brintos

brintos / linux-shallow public Read only

0
0
Text · 14.7 KiB · d6304e0 Raw
354 lines · plain
1.. SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)2 3================4bpftool-prog5================6-------------------------------------------------------------------------------7tool for inspection and simple manipulation of eBPF progs8-------------------------------------------------------------------------------9 10:Manual section: 811 12.. include:: substitutions.rst13 14SYNOPSIS15========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 COMMANDS28=============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* | **name** *MAP_NAME* }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/connect_unix** |51|     **cgroup/getpeername4** | **cgroup/getpeername6** | **cgroup/getpeername_unix** |52|     **cgroup/getsockname4** | **cgroup/getsockname6** | **cgroup/getsockname_unix** |53|     **cgroup/sendmsg4** | **cgroup/sendmsg6** | **cgroup/sendmsg_unix** |54|     **cgroup/recvmsg4** | **cgroup/recvmsg6** | **cgroup/recvmsg_unix** | **cgroup/sysctl** |55|     **cgroup/getsockopt** | **cgroup/setsockopt** | **cgroup/sock_release** |56|     **struct_ops** | **fentry** | **fexit** | **freplace** | **sk_lookup**57| }58| *ATTACH_TYPE* := {59|     **sk_msg_verdict** | **sk_skb_verdict** | **sk_skb_stream_verdict** |60|     **sk_skb_stream_parser** | **flow_dissector**61| }62| *METRICs* := {63|     **cycles** | **instructions** | **l1d_loads** | **llc_misses** |64|     **itlb_misses** | **dtlb_misses**65| }66 67 68DESCRIPTION69===========70bpftool prog { show | list } [*PROG*]71    Show information about loaded programs.  If *PROG* is specified show72    information only about given programs, otherwise list all programs73    currently loaded on the system. In case of **tag** or **name**, *PROG* may74    match several programs which will all be shown.75 76    Output will start with program ID followed by program type and zero or more77    named attributes (depending on kernel version).78 79    Since Linux 5.1 the kernel can collect statistics on BPF programs (such as80    the total time spent running the program, and the number of times it was81    run). If available, bpftool shows such statistics. However, the kernel does82    not collect them by defaults, as it slightly impacts performance on each83    program run. Activation or deactivation of the feature is performed via the84    **kernel.bpf_stats_enabled** sysctl knob.85 86    Since Linux 5.8 bpftool is able to discover information about processes87    that hold open file descriptors (FDs) against BPF programs. On such kernels88    bpftool will automatically emit this information as well.89 90bpftool prog dump xlated *PROG* [{ file *FILE* | [opcodes] [linum] [visual] }]91    Dump eBPF instructions of the programs from the kernel. By default, eBPF92    will be disassembled and printed to standard output in human-readable93    format. In this case, **opcodes** controls if raw opcodes should be printed94    as well.95 96    In case of **tag** or **name**, *PROG* may match several programs which97    will all be dumped.  However, if **file** or **visual** is specified,98    *PROG* must match a single program.99 100    If **file** is specified, the binary image will instead be written to101    *FILE*.102 103    If **visual** is specified, control flow graph (CFG) will be built instead,104    and eBPF instructions will be presented with CFG in DOT format, on standard105    output.106 107    If the programs have line_info available, the source line will be108    displayed.  If **linum** is specified, the filename, line number and line109    column will also be displayed.110 111bpftool prog dump jited  *PROG* [{ file *FILE* | [opcodes] [linum] }]112    Dump jited image (host machine code) of the program.113 114    If *FILE* is specified image will be written to a file, otherwise it will115    be disassembled and printed to stdout. *PROG* must match a single program116    when **file** is specified.117 118    **opcodes** controls if raw opcodes will be printed.119 120    If the prog has line_info available, the source line will be displayed. If121    **linum** is specified, the filename, line number and line column will also122    be displayed.123 124bpftool prog pin *PROG* *FILE*125    Pin program *PROG* as *FILE*.126 127    Note: *FILE* must be located in *bpffs* mount. It must not contain a dot128    character ('.'), which is reserved for future extensions of *bpffs*.129 130bpftool prog { load | loadall } *OBJ* *PATH* [type *TYPE*] [map { idx *IDX* | name *NAME* } *MAP*] [{ offload_dev | xdpmeta_dev } *NAME*] [pinmaps *MAP_DIR*] [autoattach]131    Load bpf program(s) from binary *OBJ* and pin as *PATH*. **bpftool prog132    load** pins only the first program from the *OBJ* as *PATH*. **bpftool prog133    loadall** pins all programs from the *OBJ* under *PATH* directory. **type**134    is optional, if not specified program type will be inferred from section135    names. By default bpftool will create new maps as declared in the ELF136    object being loaded.  **map** parameter allows for the reuse of existing137    maps. It can be specified multiple times, each time for a different map.138    *IDX* refers to index of the map to be replaced in the ELF file counting139    from 0, while *NAME* allows to replace a map by name. *MAP* specifies the140    map to use, referring to it by **id** or through a **pinned** file. If141    **offload_dev** *NAME* is specified program will be loaded onto given142    networking device (offload). If **xdpmeta_dev** *NAME* is specified program143    will become device-bound without offloading, this facilitates access to XDP144    metadata. Optional **pinmaps** argument can be provided to pin all maps145    under *MAP_DIR* directory.146 147    If **autoattach** is specified program will be attached before pin. In that148    case, only the link (representing the program attached to its hook) is149    pinned, not the program as such, so the path won't show in **bpftool prog150    show -f**, only show in **bpftool link show -f**. Also, this only works151    when bpftool (libbpf) is able to infer all necessary information from the152    object file, in particular, it's not supported for all program types. If a153    program does not support autoattach, bpftool falls back to regular pinning154    for that program instead.155 156    Note: *PATH* must be located in *bpffs* mount. It must not contain a dot157    character ('.'), which is reserved for future extensions of *bpffs*.158 159bpftool prog attach *PROG* *ATTACH_TYPE* [*MAP*]160    Attach bpf program *PROG* (with type specified by *ATTACH_TYPE*). Most161    *ATTACH_TYPEs* require a *MAP* parameter, with the exception of162    *flow_dissector* which is attached to current networking name space.163 164bpftool prog detach *PROG* *ATTACH_TYPE* [*MAP*]165    Detach bpf program *PROG* (with type specified by *ATTACH_TYPE*). Most166    *ATTACH_TYPEs* require a *MAP* parameter, with the exception of167    *flow_dissector* which is detached from the current networking name space.168 169bpftool prog tracelog170    Dump the trace pipe of the system to the console (stdout). Hit <Ctrl+C> to171    stop printing. BPF programs can write to this trace pipe at runtime with172    the **bpf_trace_printk**\ () helper. This should be used only for debugging173    purposes. For streaming data from BPF programs to user space, one can use174    perf events (see also **bpftool-map**\ (8)).175 176bpftool 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*]177    Run BPF program *PROG* in the kernel testing infrastructure for BPF,178    meaning that the program works on the data and context provided by the179    user, and not on actual packets or monitored functions etc. Return value180    and duration for the test run are printed out to the console.181 182    Input data is read from the *FILE* passed with **data_in**. If this *FILE*183    is "**-**", input data is read from standard input. Input context, if any,184    is read from *FILE* passed with **ctx_in**. Again, "**-**" can be used to185    read from standard input, but only if standard input is not already in use186    for input data. If a *FILE* is passed with **data_out**, output data is187    written to that file. Similarly, output context is written to the *FILE*188    passed with **ctx_out**. For both output flows, "**-**" can be used to189    print to the standard output (as plain text, or JSON if relevant option was190    passed). If output keywords are omitted, output data and context are191    discarded. Keywords **data_size_out** and **ctx_size_out** are used to pass192    the size (in bytes) for the output buffers to the kernel, although the193    default of 32 kB should be more than enough for most cases.194 195    Keyword **repeat** is used to indicate the number of consecutive runs to196    perform. Note that output data and context printed to files correspond to197    the last of those runs. The duration printed out at the end of the runs is198    an average over all runs performed by the command.199 200    Not all program types support test run. Among those which do, not all of201    them can take the **ctx_in**/**ctx_out** arguments. bpftool does not202    perform checks on program types.203 204bpftool prog profile *PROG* [duration *DURATION*] *METRICs*205    Profile *METRICs* for bpf program *PROG* for *DURATION* seconds or until206    user hits <Ctrl+C>. *DURATION* is optional. If *DURATION* is not specified,207    the profiling will run up to **UINT_MAX** seconds.208 209bpftool prog help210    Print short help message.211 212OPTIONS213=======214.. include:: common_options.rst215 216-f, --bpffs217    When showing BPF programs, show file names of pinned programs.218 219-m, --mapcompat220    Allow loading maps with unknown map definitions.221 222-n, --nomount223    Do not automatically attempt to mount any virtual file system (such as224    tracefs or BPF virtual file system) when necessary.225 226-L, --use-loader227    Load program as a "loader" program. This is useful to debug the generation228    of such programs. When this option is in use, bpftool attempts to load the229    programs from the object file into the kernel, but does not pin them230    (therefore, the *PATH* must not be provided).231 232    When combined with the **-d**\ \|\ **--debug** option, additional debug233    messages are generated, and the execution of the loader program will use234    the **bpf_trace_printk**\ () helper to log each step of loading BTF,235    creating the maps, and loading the programs (see **bpftool prog tracelog**236    as a way to dump those messages).237 238EXAMPLES239========240**# bpftool prog show**241 242::243 244    10: xdp  name some_prog  tag 005a3d2123620c8b  gpl run_time_ns 81632 run_cnt 10245            loaded_at 2017-09-29T20:11:00+0000  uid 0246            xlated 528B  jited 370B  memlock 4096B  map_ids 10247            pids systemd(1)248 249**# bpftool --json --pretty prog show**250 251::252 253    [{254            "id": 10,255            "type": "xdp",256            "tag": "005a3d2123620c8b",257            "gpl_compatible": true,258            "run_time_ns": 81632,259            "run_cnt": 10,260            "loaded_at": 1506715860,261            "uid": 0,262            "bytes_xlated": 528,263            "jited": true,264            "bytes_jited": 370,265            "bytes_memlock": 4096,266            "map_ids": [10267            ],268            "pids": [{269                    "pid": 1,270                    "comm": "systemd"271                }272            ]273        }274    ]275 276|277| **# bpftool prog dump xlated id 10 file /tmp/t**278| **$ ls -l /tmp/t**279 280::281 282    -rw------- 1 root root 560 Jul 22 01:42 /tmp/t283 284**# bpftool prog dump jited tag 005a3d2123620c8b**285 286::287 288    0:   push   %rbp289    1:   mov    %rsp,%rbp290    2:   sub    $0x228,%rsp291    3:   sub    $0x28,%rbp292    4:   mov    %rbx,0x0(%rbp)293 294|295| **# mount -t bpf none /sys/fs/bpf/**296| **# bpftool prog pin id 10 /sys/fs/bpf/prog**297| **# bpftool prog load ./my_prog.o /sys/fs/bpf/prog2**298| **# ls -l /sys/fs/bpf/**299 300::301 302    -rw------- 1 root root 0 Jul 22 01:43 prog303    -rw------- 1 root root 0 Jul 22 01:44 prog2304 305**# bpftool prog dump jited pinned /sys/fs/bpf/prog opcodes**306 307::308 309   0:   push   %rbp310        55311   1:   mov    %rsp,%rbp312        48 89 e5313   4:   sub    $0x228,%rsp314        48 81 ec 28 02 00 00315   b:   sub    $0x28,%rbp316        48 83 ed 28317   f:   mov    %rbx,0x0(%rbp)318        48 89 5d 00319 320|321| **# bpftool prog load xdp1_kern.o /sys/fs/bpf/xdp1 type xdp map name rxcnt id 7**322| **# bpftool prog show pinned /sys/fs/bpf/xdp1**323 324::325 326    9: xdp  name xdp_prog1  tag 539ec6ce11b52f98  gpl327            loaded_at 2018-06-25T16:17:31-0700  uid 0328            xlated 488B  jited 336B  memlock 4096B  map_ids 7329 330**# rm /sys/fs/bpf/xdp1**331 332|333| **# bpftool prog profile id 337 duration 10 cycles instructions llc_misses**334 335::336 337         51397 run_cnt338      40176203 cycles                                                 (83.05%)339      42518139 instructions    #   1.06 insns per cycle               (83.39%)340           123 llc_misses      #   2.89 LLC misses per million insns  (83.15%)341 342|343| Output below is for the trace logs.344| Run in separate terminals:345| **# bpftool prog tracelog**346| **# bpftool prog load -L -d file.o**347 348::349 350    bpftool-620059  [004] d... 2634685.517903: bpf_trace_printk: btf_load size 665 r=5351    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=6352    bpftool-620059  [004] d... 2634685.517997: bpf_trace_printk: prog_load sample insn_cnt 13 r=7353    bpftool-620059  [004] d... 2634685.517999: bpf_trace_printk: close(5) = 0354