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