1BPFTOOL-FEATURE(8) BPFTOOL-FEATURE(8)
2
6 bpftool-feature - tool for inspection of eBPF-related parameters for
7 Linux kernel or net device
8
10 bpftool [OPTIONS] feature COMMAND
11 OPTIONS := { { -j | --json } [{ -p | --pretty }] | { -d | --debug }
13 }
14 COMMANDS := { probe | help }
16
18 bpftool feature probe [COMPONENT] [full] [unprivileged] [macros [prefix PREFIX]]
19 bpftool feature list_builtins GROUP
20 bpftool feature help
21 COMPONENT := { kernel | dev NAME }
23 GROUP := { prog_types | map_types | attach_types | link_types | helpers }
24
27 bpftool feature probe [kernel] [full] [macros [prefix PREFIX]]
28 Probe the running kernel and dump a number of eBPF-related
29 parameters, such as availability of the bpf() system call,
30 JIT status, eBPF program types availability, eBPF helper
31 functions availability, and more.
32 By default, bpftool does not run probes for
34 bpf_probe_write_user() and bpf_trace_printk() helpers which
35 print warnings to kernel logs. To enable them and run all
36 probes, the full keyword should be used.
37 If the macros keyword (but not the -j option) is passed, a
39 subset of the output is dumped as a list of #define macros
40 that are ready to be included in a C header file, for exam‐
41 ple. If, additionally, prefix is used to define a PREFIX, the
42 provided string will be used as a prefix to the names of the
43 macros: this can be used to avoid conflicts on macro names
44 when including the output of this command as a header file.
45 Keyword kernel can be omitted. If no probe target is speci‐
47 fied, probing the kernel is the default behaviour.
48 When the unprivileged keyword is used, bpftool will dump only
50 the features available to a user who does not have the
51 CAP_SYS_ADMIN capability set. The features available in that
52 case usually represent a small subset of the parameters sup‐
53 ported by the system. Unprivileged users MUST use the unpriv‐
54 ileged keyword: This is to avoid misdetection if bpftool is
55 inadvertently run as non-root, for example. This keyword is
56 unavailable if bpftool was compiled without libcap.
57 bpftool feature probe dev NAME [full] [macros [prefix PREFIX]]
59 Probe network device for supported eBPF features and dump re‐
60 sults to the console.
61 The keywords full, macros and prefix have the same role as
63 when probing the kernel.
64 bpftool feature list_builtins GROUP
66 List items known to bpftool. These can be BPF program types
67 (prog_types), BPF map types (map_types), attach types (at‐
68 tach_types), link types (link_types), or BPF helper functions
69 (helpers). The command does not probe the system, but simply
70 lists the elements that bpftool knows from compilation time,
71 as provided from libbpf (for all object types) or from the
72 BPF UAPI header (list of helpers). This can be used in
73 scripts to iterate over BPF types or helpers.
74 bpftool feature help
76 Print short help message.
77
79 -h, --help
80 Print short help message (similar to bpftool help).
81 -V, --version
83 Print bpftool's version number (similar to bpftool version),
84 the number of the libbpf version in use, and optional fea‐
85 tures that were included when bpftool was compiled. Optional
86 features include linking against LLVM or libbfd to provide
87 the disassembler for JIT-ted programs (bpftool prog dump
88 jited) and usage of BPF skeletons (some features like bpftool
89 prog profile or showing pids associated to BPF objects may
90 rely on it).
91 -j, --json
93 Generate JSON output. For commands that cannot produce JSON,
94 this option has no effect.
95 -p, --pretty
97 Generate human-readable JSON output. Implies -j.
98 -d, --debug
100 Print all logs available, even debug-level information. This
101 includes logs from libbpf as well as from the verifier, when
102 attempting to load programs.
103
105 bpf(2), bpf-helpers(7), bpftool(8), bpftool-btf(8),
106 bpftool-cgroup(8), bpftool-gen(8), bpftool-iter(8), bpftool-link(8),
107 bpftool-map(8), bpftool-net(8), bpftool-perf(8), bpftool-prog(8),
108 bpftool-struct_ops(8)
109 BPFTOOL-FEATURE(8)