4 -------------------------------------------------------------------------------
5 tool for inspection and simple manipulation of eBPF progs
6 -------------------------------------------------------------------------------
13 **bpftool** [*OPTIONS*] **prog** *COMMAND*
15 *OPTIONS* := { { **-j** | **--json** } [{ **-p** | **--pretty** }] | { **-f** | **--bpffs** } }
18 { **show** | **list** | **dump xlated** | **dump jited** | **pin** | **load**
19 | **loadall** | **help** }
24 | **bpftool** **prog** { **show** | **list** } [*PROG*]
25 | **bpftool** **prog dump xlated** *PROG* [{**file** *FILE* | **opcodes** | **visual** | **linum**}]
26 | **bpftool** **prog dump jited** *PROG* [{**file** *FILE* | **opcodes** | **linum**}]
27 | **bpftool** **prog pin** *PROG* *FILE*
28 | **bpftool** **prog** { **load** | **loadall** } *OBJ* *PATH* [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** *NAME*] [**pinmaps** *MAP_DIR*]
29 | **bpftool** **prog attach** *PROG* *ATTACH_TYPE* [*MAP*]
30 | **bpftool** **prog detach** *PROG* *ATTACH_TYPE* [*MAP*]
31 | **bpftool** **prog tracelog**
32 | **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*]
33 | **bpftool** **prog profile** *PROG* [**duration** *DURATION*] *METRICs*
34 | **bpftool** **prog help**
36 | *MAP* := { **id** *MAP_ID* | **pinned** *FILE* }
37 | *PROG* := { **id** *PROG_ID* | **pinned** *FILE* | **tag** *PROG_TAG* | **name** *PROG_NAME* }
39 | **socket** | **kprobe** | **kretprobe** | **classifier** | **action** |
40 | **tracepoint** | **raw_tracepoint** | **xdp** | **perf_event** | **cgroup/skb** |
41 | **cgroup/sock** | **cgroup/dev** | **lwt_in** | **lwt_out** | **lwt_xmit** |
42 | **lwt_seg6local** | **sockops** | **sk_skb** | **sk_msg** | **lirc_mode2** |
43 | **cgroup/bind4** | **cgroup/bind6** | **cgroup/post_bind4** | **cgroup/post_bind6** |
44 | **cgroup/connect4** | **cgroup/connect6** | **cgroup/getpeername4** | **cgroup/getpeername6** |
45 | **cgroup/getsockname4** | **cgroup/getsockname6** | **cgroup/sendmsg4** | **cgroup/sendmsg6** |
46 | **cgroup/recvmsg4** | **cgroup/recvmsg6** | **cgroup/sysctl** |
47 | **cgroup/getsockopt** | **cgroup/setsockopt** |
48 | **struct_ops** | **fentry** | **fexit** | **freplace**
51 | **msg_verdict** | **stream_verdict** | **stream_parser** | **flow_dissector**
54 | **cycles** | **instructions** | **l1d_loads** | **llc_misses**
60 **bpftool prog { show | list }** [*PROG*]
61 Show information about loaded programs. If *PROG* is
62 specified show information only about given programs,
63 otherwise list all programs currently loaded on the system.
64 In case of **tag** or **name**, *PROG* may match several
65 programs which will all be shown.
67 Output will start with program ID followed by program type and
68 zero or more named attributes (depending on kernel version).
70 Since Linux 5.1 the kernel can collect statistics on BPF
71 programs (such as the total time spent running the program,
72 and the number of times it was run). If available, bpftool
73 shows such statistics. However, the kernel does not collect
74 them by defaults, as it slightly impacts performance on each
75 program run. Activation or deactivation of the feature is
76 performed via the **kernel.bpf_stats_enabled** sysctl knob.
78 **bpftool prog dump xlated** *PROG* [{ **file** *FILE* | **opcodes** | **visual** | **linum** }]
79 Dump eBPF instructions of the programs from the kernel. By
80 default, eBPF will be disassembled and printed to standard
81 output in human-readable format. In this case, **opcodes**
82 controls if raw opcodes should be printed as well.
84 In case of **tag** or **name**, *PROG* may match several
85 programs which will all be dumped. However, if **file** or
86 **visual** is specified, *PROG* must match a single program.
88 If **file** is specified, the binary image will instead be
91 If **visual** is specified, control flow graph (CFG) will be
92 built instead, and eBPF instructions will be presented with
93 CFG in DOT format, on standard output.
95 If the programs have line_info available, the source line will
96 be displayed by default. If **linum** is specified,
97 the filename, line number and line column will also be
98 displayed on top of the source line.
100 **bpftool prog dump jited** *PROG* [{ **file** *FILE* | **opcodes** | **linum** }]
101 Dump jited image (host machine code) of the program.
103 If *FILE* is specified image will be written to a file,
104 otherwise it will be disassembled and printed to stdout.
105 *PROG* must match a single program when **file** is specified.
107 **opcodes** controls if raw opcodes will be printed.
109 If the prog has line_info available, the source line will
110 be displayed by default. If **linum** is specified,
111 the filename, line number and line column will also be
112 displayed on top of the source line.
114 **bpftool prog pin** *PROG* *FILE*
115 Pin program *PROG* as *FILE*.
117 Note: *FILE* must be located in *bpffs* mount. It must not
118 contain a dot character ('.'), which is reserved for future
119 extensions of *bpffs*.
121 **bpftool prog { load | loadall }** *OBJ* *PATH* [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** *NAME*] [**pinmaps** *MAP_DIR*]
122 Load bpf program(s) from binary *OBJ* and pin as *PATH*.
123 **bpftool prog load** pins only the first program from the
124 *OBJ* as *PATH*. **bpftool prog loadall** pins all programs
125 from the *OBJ* under *PATH* directory.
126 **type** is optional, if not specified program type will be
127 inferred from section names.
128 By default bpftool will create new maps as declared in the ELF
129 object being loaded. **map** parameter allows for the reuse
130 of existing maps. It can be specified multiple times, each
131 time for a different map. *IDX* refers to index of the map
132 to be replaced in the ELF file counting from 0, while *NAME*
133 allows to replace a map by name. *MAP* specifies the map to
134 use, referring to it by **id** or through a **pinned** file.
135 If **dev** *NAME* is specified program will be loaded onto
136 given networking device (offload).
137 Optional **pinmaps** argument can be provided to pin all
138 maps under *MAP_DIR* directory.
140 Note: *PATH* must be located in *bpffs* mount. It must not
141 contain a dot character ('.'), which is reserved for future
142 extensions of *bpffs*.
144 **bpftool prog attach** *PROG* *ATTACH_TYPE* [*MAP*]
145 Attach bpf program *PROG* (with type specified by
146 *ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP*
147 parameter, with the exception of *flow_dissector* which is
148 attached to current networking name space.
150 **bpftool prog detach** *PROG* *ATTACH_TYPE* [*MAP*]
151 Detach bpf program *PROG* (with type specified by
152 *ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP*
153 parameter, with the exception of *flow_dissector* which is
154 detached from the current networking name space.
156 **bpftool prog tracelog**
157 Dump the trace pipe of the system to the console (stdout).
158 Hit <Ctrl+C> to stop printing. BPF programs can write to this
159 trace pipe at runtime with the **bpf_trace_printk**\ () helper.
160 This should be used only for debugging purposes. For
161 streaming data from BPF programs to user space, one can use
162 perf events (see also **bpftool-map**\ (8)).
164 **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*]
165 Run BPF program *PROG* in the kernel testing infrastructure
166 for BPF, meaning that the program works on the data and
167 context provided by the user, and not on actual packets or
168 monitored functions etc. Return value and duration for the
169 test run are printed out to the console.
171 Input data is read from the *FILE* passed with **data_in**.
172 If this *FILE* is "**-**", input data is read from standard
173 input. Input context, if any, is read from *FILE* passed with
174 **ctx_in**. Again, "**-**" can be used to read from standard
175 input, but only if standard input is not already in use for
176 input data. If a *FILE* is passed with **data_out**, output
177 data is written to that file. Similarly, output context is
178 written to the *FILE* passed with **ctx_out**. For both
179 output flows, "**-**" can be used to print to the standard
180 output (as plain text, or JSON if relevant option was
181 passed). If output keywords are omitted, output data and
182 context are discarded. Keywords **data_size_out** and
183 **ctx_size_out** are used to pass the size (in bytes) for the
184 output buffers to the kernel, although the default of 32 kB
185 should be more than enough for most cases.
187 Keyword **repeat** is used to indicate the number of
188 consecutive runs to perform. Note that output data and
189 context printed to files correspond to the last of those
190 runs. The duration printed out at the end of the runs is an
191 average over all runs performed by the command.
193 Not all program types support test run. Among those which do,
194 not all of them can take the **ctx_in**/**ctx_out**
195 arguments. bpftool does not perform checks on program types.
197 **bpftool prog profile** *PROG* [**duration** *DURATION*] *METRICs*
198 Profile *METRICs* for bpf program *PROG* for *DURATION*
199 seconds or until user hits <Ctrl+C>. *DURATION* is optional.
200 If *DURATION* is not specified, the profiling will run up to
201 **UINT_MAX** seconds.
203 **bpftool prog help**
204 Print short help message.
209 Print short generic help message (similar to **bpftool help**).
212 Print version number (similar to **bpftool version**).
215 Generate JSON output. For commands that cannot produce JSON, this
216 option has no effect.
219 Generate human-readable JSON output. Implies **-j**.
222 When showing BPF programs, show file names of pinned
226 Allow loading maps with unknown map definitions.
229 Do not automatically attempt to mount any virtual file system
230 (such as tracefs or BPF virtual file system) when necessary.
233 Print all logs available, even debug-level information. This
234 includes logs from libbpf as well as from the verifier, when
235 attempting to load programs.
239 **# bpftool prog show**
243 10: xdp name some_prog tag 005a3d2123620c8b gpl run_time_ns 81632 run_cnt 10
244 loaded_at 2017-09-29T20:11:00+0000 uid 0
245 xlated 528B jited 370B memlock 4096B map_ids 10
247 **# bpftool --json --pretty prog show**
254 "tag": "005a3d2123620c8b",
255 "gpl_compatible": true,
256 "run_time_ns": 81632,
258 "loaded_at": 1506715860,
263 "bytes_memlock": 4096,
270 | **# bpftool prog dump xlated id 10 file /tmp/t**
275 -rw------- 1 root root 560 Jul 22 01:42 /tmp/t
277 **# bpftool prog dump jited tag 005a3d2123620c8b**
285 4: mov %rbx,0x0(%rbp)
288 | **# mount -t bpf none /sys/fs/bpf/**
289 | **# bpftool prog pin id 10 /sys/fs/bpf/prog**
290 | **# bpftool prog load ./my_prog.o /sys/fs/bpf/prog2**
291 | **# ls -l /sys/fs/bpf/**
295 -rw------- 1 root root 0 Jul 22 01:43 prog
296 -rw------- 1 root root 0 Jul 22 01:44 prog2
298 **# bpftool prog dump jited pinned /sys/fs/bpf/prog opcodes**
310 f: mov %rbx,0x0(%rbp)
314 | **# bpftool prog load xdp1_kern.o /sys/fs/bpf/xdp1 type xdp map name rxcnt id 7**
315 | **# bpftool prog show pinned /sys/fs/bpf/xdp1**
319 9: xdp name xdp_prog1 tag 539ec6ce11b52f98 gpl
320 loaded_at 2018-06-25T16:17:31-0700 uid 0
321 xlated 488B jited 336B memlock 4096B map_ids 7
323 **# rm /sys/fs/bpf/xdp1**
326 | **# bpftool prog profile id 337 duration 10 cycles instructions llc_misses**
331 40176203 cycles (83.05%)
332 42518139 instructions # 1.06 insns per cycle (83.39%)
333 123 llc_misses # 2.89 LLC misses per million insns (83.15%)
338 **bpf-helpers**\ (7),
340 **bpftool-btf**\ (8),
341 **bpftool-cgroup**\ (8),
342 **bpftool-feature**\ (8),
343 **bpftool-gen**\ (8),
344 **bpftool-iter**\ (8),
345 **bpftool-link**\ (8),
346 **bpftool-map**\ (8),
347 **bpftool-net**\ (8),
348 **bpftool-perf**\ (8),
349 **bpftool-struct_ops**\ (8)