btf_ext_info_sec``, d 796 .BTF.ext is generated. ``btf_ext_info_sec``, defined below, is a collection of
797 func_info for each specific ELF section.:: 797 func_info for each specific ELF section.::
798 798
799 struct btf_ext_info_sec { 799 struct btf_ext_info_sec {
800 __u32 sec_name_off; /* offset to sec 800 __u32 sec_name_off; /* offset to section name */
801 __u32 num_info; 801 __u32 num_info;
802 /* Followed by num_info * record_size 802 /* Followed by num_info * record_size number of bytes */
803 __u8 data[0]; 803 __u8 data[0];
804 }; 804 };
805 805
806 Here, num_info must be greater than 0. 806 Here, num_info must be greater than 0.
807 807
808 The line_info is organized as below.:: 808 The line_info is organized as below.::
809 809
810 line_info_rec_size /* __u32 810 line_info_rec_size /* __u32 value */
811 btf_ext_info_sec for section #1 /* line_i 811 btf_ext_info_sec for section #1 /* line_info for section #1 */
812 btf_ext_info_sec for section #2 /* line_i 812 btf_ext_info_sec for section #2 /* line_info for section #2 */
813 ... 813 ...
814 814
815 ``line_info_rec_size`` specifies the size of ` 815 ``line_info_rec_size`` specifies the size of ``bpf_line_info`` structure when
816 .BTF.ext is generated. 816 .BTF.ext is generated.
817 817
818 The interpretation of ``bpf_func_info->insn_of 818 The interpretation of ``bpf_func_info->insn_off`` and
819 ``bpf_line_info->insn_off`` is different betwe 819 ``bpf_line_info->insn_off`` is different between kernel API and ELF API. For
820 kernel API, the ``insn_off`` is the instructio 820 kernel API, the ``insn_off`` is the instruction offset in the unit of ``struct
821 bpf_insn``. For ELF API, the ``insn_off`` is t 821 bpf_insn``. For ELF API, the ``insn_off`` is the byte offset from the
822 beginning of section (``btf_ext_info_sec->sec_ 822 beginning of section (``btf_ext_info_sec->sec_name_off``).
823 823
824 The core_relo is organized as below.:: 824 The core_relo is organized as below.::
825 825
826 core_relo_rec_size /* __u32 826 core_relo_rec_size /* __u32 value */
827 btf_ext_info_sec for section #1 /* core_r 827 btf_ext_info_sec for section #1 /* core_relo for section #1 */
828 btf_ext_info_sec for section #2 /* core_r 828 btf_ext_info_sec for section #2 /* core_relo for section #2 */
829 829
830 ``core_relo_rec_size`` specifies the size of ` 830 ``core_relo_rec_size`` specifies the size of ``bpf_core_relo``
831 structure when .BTF.ext is generated. All ``bp 831 structure when .BTF.ext is generated. All ``bpf_core_relo`` structures
832 within a single ``btf_ext_info_sec`` describe 832 within a single ``btf_ext_info_sec`` describe relocations applied to
833 section named by ``btf_ext_info_sec->sec_name_ 833 section named by ``btf_ext_info_sec->sec_name_off``.
834 834
835 See :ref:`Documentation/bpf/llvm_reloc.rst <bt 835 See :ref:`Documentation/bpf/llvm_reloc.rst <btf-co-re-relocations>`
836 for more information on CO-RE relocations. 836 for more information on CO-RE relocations.
837 837
838 4.2 .BTF_ids section 838 4.2 .BTF_ids section
839 -------------------- 839 --------------------
840 840
841 The .BTF_ids section encodes BTF ID values tha 841 The .BTF_ids section encodes BTF ID values that are used within the kernel.
842 842
843 This section is created during the kernel comp 843 This section is created during the kernel compilation with the help of
844 macros defined in ``include/linux/btf_ids.h`` 844 macros defined in ``include/linux/btf_ids.h`` header file. Kernel code can
845 use them to create lists and sets (sorted list 845 use them to create lists and sets (sorted lists) of BTF ID values.
846 846
847 The ``BTF_ID_LIST`` and ``BTF_ID`` macros defi 847 The ``BTF_ID_LIST`` and ``BTF_ID`` macros define unsorted list of BTF ID values,
848 with following syntax:: 848 with following syntax::
849 849
850 BTF_ID_LIST(list) 850 BTF_ID_LIST(list)
851 BTF_ID(type1, name1) 851 BTF_ID(type1, name1)
852 BTF_ID(type2, name2) 852 BTF_ID(type2, name2)
853 853
854 resulting in following layout in .BTF_ids sect 854 resulting in following layout in .BTF_ids section::
855 855
856 __BTF_ID__type1__name1__1: 856 __BTF_ID__type1__name1__1:
857 .zero 4 857 .zero 4
858 __BTF_ID__type2__name2__2: 858 __BTF_ID__type2__name2__2:
859 .zero 4 859 .zero 4
860 860
861 The ``u32 list[];`` variable is defined to acc 861 The ``u32 list[];`` variable is defined to access the list.
862 862
863 The ``BTF_ID_UNUSED`` macro defines 4 zero byt 863 The ``BTF_ID_UNUSED`` macro defines 4 zero bytes. It's used when we
864 want to define unused entry in BTF_ID_LIST, li 864 want to define unused entry in BTF_ID_LIST, like::
865 865
866 BTF_ID_LIST(bpf_skb_output_btf_ids) 866 BTF_ID_LIST(bpf_skb_output_btf_ids)
867 BTF_ID(struct, sk_buff) 867 BTF_ID(struct, sk_buff)
868 BTF_ID_UNUSED 868 BTF_ID_UNUSED
869 BTF_ID(struct, task_struct) 869 BTF_ID(struct, task_struct)
870 870
871 The ``BTF_SET_START/END`` macros pair defines 871 The ``BTF_SET_START/END`` macros pair defines sorted list of BTF ID values
872 and their count, with following syntax:: 872 and their count, with following syntax::
873 873
874 BTF_SET_START(set) 874 BTF_SET_START(set)
875 BTF_ID(type1, name1) 875 BTF_ID(type1, name1)
876 BTF_ID(type2, name2) 876 BTF_ID(type2, name2)
877 BTF_SET_END(set) 877 BTF_SET_END(set)
878 878
879 resulting in following layout in .BTF_ids sect 879 resulting in following layout in .BTF_ids section::
880 880
881 __BTF_ID__set__set: 881 __BTF_ID__set__set:
882 .zero 4 882 .zero 4
883 __BTF_ID__type1__name1__3: 883 __BTF_ID__type1__name1__3:
884 .zero 4 884 .zero 4
885 __BTF_ID__type2__name2__4: 885 __BTF_ID__type2__name2__4:
886 .zero 4 886 .zero 4
887 887
888 The ``struct btf_id_set set;`` variable is def 888 The ``struct btf_id_set set;`` variable is defined to access the list.
889 889
890 The ``typeX`` name can be one of following:: 890 The ``typeX`` name can be one of following::
891 891
892 struct, union, typedef, func 892 struct, union, typedef, func
893 893
894 and is used as a filter when resolving the BTF 894 and is used as a filter when resolving the BTF ID value.
895 895
896 All the BTF ID lists and sets are compiled in 896 All the BTF ID lists and sets are compiled in the .BTF_ids section and
897 resolved during the linking phase of kernel bu 897 resolved during the linking phase of kernel build by ``resolve_btfids`` tool.
898 898
899 5. Using BTF 899 5. Using BTF
900 ============ 900 ============
901 901
902 5.1 bpftool map pretty print 902 5.1 bpftool map pretty print
903 ---------------------------- 903 ----------------------------
904 904
905 With BTF, the map key/value can be printed bas 905 With BTF, the map key/value can be printed based on fields rather than simply
906 raw bytes. This is especially valuable for lar 906 raw bytes. This is especially valuable for large structure or if your data
907 structure has bitfields. For example, for the 907 structure has bitfields. For example, for the following map,::
908 908
909 enum A { A1, A2, A3, A4, A5 }; 909 enum A { A1, A2, A3, A4, A5 };
910 typedef enum A ___A; 910 typedef enum A ___A;
911 struct tmp_t { 911 struct tmp_t {
912 char a1:4; 912 char a1:4;
913 int a2:4; 913 int a2:4;
914 int :4; 914 int :4;
915 __u32 a3:4; 915 __u32 a3:4;
916 int b; 916 int b;
917 ___A b1:4; 917 ___A b1:4;
918 enum A b2:4; 918 enum A b2:4;
919 }; 919 };
920 struct { 920 struct {
921 __uint(type, BPF_MAP_TYPE_ARRAY); 921 __uint(type, BPF_MAP_TYPE_ARRAY);
922 __type(key, int); 922 __type(key, int);
923 __type(value, struct tmp_t); 923 __type(value, struct tmp_t);
924 __uint(max_entries, 1); 924 __uint(max_entries, 1);
925 } tmpmap SEC(".maps"); 925 } tmpmap SEC(".maps");
926 926
927 bpftool is able to pretty print like below: 927 bpftool is able to pretty print like below:
928 :: 928 ::
929 929
930 [{ 930 [{
931 "key": 0, 931 "key": 0,
932 "value": { 932 "value": {
933 "a1": 0x2, 933 "a1": 0x2,
934 "a2": 0x4, 934 "a2": 0x4,
935 "a3": 0x6, 935 "a3": 0x6,
936 "b": 7, 936 "b": 7,
937 "b1": 0x8, 937 "b1": 0x8,
938 "b2": 0xa 938 "b2": 0xa
939 } 939 }
940 } 940 }
941 ] 941 ]
942 942
943 5.2 bpftool prog dump 943 5.2 bpftool prog dump
944 --------------------- 944 ---------------------
945 945
946 The following is an example showing how func_i 946 The following is an example showing how func_info and line_info can help prog
947 dump with better kernel symbol names, function 947 dump with better kernel symbol names, function prototypes and line
948 information.:: 948 information.::
949 949
950 $ bpftool prog dump jited pinned /sys/fs/b 950 $ bpftool prog dump jited pinned /sys/fs/bpf/test_btf_haskv
951 [...] 951 [...]
952 int test_long_fname_2(struct dummy_tracepo 952 int test_long_fname_2(struct dummy_tracepoint_args * arg):
953 bpf_prog_44a040bf25481309_test_long_fname_ 953 bpf_prog_44a040bf25481309_test_long_fname_2:
954 ; static int test_long_fname_2(struct dumm 954 ; static int test_long_fname_2(struct dummy_tracepoint_args *arg)
955 0: push %rbp 955 0: push %rbp
956 1: mov %rsp,%rbp 956 1: mov %rsp,%rbp
957 4: sub $0x30,%rsp 957 4: sub $0x30,%rsp
958 b: sub $0x28,%rbp 958 b: sub $0x28,%rbp
959 f: mov %rbx,0x0(%rbp) 959 f: mov %rbx,0x0(%rbp)
960 13: mov %r13,0x8(%rbp) 960 13: mov %r13,0x8(%rbp)
961 17: mov %r14,0x10(%rbp) 961 17: mov %r14,0x10(%rbp)
962 1b: mov %r15,0x18(%rbp) 962 1b: mov %r15,0x18(%rbp)
963 1f: xor %eax,%eax 963 1f: xor %eax,%eax
964 21: mov %rax,0x20(%rbp) 964 21: mov %rax,0x20(%rbp)
965 25: xor %esi,%esi 965 25: xor %esi,%esi
966 ; int key = 0; 966 ; int key = 0;
967 27: mov %esi,-0x4(%rbp) 967 27: mov %esi,-0x4(%rbp)
968 ; if (!arg->sock) 968 ; if (!arg->sock)
969 2a: mov 0x8(%rdi),%rdi 969 2a: mov 0x8(%rdi),%rdi
970 ; if (!arg->sock) 970 ; if (!arg->sock)
971 2e: cmp $0x0,%rdi 971 2e: cmp $0x0,%rdi
972 32: je 0x0000000000000070 972 32: je 0x0000000000000070
973 34: mov %rbp,%rsi 973 34: mov %rbp,%rsi
974 ; counts = bpf_map_lookup_elem(&btf_map, & 974 ; counts = bpf_map_lookup_elem(&btf_map, &key);
975 [...] 975 [...]
976 976
977 5.3 Verifier Log 977 5.3 Verifier Log
978 ---------------- 978 ----------------
979 979
980 The following is an example of how line_info c 980 The following is an example of how line_info can help debugging verification
981 failure.:: 981 failure.::
982 982
983 /* The code at tools/testing/selftests/ 983 /* The code at tools/testing/selftests/bpf/test_xdp_noinline.c
984 * is modified as below. 984 * is modified as below.
985 */ 985 */
986 data = (void *)(long)xdp->data; 986 data = (void *)(long)xdp->data;
987 data_end = (void *)(long)xdp->data_end; 987 data_end = (void *)(long)xdp->data_end;
988 /* 988 /*
989 if (data + 4 > data_end) 989 if (data + 4 > data_end)
990 return XDP_DROP; 990 return XDP_DROP;
991 */ 991 */
992 *(u32 *)data = dst->dst; 992 *(u32 *)data = dst->dst;
993 993
994 $ bpftool prog load ./test_xdp_noinline.o 994 $ bpftool prog load ./test_xdp_noinline.o /sys/fs/bpf/test_xdp_noinline type xdp
995 ; data = (void *)(long)xdp->data; 995 ; data = (void *)(long)xdp->data;
996 224: (79) r2 = *(u64 *)(r10 -112) 996 224: (79) r2 = *(u64 *)(r10 -112)
997 225: (61) r2 = *(u32 *)(r2 +0) 997 225: (61) r2 = *(u32 *)(r2 +0)
998 ; *(u32 *)data = dst->dst; 998 ; *(u32 *)data = dst->dst;
999 226: (63) *(u32 *)(r2 +0) = r1 999 226: (63) *(u32 *)(r2 +0) = r1
1000 invalid access to packet, off=0 size= 1000 invalid access to packet, off=0 size=4, R2(id=0,off=0,r=0)
1001 R2 offset is outside of the packet 1001 R2 offset is outside of the packet
1002 1002
1003 6. BTF Generation 1003 6. BTF Generation
1004 ================= 1004 =================
1005 1005
1006 You need latest pahole 1006 You need latest pahole
1007 1007
1008 https://git.kernel.org/pub/scm/devel/pahole 1008 https://git.kernel.org/pub/scm/devel/pahole/pahole.git/
1009 1009
1010 or llvm (8.0 or later). The pahole acts as a 1010 or llvm (8.0 or later). The pahole acts as a dwarf2btf converter. It doesn't
1011 support .BTF.ext and btf BTF_KIND_FUNC type y 1011 support .BTF.ext and btf BTF_KIND_FUNC type yet. For example,::
1012 1012
1013 -bash-4.4$ cat t.c 1013 -bash-4.4$ cat t.c
1014 struct t { 1014 struct t {
1015 int a:2; 1015 int a:2;
1016 int b:3; 1016 int b:3;
1017 int c:2; 1017 int c:2;
1018 } g; 1018 } g;
1019 -bash-4.4$ gcc -c -O2 -g t.c 1019 -bash-4.4$ gcc -c -O2 -g t.c
1020 -bash-4.4$ pahole -JV t.o 1020 -bash-4.4$ pahole -JV t.o
1021 File t.o: 1021 File t.o:
1022 [1] STRUCT t kind_flag=1 size=4 vlen=3 1022 [1] STRUCT t kind_flag=1 size=4 vlen=3
1023 a type_id=2 bitfield_size=2 bit 1023 a type_id=2 bitfield_size=2 bits_offset=0
1024 b type_id=2 bitfield_size=3 bit 1024 b type_id=2 bitfield_size=3 bits_offset=2
1025 c type_id=2 bitfield_size=2 bit 1025 c type_id=2 bitfield_size=2 bits_offset=5
1026 [2] INT int size=4 bit_offset=0 nr_bits 1026 [2] INT int size=4 bit_offset=0 nr_bits=32 encoding=SIGNED
1027 1027
1028 The llvm is able to generate .BTF and .BTF.ex 1028 The llvm is able to generate .BTF and .BTF.ext directly with -g for bpf target
1029 only. The assembly code (-S) is able to show 1029 only. The assembly code (-S) is able to show the BTF encoding in assembly
1030 format.:: 1030 format.::
1031 1031
1032 -bash-4.4$ cat t2.c 1032 -bash-4.4$ cat t2.c
1033 typedef int __int32; 1033 typedef int __int32;
1034 struct t2 { 1034 struct t2 {
1035 int a2; 1035 int a2;
1036 int (*f2)(char q1, __int32 q2, ...); 1036 int (*f2)(char q1, __int32 q2, ...);
1037 int (*f3)(); 1037 int (*f3)();
1038 } g2; 1038 } g2;
1039 int main() { return 0; } 1039 int main() { return 0; }
1040 int test() { return 0; } 1040 int test() { return 0; }
1041 -bash-4.4$ clang -c -g -O2 --target=bpf t 1041 -bash-4.4$ clang -c -g -O2 --target=bpf t2.c
1042 -bash-4.4$ readelf -S t2.o 1042 -bash-4.4$ readelf -S t2.o
1043 ...... 1043 ......
1044 [ 8] .BTF PROGBITS 1044 [ 8] .BTF PROGBITS 0000000000000000 00000247
1045 000000000000016e 0000000000000000 1045 000000000000016e 0000000000000000 0 0 1
1046 [ 9] .BTF.ext PROGBITS 1046 [ 9] .BTF.ext PROGBITS 0000000000000000 000003b5
1047 0000000000000060 0000000000000000 1047 0000000000000060 0000000000000000 0 0 1
1048 [10] .rel.BTF.ext REL 1048 [10] .rel.BTF.ext REL 0000000000000000 000007e0
1049 0000000000000040 0000000000000010 1049 0000000000000040 0000000000000010 16 9 8
1050 ...... 1050 ......
1051 -bash-4.4$ clang -S -g -O2 --target=bpf t 1051 -bash-4.4$ clang -S -g -O2 --target=bpf t2.c
1052 -bash-4.4$ cat t2.s 1052 -bash-4.4$ cat t2.s
1053 ...... 1053 ......
1054 .section .BTF,"",@progbits 1054 .section .BTF,"",@progbits
1055 .short 60319 # 1055 .short 60319 # 0xeb9f
1056 .byte 1 1056 .byte 1
1057 .byte 0 1057 .byte 0
1058 .long 24 1058 .long 24
1059 .long 0 1059 .long 0
1060 .long 220 1060 .long 220
1061 .long 220 1061 .long 220
1062 .long 122 1062 .long 122
1063 .long 0 # 1063 .long 0 # BTF_KIND_FUNC_PROTO(id = 1)
1064 .long 218103808 # 1064 .long 218103808 # 0xd000000
1065 .long 2 1065 .long 2
1066 .long 83 # 1066 .long 83 # BTF_KIND_INT(id = 2)
1067 .long 16777216 # 1067 .long 16777216 # 0x1000000
1068 .long 4 1068 .long 4
1069 .long 16777248 # 1069 .long 16777248 # 0x1000020
1070 ...... 1070 ......
1071 .byte 0 # 1071 .byte 0 # string offset=0
1072 .ascii ".text" # 1072 .ascii ".text" # string offset=1
1073 .byte 0 1073 .byte 0
1074 .ascii "/home/yhs/tmp-pahole/t2. 1074 .ascii "/home/yhs/tmp-pahole/t2.c" # string offset=7
1075 .byte 0 1075 .byte 0
1076 .ascii "int main() { return 0; } 1076 .ascii "int main() { return 0; }" # string offset=33
1077 .byte 0 1077 .byte 0
1078 .ascii "int test() { return 0; } 1078 .ascii "int test() { return 0; }" # string offset=58
1079 .byte 0 1079 .byte 0
1080 .ascii "int" # 1080 .ascii "int" # string offset=83
1081 ...... 1081 ......
1082 .section .BTF.ext,"",@prog 1082 .section .BTF.ext,"",@progbits
1083 .short 60319 # 1083 .short 60319 # 0xeb9f
1084 .byte 1 1084 .byte 1
1085 .byte 0 1085 .byte 0
1086 .long 24 1086 .long 24
1087 .long 0 1087 .long 0
1088 .long 28 1088 .long 28
1089 .long 28 1089 .long 28
1090 .long 44 1090 .long 44
1091 .long 8 # 1091 .long 8 # FuncInfo
1092 .long 1 # 1092 .long 1 # FuncInfo section string offset=1
1093 .long 2 1093 .long 2
1094 .long .Lfunc_begin0 1094 .long .Lfunc_begin0
1095 .long 3 1095 .long 3
1096 .long .Lfunc_begin1 1096 .long .Lfunc_begin1
1097 .long 5 1097 .long 5
1098 .long 16 # 1098 .long 16 # LineInfo
1099 .long 1 # 1099 .long 1 # LineInfo section string offset=1
1100 .long 2 1100 .long 2
1101 .long .Ltmp0 1101 .long .Ltmp0
1102 .long 7 1102 .long 7
1103 .long 33 1103 .long 33
1104 .long 7182 # 1104 .long 7182 # Line 7 Col 14
1105 .long .Ltmp3 1105 .long .Ltmp3
1106 .long 7 1106 .long 7
1107 .long 58 1107 .long 58
1108 .long 8206 # 1108 .long 8206 # Line 8 Col 14
1109 1109
1110 7. Testing 1110 7. Testing
1111 ========== 1111 ==========
1112 1112
1113 The kernel BPF selftest `tools/testing/selfte 1113 The kernel BPF selftest `tools/testing/selftests/bpf/prog_tests/btf.c`_
1114 provides an extensive set of BTF-related test 1114 provides an extensive set of BTF-related tests.
1115 1115
1116 .. Links 1116 .. Links
1117 .. _tools/testing/selftests/bpf/prog_tests/bt 1117 .. _tools/testing/selftests/bpf/prog_tests/btf.c:
1118 https://git.kernel.org/pub/scm/linux/kerne 1118 https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/tools/testing/selftests/bpf/prog_tests/btf.c
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.