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