1 .. SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
6 -------------------------------------------------------------------------------
7 tool for inspection and simple manipulation of eBPF maps
8 -------------------------------------------------------------------------------
12 .. include:: substitutions.rst
17 **bpftool** [*OPTIONS*] **map** *COMMAND*
19 *OPTIONS* := { |COMMON_OPTIONS| | { **-f** | **--bpffs** } | { **-n** | **--nomount** } }
22 { **show** | **list** | **create** | **dump** | **update** | **lookup** | **getnext** |
23 **delete** | **pin** | **help** }
28 | **bpftool** **map** { **show** | **list** } [*MAP*]
29 | **bpftool** **map create** *FILE* **type** *TYPE* **key** *KEY_SIZE* **value** *VALUE_SIZE* \
30 | **entries** *MAX_ENTRIES* **name** *NAME* [**flags** *FLAGS*] [**inner_map** *MAP*] \
31 | [**offload_dev** *NAME*]
32 | **bpftool** **map dump** *MAP*
33 | **bpftool** **map update** *MAP* [**key** *DATA*] [**value** *VALUE*] [*UPDATE_FLAGS*]
34 | **bpftool** **map lookup** *MAP* [**key** *DATA*]
35 | **bpftool** **map getnext** *MAP* [**key** *DATA*]
36 | **bpftool** **map delete** *MAP* **key** *DATA*
37 | **bpftool** **map pin** *MAP* *FILE*
38 | **bpftool** **map event_pipe** *MAP* [**cpu** *N* **index** *M*]
39 | **bpftool** **map peek** *MAP*
40 | **bpftool** **map push** *MAP* **value** *VALUE*
41 | **bpftool** **map pop** *MAP*
42 | **bpftool** **map enqueue** *MAP* **value** *VALUE*
43 | **bpftool** **map dequeue** *MAP*
44 | **bpftool** **map freeze** *MAP*
45 | **bpftool** **map help**
47 | *MAP* := { **id** *MAP_ID* | **pinned** *FILE* | **name** *MAP_NAME* }
48 | *DATA* := { [**hex**] *BYTES* }
49 | *PROG* := { **id** *PROG_ID* | **pinned** *FILE* | **tag** *PROG_TAG* | **name** *PROG_NAME* }
50 | *VALUE* := { *DATA* | *MAP* | *PROG* }
51 | *UPDATE_FLAGS* := { **any** | **exist** | **noexist** }
52 | *TYPE* := { **hash** | **array** | **prog_array** | **perf_event_array** | **percpu_hash**
53 | | **percpu_array** | **stack_trace** | **cgroup_array** | **lru_hash**
54 | | **lru_percpu_hash** | **lpm_trie** | **array_of_maps** | **hash_of_maps**
55 | | **devmap** | **devmap_hash** | **sockmap** | **cpumap** | **xskmap** | **sockhash**
56 | | **cgroup_storage** | **reuseport_sockarray** | **percpu_cgroup_storage**
57 | | **queue** | **stack** | **sk_storage** | **struct_ops** | **ringbuf** | **inode_storage**
58 | | **task_storage** | **bloom_filter** | **user_ringbuf** | **cgrp_storage** | **arena** }
62 bpftool map { show | list } [*MAP*]
63 Show information about loaded maps. If *MAP* is specified show information
64 only about given maps, otherwise list all maps currently loaded on the
65 system. In case of **name**, *MAP* may match several maps which will all
68 Output will start with map ID followed by map type and zero or more named
69 attributes (depending on kernel version).
71 Since Linux 5.8 bpftool is able to discover information about processes
72 that hold open file descriptors (FDs) against BPF maps. On such kernels
73 bpftool will automatically emit this information as well.
75 bpftool map create *FILE* type *TYPE* key *KEY_SIZE* value *VALUE_SIZE* entries *MAX_ENTRIES* name *NAME* [flags *FLAGS*] [inner_map *MAP*] [offload_dev *NAME*]
76 Create a new map with given parameters and pin it to *bpffs* as *FILE*.
78 *FLAGS* should be an integer which is the combination of desired flags,
79 e.g. 1024 for **BPF_F_MMAPABLE** (see bpf.h UAPI header for existing
82 To create maps of type array-of-maps or hash-of-maps, the **inner_map**
83 keyword must be used to pass an inner map. The kernel needs it to collect
84 metadata related to the inner maps that the new map will work with.
86 Keyword **offload_dev** expects a network interface name, and is used to
87 request hardware offload for the map.
89 bpftool map dump *MAP*
90 Dump all entries in a given *MAP*. In case of **name**, *MAP* may match
91 several maps which will all be dumped.
93 bpftool map update *MAP* [key *DATA*] [value *VALUE*] [*UPDATE_FLAGS*]
94 Update map entry for a given *KEY*.
96 *UPDATE_FLAGS* can be one of: **any** update existing entry or add if
97 doesn't exit; **exist** update only if entry already exists; **noexist**
98 update only if entry doesn't exist.
100 If the **hex** keyword is provided in front of the bytes sequence, the
101 bytes are parsed as hexadecimal values, even if no "0x" prefix is added. If
102 the keyword is not provided, then the bytes are parsed as decimal values,
103 unless a "0x" prefix (for hexadecimal) or a "0" prefix (for octal) is
106 bpftool map lookup *MAP* [key *DATA*]
107 Lookup **key** in the map.
109 bpftool map getnext *MAP* [key *DATA*]
110 Get next key. If *key* is not specified, get first key.
112 bpftool map delete *MAP* key *DATA*
113 Remove entry from the map.
115 bpftool map pin *MAP* *FILE*
116 Pin map *MAP* as *FILE*.
118 Note: *FILE* must be located in *bpffs* mount. It must not contain a dot
119 character ('.'), which is reserved for future extensions of *bpffs*.
121 bpftool map event_pipe *MAP* [cpu *N* index *M*]
122 Read events from a **BPF_MAP_TYPE_PERF_EVENT_ARRAY** map.
124 Install perf rings into a perf event array map and dump output of any
125 **bpf_perf_event_output**\ () call in the kernel. By default read the
126 number of CPUs on the system and install perf ring for each CPU in the
127 corresponding index in the array.
129 If **cpu** and **index** are specified, install perf ring for given **cpu**
130 at **index** in the array (single ring).
132 Note that installing a perf ring into an array will silently replace any
133 existing ring. Any other application will stop receiving events if it
134 installed its rings earlier.
136 bpftool map peek *MAP*
137 Peek next value in the queue or stack.
139 bpftool map push *MAP* value *VALUE*
140 Push *VALUE* onto the stack.
142 bpftool map pop *MAP*
143 Pop and print value from the stack.
145 bpftool map enqueue *MAP* value *VALUE*
146 Enqueue *VALUE* into the queue.
148 bpftool map dequeue *MAP*
149 Dequeue and print value from the queue.
151 bpftool map freeze *MAP*
152 Freeze the map as read-only from user space. Entries from a frozen map can
153 not longer be updated or deleted with the **bpf**\ () system call. This
154 operation is not reversible, and the map remains immutable from user space
155 until its destruction. However, read and write permissions for BPF programs
156 to the map remain unchanged.
159 Print short help message.
163 .. include:: common_options.rst
166 Show file names of pinned maps.
169 Do not automatically attempt to mount any virtual file system (such as
170 tracefs or BPF virtual file system) when necessary.
174 **# bpftool map show**
178 10: hash name some_map flags 0x0
179 key 4B value 8B max_entries 2048 memlock 167936B
182 The following three commands are equivalent:
185 | **# bpftool map update id 10 key hex 20 c4 b7 00 value hex 0f ff ff ab 01 02 03 4c**
186 | **# bpftool map update id 10 key 0x20 0xc4 0xb7 0x00 value 0x0f 0xff 0xff 0xab 0x01 0x02 0x03 0x4c**
187 | **# bpftool map update id 10 key 32 196 183 0 value 15 255 255 171 1 2 3 76**
189 **# bpftool map lookup id 10 key 0 1 2 3**
193 key: 00 01 02 03 value: 00 01 02 03 04 05 06 07
196 **# bpftool map dump id 10**
200 key: 00 01 02 03 value: 00 01 02 03 04 05 06 07
201 key: 0d 00 07 00 value: 02 00 00 00 01 02 03 04
204 **# bpftool map getnext id 10 key 0 1 2 3**
214 | **# mount -t bpf none /sys/fs/bpf/**
215 | **# bpftool map pin id 10 /sys/fs/bpf/map**
216 | **# bpftool map del pinned /sys/fs/bpf/map key 13 00 07 00**
218 Note that map update can also be used in order to change the program references
219 hold by a program array map. This can be used, for example, to change the
220 programs used for tail-call jumps at runtime, without having to reload the
221 entry-point program. Below is an example for this use case: we load a program
222 defining a prog array map, and with a main function that contains a tail call
223 to other programs that can be used either to "process" packets or to "debug"
224 processing. Note that the prog array map MUST be pinned into the BPF virtual
225 file system for the map update to work successfully, as kernel flushes prog
226 array maps when they have no more references from user space (and the update
227 would be lost as soon as bpftool exits).
230 | **# bpftool prog loadall tail_calls.o /sys/fs/bpf/foo type xdp**
231 | **# bpftool prog --bpffs**
235 545: xdp name main_func tag 674b4b5597193dc3 gpl
236 loaded_at 2018-12-12T15:02:58+0000 uid 0
237 xlated 240B jited 257B memlock 4096B map_ids 294
238 pinned /sys/fs/bpf/foo/xdp
239 546: xdp name bpf_func_process tag e369a529024751fc gpl
240 loaded_at 2018-12-12T15:02:58+0000 uid 0
241 xlated 200B jited 164B memlock 4096B
242 pinned /sys/fs/bpf/foo/process
243 547: xdp name bpf_func_debug tag 0b597868bc7f0976 gpl
244 loaded_at 2018-12-12T15:02:58+0000 uid 0
245 xlated 200B jited 164B memlock 4096B
246 pinned /sys/fs/bpf/foo/debug
252 294: prog_array name jmp_table flags 0x0
253 key 4B value 4B max_entries 1 memlock 4096B
254 owner_prog_type xdp owner jited
257 | **# bpftool map pin id 294 /sys/fs/bpf/bar**
258 | **# bpftool map dump pinned /sys/fs/bpf/bar**
265 | **# bpftool map update pinned /sys/fs/bpf/bar key 0 0 0 0 value pinned /sys/fs/bpf/foo/debug**
266 | **# bpftool map dump pinned /sys/fs/bpf/bar**
270 key: 00 00 00 00 value: 22 02 00 00