bpftool-prog.rst (14374B)
1.. SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) 2 3================ 4bpftool-prog 5================ 6------------------------------------------------------------------------------- 7tool for inspection and simple manipulation of eBPF progs 8------------------------------------------------------------------------------- 9 10:Manual section: 8 11 12.. include:: substitutions.rst 13 14SYNOPSIS 15======== 16 17 **bpftool** [*OPTIONS*] **prog** *COMMAND* 18 19 *OPTIONS* := { |COMMON_OPTIONS| | 20 { **-f** | **--bpffs** } | { **-m** | **--mapcompat** } | { **-n** | **--nomount** } | 21 { **-L** | **--use-loader** } } 22 23 *COMMANDS* := 24 { **show** | **list** | **dump xlated** | **dump jited** | **pin** | **load** | 25 **loadall** | **help** } 26 27PROG COMMANDS 28============= 29 30| **bpftool** **prog** { **show** | **list** } [*PROG*] 31| **bpftool** **prog dump xlated** *PROG* [{**file** *FILE* | **opcodes** | **visual** | **linum**}] 32| **bpftool** **prog dump jited** *PROG* [{**file** *FILE* | **opcodes** | **linum**}] 33| **bpftool** **prog pin** *PROG* *FILE* 34| **bpftool** **prog** { **load** | **loadall** } *OBJ* *PATH* [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** *NAME*] [**pinmaps** *MAP_DIR*] 35| **bpftool** **prog attach** *PROG* *ATTACH_TYPE* [*MAP*] 36| **bpftool** **prog detach** *PROG* *ATTACH_TYPE* [*MAP*] 37| **bpftool** **prog tracelog** 38| **bpftool** **prog run** *PROG* **data_in** *FILE* [**data_out** *FILE* [**data_size_out** *L*]] [**ctx_in** *FILE* [**ctx_out** *FILE* [**ctx_size_out** *M*]]] [**repeat** *N*] 39| **bpftool** **prog profile** *PROG* [**duration** *DURATION*] *METRICs* 40| **bpftool** **prog help** 41| 42| *MAP* := { **id** *MAP_ID* | **pinned** *FILE* } 43| *PROG* := { **id** *PROG_ID* | **pinned** *FILE* | **tag** *PROG_TAG* | **name** *PROG_NAME* } 44| *TYPE* := { 45| **socket** | **kprobe** | **kretprobe** | **classifier** | **action** | 46| **tracepoint** | **raw_tracepoint** | **xdp** | **perf_event** | **cgroup/skb** | 47| **cgroup/sock** | **cgroup/dev** | **lwt_in** | **lwt_out** | **lwt_xmit** | 48| **lwt_seg6local** | **sockops** | **sk_skb** | **sk_msg** | **lirc_mode2** | 49| **cgroup/bind4** | **cgroup/bind6** | **cgroup/post_bind4** | **cgroup/post_bind6** | 50| **cgroup/connect4** | **cgroup/connect6** | **cgroup/getpeername4** | **cgroup/getpeername6** | 51| **cgroup/getsockname4** | **cgroup/getsockname6** | **cgroup/sendmsg4** | **cgroup/sendmsg6** | 52| **cgroup/recvmsg4** | **cgroup/recvmsg6** | **cgroup/sysctl** | 53| **cgroup/getsockopt** | **cgroup/setsockopt** | **cgroup/sock_release** | 54| **struct_ops** | **fentry** | **fexit** | **freplace** | **sk_lookup** 55| } 56| *ATTACH_TYPE* := { 57| **msg_verdict** | **skb_verdict** | **stream_verdict** | **stream_parser** | **flow_dissector** 58| } 59| *METRICs* := { 60| **cycles** | **instructions** | **l1d_loads** | **llc_misses** | 61| **itlb_misses** | **dtlb_misses** 62| } 63 64 65DESCRIPTION 66=========== 67 **bpftool prog { show | list }** [*PROG*] 68 Show information about loaded programs. If *PROG* is 69 specified show information only about given programs, 70 otherwise list all programs currently loaded on the system. 71 In case of **tag** or **name**, *PROG* may match several 72 programs which will all be shown. 73 74 Output will start with program ID followed by program type and 75 zero or more named attributes (depending on kernel version). 76 77 Since Linux 5.1 the kernel can collect statistics on BPF 78 programs (such as the total time spent running the program, 79 and the number of times it was run). If available, bpftool 80 shows such statistics. However, the kernel does not collect 81 them by defaults, as it slightly impacts performance on each 82 program run. Activation or deactivation of the feature is 83 performed via the **kernel.bpf_stats_enabled** sysctl knob. 84 85 Since Linux 5.8 bpftool is able to discover information about 86 processes that hold open file descriptors (FDs) against BPF 87 programs. On such kernels bpftool will automatically emit this 88 information as well. 89 90 **bpftool prog dump xlated** *PROG* [{ **file** *FILE* | **opcodes** | **visual** | **linum** }] 91 Dump eBPF instructions of the programs from the kernel. By 92 default, eBPF will be disassembled and printed to standard 93 output in human-readable format. In this case, **opcodes** 94 controls if raw opcodes should be printed as well. 95 96 In case of **tag** or **name**, *PROG* may match several 97 programs which will all be dumped. However, if **file** or 98 **visual** is specified, *PROG* must match a single program. 99 100 If **file** is specified, the binary image will instead be 101 written to *FILE*. 102 103 If **visual** is specified, control flow graph (CFG) will be 104 built instead, and eBPF instructions will be presented with 105 CFG in DOT format, on standard output. 106 107 If the programs have line_info available, the source line will 108 be displayed by default. If **linum** is specified, 109 the filename, line number and line column will also be 110 displayed on top of the source line. 111 112 **bpftool prog dump jited** *PROG* [{ **file** *FILE* | **opcodes** | **linum** }] 113 Dump jited image (host machine code) of the program. 114 115 If *FILE* is specified image will be written to a file, 116 otherwise it will be disassembled and printed to stdout. 117 *PROG* must match a single program when **file** is specified. 118 119 **opcodes** controls if raw opcodes will be printed. 120 121 If the prog has line_info available, the source line will 122 be displayed by default. If **linum** is specified, 123 the filename, line number and line column will also be 124 displayed on top of the source line. 125 126 **bpftool prog pin** *PROG* *FILE* 127 Pin program *PROG* as *FILE*. 128 129 Note: *FILE* must be located in *bpffs* mount. It must not 130 contain a dot character ('.'), which is reserved for future 131 extensions of *bpffs*. 132 133 **bpftool prog { load | loadall }** *OBJ* *PATH* [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** *NAME*] [**pinmaps** *MAP_DIR*] 134 Load bpf program(s) from binary *OBJ* and pin as *PATH*. 135 **bpftool prog load** pins only the first program from the 136 *OBJ* as *PATH*. **bpftool prog loadall** pins all programs 137 from the *OBJ* under *PATH* directory. 138 **type** is optional, if not specified program type will be 139 inferred from section names. 140 By default bpftool will create new maps as declared in the ELF 141 object being loaded. **map** parameter allows for the reuse 142 of existing maps. It can be specified multiple times, each 143 time for a different map. *IDX* refers to index of the map 144 to be replaced in the ELF file counting from 0, while *NAME* 145 allows to replace a map by name. *MAP* specifies the map to 146 use, referring to it by **id** or through a **pinned** file. 147 If **dev** *NAME* is specified program will be loaded onto 148 given networking device (offload). 149 Optional **pinmaps** argument can be provided to pin all 150 maps under *MAP_DIR* directory. 151 152 Note: *PATH* must be located in *bpffs* mount. It must not 153 contain a dot character ('.'), which is reserved for future 154 extensions of *bpffs*. 155 156 **bpftool prog attach** *PROG* *ATTACH_TYPE* [*MAP*] 157 Attach bpf program *PROG* (with type specified by 158 *ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP* 159 parameter, with the exception of *flow_dissector* which is 160 attached to current networking name space. 161 162 **bpftool prog detach** *PROG* *ATTACH_TYPE* [*MAP*] 163 Detach bpf program *PROG* (with type specified by 164 *ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP* 165 parameter, with the exception of *flow_dissector* which is 166 detached from the current networking name space. 167 168 **bpftool prog tracelog** 169 Dump the trace pipe of the system to the console (stdout). 170 Hit <Ctrl+C> to stop printing. BPF programs can write to this 171 trace pipe at runtime with the **bpf_trace_printk**\ () helper. 172 This should be used only for debugging purposes. For 173 streaming data from BPF programs to user space, one can use 174 perf events (see also **bpftool-map**\ (8)). 175 176 **bpftool prog run** *PROG* **data_in** *FILE* [**data_out** *FILE* [**data_size_out** *L*]] [**ctx_in** *FILE* [**ctx_out** *FILE* [**ctx_size_out** *M*]]] [**repeat** *N*] 177 Run BPF program *PROG* in the kernel testing infrastructure 178 for BPF, meaning that the program works on the data and 179 context provided by the user, and not on actual packets or 180 monitored functions etc. Return value and duration for the 181 test run are printed out to the console. 182 183 Input data is read from the *FILE* passed with **data_in**. 184 If this *FILE* is "**-**", input data is read from standard 185 input. Input context, if any, is read from *FILE* passed with 186 **ctx_in**. Again, "**-**" can be used to read from standard 187 input, but only if standard input is not already in use for 188 input data. If a *FILE* is passed with **data_out**, output 189 data is written to that file. Similarly, output context is 190 written to the *FILE* passed with **ctx_out**. For both 191 output flows, "**-**" can be used to print to the standard 192 output (as plain text, or JSON if relevant option was 193 passed). If output keywords are omitted, output data and 194 context are discarded. Keywords **data_size_out** and 195 **ctx_size_out** are used to pass the size (in bytes) for the 196 output buffers to the kernel, although the default of 32 kB 197 should be more than enough for most cases. 198 199 Keyword **repeat** is used to indicate the number of 200 consecutive runs to perform. Note that output data and 201 context printed to files correspond to the last of those 202 runs. The duration printed out at the end of the runs is an 203 average over all runs performed by the command. 204 205 Not all program types support test run. Among those which do, 206 not all of them can take the **ctx_in**/**ctx_out** 207 arguments. bpftool does not perform checks on program types. 208 209 **bpftool prog profile** *PROG* [**duration** *DURATION*] *METRICs* 210 Profile *METRICs* for bpf program *PROG* for *DURATION* 211 seconds or until user hits <Ctrl+C>. *DURATION* is optional. 212 If *DURATION* is not specified, the profiling will run up to 213 **UINT_MAX** seconds. 214 215 **bpftool prog help** 216 Print short help message. 217 218OPTIONS 219======= 220 .. include:: common_options.rst 221 222 -f, --bpffs 223 When showing BPF programs, show file names of pinned 224 programs. 225 226 -m, --mapcompat 227 Allow loading maps with unknown map definitions. 228 229 -n, --nomount 230 Do not automatically attempt to mount any virtual file system 231 (such as tracefs or BPF virtual file system) when necessary. 232 233 -L, --use-loader 234 Load program as a "loader" program. This is useful to debug 235 the generation of such programs. When this option is in 236 use, bpftool attempts to load the programs from the object 237 file into the kernel, but does not pin them (therefore, the 238 *PATH* must not be provided). 239 240 When combined with the **-d**\ \|\ **--debug** option, 241 additional debug messages are generated, and the execution 242 of the loader program will use the **bpf_trace_printk**\ () 243 helper to log each step of loading BTF, creating the maps, 244 and loading the programs (see **bpftool prog tracelog** as 245 a way to dump those messages). 246 247EXAMPLES 248======== 249**# bpftool prog show** 250 251:: 252 253 10: xdp name some_prog tag 005a3d2123620c8b gpl run_time_ns 81632 run_cnt 10 254 loaded_at 2017-09-29T20:11:00+0000 uid 0 255 xlated 528B jited 370B memlock 4096B map_ids 10 256 pids systemd(1) 257 258**# bpftool --json --pretty prog show** 259 260:: 261 262 [{ 263 "id": 10, 264 "type": "xdp", 265 "tag": "005a3d2123620c8b", 266 "gpl_compatible": true, 267 "run_time_ns": 81632, 268 "run_cnt": 10, 269 "loaded_at": 1506715860, 270 "uid": 0, 271 "bytes_xlated": 528, 272 "jited": true, 273 "bytes_jited": 370, 274 "bytes_memlock": 4096, 275 "map_ids": [10 276 ], 277 "pids": [{ 278 "pid": 1, 279 "comm": "systemd" 280 } 281 ] 282 } 283 ] 284 285| 286| **# bpftool prog dump xlated id 10 file /tmp/t** 287| **$ ls -l /tmp/t** 288 289:: 290 291 -rw------- 1 root root 560 Jul 22 01:42 /tmp/t 292 293**# bpftool prog dump jited tag 005a3d2123620c8b** 294 295:: 296 297 0: push %rbp 298 1: mov %rsp,%rbp 299 2: sub $0x228,%rsp 300 3: sub $0x28,%rbp 301 4: mov %rbx,0x0(%rbp) 302 303| 304| **# mount -t bpf none /sys/fs/bpf/** 305| **# bpftool prog pin id 10 /sys/fs/bpf/prog** 306| **# bpftool prog load ./my_prog.o /sys/fs/bpf/prog2** 307| **# ls -l /sys/fs/bpf/** 308 309:: 310 311 -rw------- 1 root root 0 Jul 22 01:43 prog 312 -rw------- 1 root root 0 Jul 22 01:44 prog2 313 314**# bpftool prog dump jited pinned /sys/fs/bpf/prog opcodes** 315 316:: 317 318 0: push %rbp 319 55 320 1: mov %rsp,%rbp 321 48 89 e5 322 4: sub $0x228,%rsp 323 48 81 ec 28 02 00 00 324 b: sub $0x28,%rbp 325 48 83 ed 28 326 f: mov %rbx,0x0(%rbp) 327 48 89 5d 00 328 329| 330| **# bpftool prog load xdp1_kern.o /sys/fs/bpf/xdp1 type xdp map name rxcnt id 7** 331| **# bpftool prog show pinned /sys/fs/bpf/xdp1** 332 333:: 334 335 9: xdp name xdp_prog1 tag 539ec6ce11b52f98 gpl 336 loaded_at 2018-06-25T16:17:31-0700 uid 0 337 xlated 488B jited 336B memlock 4096B map_ids 7 338 339**# rm /sys/fs/bpf/xdp1** 340 341| 342| **# bpftool prog profile id 337 duration 10 cycles instructions llc_misses** 343 344:: 345 346 51397 run_cnt 347 40176203 cycles (83.05%) 348 42518139 instructions # 1.06 insns per cycle (83.39%) 349 123 llc_misses # 2.89 LLC misses per million insns (83.15%) 350 351| 352| Output below is for the trace logs. 353| Run in separate terminals: 354| **# bpftool prog tracelog** 355| **# bpftool prog load -L -d file.o** 356 357:: 358 359 bpftool-620059 [004] d... 2634685.517903: bpf_trace_printk: btf_load size 665 r=5 360 bpftool-620059 [004] d... 2634685.517912: bpf_trace_printk: map_create sample_map idx 0 type 2 value_size 4 value_btf_id 0 r=6 361 bpftool-620059 [004] d... 2634685.517997: bpf_trace_printk: prog_load sample insn_cnt 13 r=7 362 bpftool-620059 [004] d... 2634685.517999: bpf_trace_printk: close(5) = 0