1BPFTOOL-PROG(8) BPFTOOL-PROG(8)
2
3
4
6 bpftool-prog - tool for inspection and simple manipulation of eBPF
7 progs
8
10 bpftool [OPTIONS] prog COMMAND
11
12 OPTIONS := { { -j | --json } [{ -p | --pretty }] | { -f | --bpffs }
13 }
14
15 COMMANDS := { show | list | dump xlated | dump jited | pin | load |
16 loadall | help }
17
19 bpftool prog { show | list } [PROG]
20 bpftool prog dump xlated PROG [{file FILE | opcodes | visual | linum}]
21 bpftool prog dump jited PROG [{file FILE | opcodes | linum}]
22 bpftool prog pin PROG FILE
23 bpftool prog { load | loadall } OBJ PATH [type TYPE] [map {idx IDX | name NAME} MAP] [dev NAME] [pinmaps MAP_DIR]
24 bpftool prog attach PROG ATTACH_TYPE [MAP]
25 bpftool prog detach PROG ATTACH_TYPE [MAP]
26 bpftool prog tracelog
27 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]
28 bpftool prog help
29
30 MAP := { id MAP_ID | pinned FILE }
31 PROG := { id PROG_ID | pinned FILE | tag PROG_TAG }
32 TYPE := {
33 socket | kprobe | kretprobe | classifier | action |
34 tracepoint | raw_tracepoint | xdp | perf_event | cgroup/skb |
35 cgroup/sock | cgroup/dev | lwt_in | lwt_out | lwt_xmit |
36 lwt_seg6local | sockops | sk_skb | sk_msg | lirc_mode2 |
37 cgroup/bind4 | cgroup/bind6 | cgroup/post_bind4 | cgroup/post_bind6 |
38 cgroup/connect4 | cgroup/connect6 | cgroup/sendmsg4 | cgroup/sendmsg6 |
39 cgroup/recvmsg4 | cgroup/recvmsg6 | cgroup/sysctl |
40 cgroup/getsockopt | cgroup/setsockopt
41 }
42 ATTACH_TYPE := {
43 msg_verdict | stream_verdict | stream_parser | flow_dissector
44 }
45
46
48 bpftool prog { show | list } [PROG]
49 Show information about loaded programs. If PROG is specified
50 show information only about given program, otherwise list all
51 programs currently loaded on the system.
52
53 Output will start with program ID followed by program type
54 and zero or more named attributes (depending on kernel ver‐
55 sion).
56
57 Since Linux 5.1 the kernel can collect statistics on BPF pro‐
58 grams (such as the total time spent running the program, and
59 the number of times it was run). If available, bpftool shows
60 such statistics. However, the kernel does not collect them by
61 defaults, as it slightly impacts performance on each program
62 run. Activation or deactivation of the feature is performed
63 via the kernel.bpf_stats_enabled sysctl knob.
64
65 bpftool prog dump xlated PROG [{ file FILE | opcodes | visual |
66 linum }]
67 Dump eBPF instructions of the program from the kernel. By
68 default, eBPF will be disassembled and printed to standard
69 output in human-readable format. In this case, opcodes con‐
70 trols if raw opcodes should be printed as well.
71
72 If file is specified, the binary image will instead be writ‐
73 ten to FILE.
74
75 If visual is specified, control flow graph (CFG) will be
76 built instead, and eBPF instructions will be presented with
77 CFG in DOT format, on standard output.
78
79 If the prog has line_info available, the source line will be
80 displayed by default. If linum is specified, the filename,
81 line number and line column will also be displayed on top of
82 the source line.
83
84 bpftool prog dump jited PROG [{ file FILE | opcodes | linum }]
85 Dump jited image (host machine code) of the program. If FILE
86 is specified image will be written to a file, otherwise it
87 will be disassembled and printed to stdout.
88
89 opcodes controls if raw opcodes will be printed.
90
91 If the prog has line_info available, the source line will be
92 displayed by default. If linum is specified, the filename,
93 line number and line column will also be displayed on top of
94 the source line.
95
96 bpftool prog pin PROG FILE
97 Pin program PROG as FILE.
98
99 Note: FILE must be located in bpffs mount. It must not con‐
100 tain a dot character ('.'), which is reserved for future
101 extensions of bpffs.
102
103 bpftool prog { load | loadall } OBJ PATH [type TYPE] [map {idx IDX |
104 name NAME} MAP] [dev NAME] [pinmaps MAP_DIR]
105 Load bpf program(s) from binary OBJ and pin as PATH. bpftool
106 prog load pins only the first program from the OBJ as PATH.
107 bpftool prog loadall pins all programs from the OBJ under
108 PATH directory. type is optional, if not specified program
109 type will be inferred from section names. By default bpftool
110 will create new maps as declared in the ELF object being
111 loaded. map parameter allows for the reuse of existing maps.
112 It can be specified multiple times, each time for a different
113 map. IDX refers to index of the map to be replaced in the
114 ELF file counting from 0, while NAME allows to replace a map
115 by name. MAP specifies the map to use, referring to it by id
116 or through a pinned file. If dev NAME is specified program
117 will be loaded onto given networking device (offload).
118 Optional pinmaps argument can be provided to pin all maps
119 under MAP_DIR directory.
120
121 Note: PATH must be located in bpffs mount. It must not con‐
122 tain a dot character ('.'), which is reserved for future
123 extensions of bpffs.
124
125 bpftool prog attach PROG ATTACH_TYPE [MAP]
126 Attach bpf program PROG (with type specified by ATTACH_TYPE).
127 Most ATTACH_TYPEs require a MAP parameter, with the exception
128 of flow_dissector which is attached to current networking
129 name space.
130
131 bpftool prog detach PROG ATTACH_TYPE [MAP]
132 Detach bpf program PROG (with type specified by ATTACH_TYPE).
133 Most ATTACH_TYPEs require a MAP parameter, with the exception
134 of flow_dissector which is detached from the current network‐
135 ing name space.
136
137 bpftool prog tracelog
138 Dump the trace pipe of the system to the console (stdout).
139 Hit <Ctrl+C> to stop printing. BPF programs can write to this
140 trace pipe at runtime with the bpf_trace_printk() helper.
141 This should be used only for debugging purposes. For stream‐
142 ing data from BPF programs to user space, one can use perf
143 events (see also bpftool-map(8)).
144
145 bpftool prog run PROG data_in FILE [data_out FILE [data_size_out L]]
146 [ctx_in FILE [ctx_out FILE [ctx_size_out M]]] [repeat N]
147 Run BPF program PROG in the kernel testing infrastructure for
148 BPF, meaning that the program works on the data and context
149 provided by the user, and not on actual packets or monitored
150 functions etc. Return value and duration for the test run are
151 printed out to the console.
152
153 Input data is read from the FILE passed with data_in. If
154 this FILE is "-", input data is read from standard input.
155 Input context, if any, is read from FILE passed with ctx_in.
156 Again, "-" can be used to read from standard input, but only
157 if standard input is not already in use for input data. If a
158 FILE is passed with data_out, output data is written to that
159 file. Similarly, output context is written to the FILE passed
160 with ctx_out. For both output flows, "-" can be used to print
161 to the standard output (as plain text, or JSON if relevant
162 option was passed). If output keywords are omitted, output
163 data and context are discarded. Keywords data_size_out and
164 ctx_size_out are used to pass the size (in bytes) for the
165 output buffers to the kernel, although the default of 32 kB
166 should be more than enough for most cases.
167
168 Keyword repeat is used to indicate the number of consecutive
169 runs to perform. Note that output data and context printed to
170 files correspond to the last of those runs. The duration
171 printed out at the end of the runs is an average over all
172 runs performed by the command.
173
174 Not all program types support test run. Among those which do,
175 not all of them can take the ctx_in/ctx_out arguments.
176 bpftool does not perform checks on program types.
177
178 bpftool prog help
179 Print short help message.
180
182 -h, --help
183 Print short generic help message (similar to bpftool help).
184
185 -V, --version
186 Print version number (similar to bpftool version).
187
188 -j, --json
189 Generate JSON output. For commands that cannot produce JSON,
190 this option has no effect.
191
192 -p, --pretty
193 Generate human-readable JSON output. Implies -j.
194
195 -f, --bpffs
196 When showing BPF programs, show file names of pinned pro‐
197 grams.
198
199 -m, --mapcompat
200 Allow loading maps with unknown map definitions.
201
202 -n, --nomount
203 Do not automatically attempt to mount any virtual file system
204 (such as tracefs or BPF virtual file system) when necessary.
205
206 -d, --debug
207 Print all logs available, even debug-level information. This
208 includes logs from libbpf as well as from the verifier, when
209 attempting to load programs.
210
212 # bpftool prog show
213
214 10: xdp name some_prog tag 005a3d2123620c8b gpl run_time_ns 81632 run_cnt 10
215 loaded_at 2017-09-29T20:11:00+0000 uid 0
216 xlated 528B jited 370B memlock 4096B map_ids 10
217
218 # bpftool --json --pretty prog show
219
220 [{
221 "id": 10,
222 "type": "xdp",
223 "tag": "005a3d2123620c8b",
224 "gpl_compatible": true,
225 "run_time_ns": 81632,
226 "run_cnt": 10,
227 "loaded_at": 1506715860,
228 "uid": 0,
229 "bytes_xlated": 528,
230 "jited": true,
231 "bytes_jited": 370,
232 "bytes_memlock": 4096,
233 "map_ids": [10
234 ]
235 }
236 ]
237
238 # bpftool prog dump xlated id 10 file /tmp/t
239 # ls -l /tmp/t
240
241
242 -rw------- 1 root root 560 Jul 22 01:42 /tmp/t
243
244 # bpftool prog dump jited tag 005a3d2123620c8b
245
246 0: push %rbp
247 1: mov %rsp,%rbp
248 2: sub $0x228,%rsp
249 3: sub $0x28,%rbp
250 4: mov %rbx,0x0(%rbp)
251
252 # mount -t bpf none /sys/fs/bpf/
253 # bpftool prog pin id 10 /sys/fs/bpf/prog
254 # bpftool prog load ./my_prog.o /sys/fs/bpf/prog2
255 # ls -l /sys/fs/bpf/
256
257
258 -rw------- 1 root root 0 Jul 22 01:43 prog
259 -rw------- 1 root root 0 Jul 22 01:44 prog2
260
261 # bpftool prog dump jited pinned /sys/fs/bpf/prog opcodes
262
263 0: push %rbp
264 55
265 1: mov %rsp,%rbp
266 48 89 e5
267 4: sub $0x228,%rsp
268 48 81 ec 28 02 00 00
269 b: sub $0x28,%rbp
270 48 83 ed 28
271 f: mov %rbx,0x0(%rbp)
272 48 89 5d 00
273
274 # bpftool prog load xdp1_kern.o /sys/fs/bpf/xdp1 type xdp map name rxcnt id 7
275 # bpftool prog show pinned /sys/fs/bpf/xdp1
276
277
278 9: xdp name xdp_prog1 tag 539ec6ce11b52f98 gpl
279 loaded_at 2018-06-25T16:17:31-0700 uid 0
280 xlated 488B jited 336B memlock 4096B map_ids 7
281
282 # rm /sys/fs/bpf/xdp1
283
285 bpf(2), bpf-helpers(7), bpftool(8), bpftool-map(8),
286 bpftool-cgroup(8), bpftool-feature(8), bpftool-net(8),
287 bpftool-perf(8), bpftool-btf(8)
288
289
290
291
292 BPFTOOL-PROG(8)