1BPFTOOL-CGROUP(8) BPFTOOL-CGROUP(8)
2
3
4
6 bpftool-cgroup - tool for inspection and simple manipulation of eBPF
7 progs
8
10 bpftool [OPTIONS] cgroup COMMAND
11
12 OPTIONS := { { -j | --json } [{ -p | --pretty }] | { -d | --debug }
13 | { -l | --legacy } | { -f | --bpffs } }
14
15 COMMANDS := { show | list | tree | attach | detach | help }
16
18 bpftool cgroup { show | list } CGROUP [effective]
19 bpftool cgroup tree [CGROUP_ROOT] [effective]
20 bpftool cgroup attach CGROUP ATTACH_TYPE PROG [ATTACH_FLAGS]
21 bpftool cgroup detach CGROUP ATTACH_TYPE PROG
22 bpftool cgroup help
23
24 PROG := { id PROG_ID | pinned FILE | tag PROG_TAG }
25 ATTACH_TYPE := { cgroup_inet_ingress | cgroup_inet_egress |
26 cgroup_inet_sock_create | cgroup_sock_ops |
27 cgroup_device | cgroup_inet4_bind | cgroup_inet6_bind |
28 cgroup_inet4_post_bind | cgroup_inet6_post_bind |
29 cgroup_inet4_connect | cgroup_inet6_connect |
30 cgroup_inet4_getpeername | cgroup_inet6_getpeername |
31 cgroup_inet4_getsockname | cgroup_inet6_getsockname |
32 cgroup_udp4_sendmsg | cgroup_udp6_sendmsg |
33 cgroup_udp4_recvmsg | cgroup_udp6_recvmsg |
34 cgroup_sysctl | cgroup_getsockopt | cgroup_setsockopt |
35 cgroup_inet_sock_release }
36 ATTACH_FLAGS := { multi | override }
37
38
40 bpftool cgroup { show | list } CGROUP [effective]
41 List all programs attached to the cgroup CGROUP.
42
43 Output will start with program ID followed by attach type,
44 attach flags and program name.
45
46 If effective is specified retrieve effective programs that
47 will execute for events within a cgroup. This includes inher‐
48 ited along with attached ones.
49
50 bpftool cgroup tree [CGROUP_ROOT] [effective]
51 Iterate over all cgroups in CGROUP_ROOT and list all attached
52 programs. If CGROUP_ROOT is not specified, bpftool uses
53 cgroup v2 mountpoint.
54
55 The output is similar to the output of cgroup show/list com‐
56 mands: it starts with absolute cgroup path, followed by pro‐
57 gram ID, attach type, attach flags and program name.
58
59 If effective is specified retrieve effective programs that
60 will execute for events within a cgroup. This includes inher‐
61 ited along with attached ones.
62
63 bpftool cgroup attach CGROUP ATTACH_TYPE PROG [ATTACH_FLAGS]
64 Attach program PROG to the cgroup CGROUP with attach type AT‐
65 TACH_TYPE and optional ATTACH_FLAGS.
66
67 ATTACH_FLAGS can be one of: override if a sub-cgroup installs
68 some bpf program, the program in this cgroup yields to
69 sub-cgroup program; multi if a sub-cgroup installs some bpf
70 program, that cgroup program gets run in addition to the pro‐
71 gram in this cgroup.
72
73 Only one program is allowed to be attached to a cgroup with
74 no attach flags or the override flag. Attaching another pro‐
75 gram will release old program and attach the new one.
76
77 Multiple programs are allowed to be attached to a cgroup with
78 multi. They are executed in FIFO order (those that were at‐
79 tached first, run first).
80
81 Non-default ATTACH_FLAGS are supported by kernel version 4.14
82 and later.
83
84 ATTACH_TYPE can be on of: ingress ingress path of the inet
85 socket (since 4.10); egress egress path of the inet socket
86 (since 4.10); sock_create opening of an inet socket (since
87 4.10); sock_ops various socket operations (since 4.12); de‐
88 vice device access (since 4.15); bind4 call to bind(2) for an
89 inet4 socket (since 4.17); bind6 call to bind(2) for an inet6
90 socket (since 4.17); post_bind4 return from bind(2) for an
91 inet4 socket (since 4.17); post_bind6 return from bind(2) for
92 an inet6 socket (since 4.17); connect4 call to connect(2) for
93 an inet4 socket (since 4.17); connect6 call to connect(2) for
94 an inet6 socket (since 4.17); sendmsg4 call to sendto(2),
95 sendmsg(2), sendmmsg(2) for an unconnected udp4 socket (since
96 4.18); sendmsg6 call to sendto(2), sendmsg(2), sendmmsg(2)
97 for an unconnected udp6 socket (since 4.18); recvmsg4 call to
98 recvfrom(2), recvmsg(2), recvmmsg(2) for an unconnected udp4
99 socket (since 5.2); recvmsg6 call to recvfrom(2), recvmsg(2),
100 recvmmsg(2) for an unconnected udp6 socket (since 5.2);
101 sysctl sysctl access (since 5.2); getsockopt call to getsock‐
102 opt (since 5.3); setsockopt call to setsockopt (since 5.3);
103 getpeername4 call to getpeername(2) for an inet4 socket
104 (since 5.8); getpeername6 call to getpeername(2) for an inet6
105 socket (since 5.8); getsockname4 call to getsockname(2) for
106 an inet4 socket (since 5.8); getsockname6 call to getsock‐
107 name(2) for an inet6 socket (since 5.8). sock_release clos‐
108 ing an userspace inet socket (since 5.9).
109
110 bpftool cgroup detach CGROUP ATTACH_TYPE PROG
111 Detach PROG from the cgroup CGROUP and attach type AT‐
112 TACH_TYPE.
113
114 bpftool prog help
115 Print short help message.
116
118 -h, --help
119 Print short help message (similar to bpftool help).
120
121 -V, --version
122 Print bpftool's version number (similar to bpftool version),
123 the number of the libbpf version in use, and optional fea‐
124 tures that were included when bpftool was compiled. Optional
125 features include linking against libbfd to provide the disas‐
126 sembler for JIT-ted programs (bpftool prog dump jited) and
127 usage of BPF skeletons (some features like bpftool prog pro‐
128 file or showing pids associated to BPF objects may rely on
129 it).
130
131 -j, --json
132 Generate JSON output. For commands that cannot produce JSON,
133 this option has no effect.
134
135 -p, --pretty
136 Generate human-readable JSON output. Implies -j.
137
138 -d, --debug
139 Print all logs available, even debug-level information. This
140 includes logs from libbpf as well as from the verifier, when
141 attempting to load programs.
142
143 -l, --legacy
144 Use legacy libbpf mode which has more relaxed BPF program re‐
145 quirements. By default, bpftool has more strict requirements
146 about section names, changes pinning logic and doesn't sup‐
147 port some of the older non-BTF map declarations.
148
149 See
150 https://github.com/libbpf/libbpf/wiki/Libbpf:-the-road-to-v1.0
151 for details.
152
153 -f, --bpffs
154 Show file names of pinned programs.
155
157 # mount -t bpf none /sys/fs/bpf/
158 # mkdir /sys/fs/cgroup/test.slice
159 # bpftool prog load ./device_cgroup.o /sys/fs/bpf/prog
160 # bpftool cgroup attach /sys/fs/cgroup/test.slice/ device id 1 allow_multi
161
162
163 # bpftool cgroup list /sys/fs/cgroup/test.slice/
164
165 ID AttachType AttachFlags Name
166 1 device allow_multi bpf_prog1
167
168 # bpftool cgroup detach /sys/fs/cgroup/test.slice/ device id 1
169 # bpftool cgroup list /sys/fs/cgroup/test.slice/
170
171
172 ID AttachType AttachFlags Name
173
175 bpf(2), bpf-helpers(7), bpftool(8), bpftool-btf(8), bpftool-fea‐
176 ture(8), bpftool-gen(8), bpftool-iter(8), bpftool-link(8),
177 bpftool-map(8), bpftool-net(8), bpftool-perf(8), bpftool-prog(8),
178 bpftool-struct_ops(8)
179
180
181
182
183 BPFTOOL-CGROUP(8)