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**] [**linum**] [**visual**] }]
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*] [{ **offload_dev** | **xdpmeta_dev** } *NAME*] [**pinmaps** *MAP_DIR*] [**autoattach**]
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|		**sk_msg_verdict** | **sk_skb_verdict** | **sk_skb_stream_verdict** |
58|		**sk_skb_stream_parser** | **flow_dissector**
59|	}
60|	*METRICs* := {
61|		**cycles** | **instructions** | **l1d_loads** | **llc_misses** |
62|		**itlb_misses** | **dtlb_misses**
63|	}
64
65
66DESCRIPTION
67===========
68	**bpftool prog { show | list }** [*PROG*]
69		  Show information about loaded programs.  If *PROG* is
70		  specified show information only about given programs,
71		  otherwise list all programs currently loaded on the system.
72		  In case of **tag** or **name**, *PROG* may match several
73		  programs which will all be shown.
74
75		  Output will start with program ID followed by program type and
76		  zero or more named attributes (depending on kernel version).
77
78		  Since Linux 5.1 the kernel can collect statistics on BPF
79		  programs (such as the total time spent running the program,
80		  and the number of times it was run). If available, bpftool
81		  shows such statistics. However, the kernel does not collect
82		  them by defaults, as it slightly impacts performance on each
83		  program run. Activation or deactivation of the feature is
84		  performed via the **kernel.bpf_stats_enabled** sysctl knob.
85
86		  Since Linux 5.8 bpftool is able to discover information about
87		  processes that hold open file descriptors (FDs) against BPF
88		  programs. On such kernels bpftool will automatically emit this
89		  information as well.
90
91	**bpftool prog dump xlated** *PROG* [{ **file** *FILE* | [**opcodes**] [**linum**] [**visual**] }]
92		  Dump eBPF instructions of the programs from the kernel. By
93		  default, eBPF will be disassembled and printed to standard
94		  output in human-readable format. In this case, **opcodes**
95		  controls if raw opcodes should be printed as well.
96
97		  In case of **tag** or **name**, *PROG* may match several
98		  programs which will all be dumped.  However, if **file** or
99		  **visual** is specified, *PROG* must match a single program.
100
101		  If **file** is specified, the binary image will instead be
102		  written to *FILE*.
103
104		  If **visual** is specified, control flow graph (CFG) will be
105		  built instead, and eBPF instructions will be presented with
106		  CFG in DOT format, on standard output.
107
108		  If the programs have line_info available, the source line will
109		  be displayed.  If **linum** is specified, the filename, line
110		  number and line column will also be displayed.
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.  If **linum** is specified, the filename, line
123		  number and line column will also be displayed.
124
125	**bpftool prog pin** *PROG* *FILE*
126		  Pin program *PROG* as *FILE*.
127
128		  Note: *FILE* must be located in *bpffs* mount. It must not
129		  contain a dot character ('.'), which is reserved for future
130		  extensions of *bpffs*.
131
132	**bpftool prog { load | loadall }** *OBJ* *PATH* [**type** *TYPE*] [**map** { **idx** *IDX* | **name** *NAME* } *MAP*] [{ **offload_dev** | **xdpmeta_dev** } *NAME*] [**pinmaps** *MAP_DIR*] [**autoattach**]
133		  Load bpf program(s) from binary *OBJ* and pin as *PATH*.
134		  **bpftool prog load** pins only the first program from the
135		  *OBJ* as *PATH*. **bpftool prog loadall** pins all programs
136		  from the *OBJ* under *PATH* directory.
137		  **type** is optional, if not specified program type will be
138		  inferred from section names.
139		  By default bpftool will create new maps as declared in the ELF
140		  object being loaded.  **map** parameter allows for the reuse
141		  of existing maps.  It can be specified multiple times, each
142		  time for a different map.  *IDX* refers to index of the map
143		  to be replaced in the ELF file counting from 0, while *NAME*
144		  allows to replace a map by name.  *MAP* specifies the map to
145		  use, referring to it by **id** or through a **pinned** file.
146		  If **offload_dev** *NAME* is specified program will be loaded
147		  onto given networking device (offload).
148		  If **xdpmeta_dev** *NAME* is specified program will become
149		  device-bound without offloading, this facilitates access
150		  to XDP metadata.
151		  Optional **pinmaps** argument can be provided to pin all
152		  maps under *MAP_DIR* directory.
153
154		  If **autoattach** is specified program will be attached
155		  before pin. In that case, only the link (representing the
156		  program attached to its hook) is pinned, not the program as
157		  such, so the path won't show in **bpftool prog show -f**,
158		  only show in **bpftool link show -f**. Also, this only works
159		  when bpftool (libbpf) is able to infer all necessary
160		  information from the object file, in particular, it's not
161		  supported for all program types. If a program does not
162		  support autoattach, bpftool falls back to regular pinning
163		  for that program instead.
164
165		  Note: *PATH* must be located in *bpffs* mount. It must not
166		  contain a dot character ('.'), which is reserved for future
167		  extensions of *bpffs*.
168
169	**bpftool prog attach** *PROG* *ATTACH_TYPE* [*MAP*]
170		  Attach bpf program *PROG* (with type specified by
171		  *ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP*
172		  parameter, with the exception of *flow_dissector* which is
173		  attached to current networking name space.
174
175	**bpftool prog detach** *PROG* *ATTACH_TYPE* [*MAP*]
176		  Detach bpf program *PROG* (with type specified by
177		  *ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP*
178		  parameter, with the exception of *flow_dissector* which is
179		  detached from the current networking name space.
180
181	**bpftool prog tracelog**
182		  Dump the trace pipe of the system to the console (stdout).
183		  Hit <Ctrl+C> to stop printing. BPF programs can write to this
184		  trace pipe at runtime with the **bpf_trace_printk**\ () helper.
185		  This should be used only for debugging purposes. For
186		  streaming data from BPF programs to user space, one can use
187		  perf events (see also **bpftool-map**\ (8)).
188
189	**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*]
190		  Run BPF program *PROG* in the kernel testing infrastructure
191		  for BPF, meaning that the program works on the data and
192		  context provided by the user, and not on actual packets or
193		  monitored functions etc. Return value and duration for the
194		  test run are printed out to the console.
195
196		  Input data is read from the *FILE* passed with **data_in**.
197		  If this *FILE* is "**-**", input data is read from standard
198		  input. Input context, if any, is read from *FILE* passed with
199		  **ctx_in**. Again, "**-**" can be used to read from standard
200		  input, but only if standard input is not already in use for
201		  input data. If a *FILE* is passed with **data_out**, output
202		  data is written to that file. Similarly, output context is
203		  written to the *FILE* passed with **ctx_out**. For both
204		  output flows, "**-**" can be used to print to the standard
205		  output (as plain text, or JSON if relevant option was
206		  passed). If output keywords are omitted, output data and
207		  context are discarded. Keywords **data_size_out** and
208		  **ctx_size_out** are used to pass the size (in bytes) for the
209		  output buffers to the kernel, although the default of 32 kB
210		  should be more than enough for most cases.
211
212		  Keyword **repeat** is used to indicate the number of
213		  consecutive runs to perform. Note that output data and
214		  context printed to files correspond to the last of those
215		  runs. The duration printed out at the end of the runs is an
216		  average over all runs performed by the command.
217
218		  Not all program types support test run. Among those which do,
219		  not all of them can take the **ctx_in**/**ctx_out**
220		  arguments. bpftool does not perform checks on program types.
221
222	**bpftool prog profile** *PROG* [**duration** *DURATION*] *METRICs*
223		  Profile *METRICs* for bpf program *PROG* for *DURATION*
224		  seconds or until user hits <Ctrl+C>. *DURATION* is optional.
225		  If *DURATION* is not specified, the profiling will run up to
226		  **UINT_MAX** seconds.
227
228	**bpftool prog help**
229		  Print short help message.
230
231OPTIONS
232=======
233	.. include:: common_options.rst
234
235	-f, --bpffs
236		  When showing BPF programs, show file names of pinned
237		  programs.
238
239	-m, --mapcompat
240		  Allow loading maps with unknown map definitions.
241
242	-n, --nomount
243		  Do not automatically attempt to mount any virtual file system
244		  (such as tracefs or BPF virtual file system) when necessary.
245
246	-L, --use-loader
247		  Load program as a "loader" program. This is useful to debug
248		  the generation of such programs. When this option is in
249		  use, bpftool attempts to load the programs from the object
250		  file into the kernel, but does not pin them (therefore, the
251		  *PATH* must not be provided).
252
253		  When combined with the **-d**\ \|\ **--debug** option,
254		  additional debug messages are generated, and the execution
255		  of the loader program will use the **bpf_trace_printk**\ ()
256		  helper to log each step of loading BTF, creating the maps,
257		  and loading the programs (see **bpftool prog tracelog** as
258		  a way to dump those messages).
259
260EXAMPLES
261========
262**# bpftool prog show**
263
264::
265
266    10: xdp  name some_prog  tag 005a3d2123620c8b  gpl run_time_ns 81632 run_cnt 10
267            loaded_at 2017-09-29T20:11:00+0000  uid 0
268            xlated 528B  jited 370B  memlock 4096B  map_ids 10
269            pids systemd(1)
270
271**# bpftool --json --pretty prog show**
272
273::
274
275    [{
276            "id": 10,
277            "type": "xdp",
278            "tag": "005a3d2123620c8b",
279            "gpl_compatible": true,
280            "run_time_ns": 81632,
281            "run_cnt": 10,
282            "loaded_at": 1506715860,
283            "uid": 0,
284            "bytes_xlated": 528,
285            "jited": true,
286            "bytes_jited": 370,
287            "bytes_memlock": 4096,
288            "map_ids": [10
289            ],
290            "pids": [{
291                    "pid": 1,
292                    "comm": "systemd"
293                }
294            ]
295        }
296    ]
297
298|
299| **# bpftool prog dump xlated id 10 file /tmp/t**
300| **$ ls -l /tmp/t**
301
302::
303
304    -rw------- 1 root root 560 Jul 22 01:42 /tmp/t
305
306**# bpftool prog dump jited tag 005a3d2123620c8b**
307
308::
309
310    0:   push   %rbp
311    1:   mov    %rsp,%rbp
312    2:   sub    $0x228,%rsp
313    3:   sub    $0x28,%rbp
314    4:   mov    %rbx,0x0(%rbp)
315
316|
317| **# mount -t bpf none /sys/fs/bpf/**
318| **# bpftool prog pin id 10 /sys/fs/bpf/prog**
319| **# bpftool prog load ./my_prog.o /sys/fs/bpf/prog2**
320| **# ls -l /sys/fs/bpf/**
321
322::
323
324    -rw------- 1 root root 0 Jul 22 01:43 prog
325    -rw------- 1 root root 0 Jul 22 01:44 prog2
326
327**# bpftool prog dump jited pinned /sys/fs/bpf/prog opcodes**
328
329::
330
331   0:   push   %rbp
332        55
333   1:   mov    %rsp,%rbp
334        48 89 e5
335   4:   sub    $0x228,%rsp
336        48 81 ec 28 02 00 00
337   b:   sub    $0x28,%rbp
338        48 83 ed 28
339   f:   mov    %rbx,0x0(%rbp)
340        48 89 5d 00
341
342|
343| **# bpftool prog load xdp1_kern.o /sys/fs/bpf/xdp1 type xdp map name rxcnt id 7**
344| **# bpftool prog show pinned /sys/fs/bpf/xdp1**
345
346::
347
348    9: xdp  name xdp_prog1  tag 539ec6ce11b52f98  gpl
349            loaded_at 2018-06-25T16:17:31-0700  uid 0
350            xlated 488B  jited 336B  memlock 4096B  map_ids 7
351
352**# rm /sys/fs/bpf/xdp1**
353
354|
355| **# bpftool prog profile id 337 duration 10 cycles instructions llc_misses**
356
357::
358
359         51397 run_cnt
360      40176203 cycles                                                 (83.05%)
361      42518139 instructions    #   1.06 insns per cycle               (83.39%)
362           123 llc_misses      #   2.89 LLC misses per million insns  (83.15%)
363
364|
365| Output below is for the trace logs.
366| Run in separate terminals:
367| **# bpftool prog tracelog**
368| **# bpftool prog load -L -d file.o**
369
370::
371
372    bpftool-620059  [004] d... 2634685.517903: bpf_trace_printk: btf_load size 665 r=5
373    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
374    bpftool-620059  [004] d... 2634685.517997: bpf_trace_printk: prog_load sample insn_cnt 13 r=7
375    bpftool-620059  [004] d... 2634685.517999: bpf_trace_printk: close(5) = 0
376