mirrors/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2025-11-04 02:30:34 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions
linux/tools/bpf/bpftool/Documentation/bpftool-prog.rst
Quentin Monnet 475a23c2c1 tools: bpftool: Complete metrics list in "bpftool prog profile" doc 
Profiling programs with bpftool was extended some time ago to support
two new metrics, namely itlb_misses and dtlb_misses (misses for the
instruction/data translation lookaside buffer). Update the manual page
and bash completion accordingly.

Fixes: 450d060e8f ("bpftool: Add {i,d}tlb_misses support for bpftool profile")
Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210730215435.7095-8-quentin@isovalent.com
2021-07-30 15:40:28 -07:00

358 lines
14 KiB
ReStructuredText
Raw Blame History

================
bpftool-prog
================
-------------------------------------------------------------------------------
tool for inspection and simple manipulation of eBPF progs
-------------------------------------------------------------------------------
:Manual section: 8
SYNOPSIS
========
**bpftool** [*OPTIONS*] **prog** *COMMAND*
*OPTIONS* := { { **-j** | **--json** } [{ **-p** | **--pretty** }] | { **-d** | **--debug** } |
{ **-f** | **--bpffs** } | { **-m** | **--mapcompat** } | { **-n** | **--nomount** } |
{ **-L** | **--use-loader** } }
*COMMANDS* :=
{ **show** | **list** | **dump xlated** | **dump jited** | **pin** | **load**
| **loadall** | **help** }
PROG COMMANDS
=============
| **bpftool** **prog** { **show** | **list** } [*PROG*]
| **bpftool** **prog dump xlated** *PROG* [{**file** *FILE* | **opcodes** | **visual** | **linum**}]
| **bpftool** **prog dump jited** *PROG* [{**file** *FILE* | **opcodes** | **linum**}]
| **bpftool** **prog pin** *PROG* *FILE*
| **bpftool** **prog** { **load** | **loadall** } *OBJ* *PATH* [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** *NAME*] [**pinmaps** *MAP_DIR*]
| **bpftool** **prog attach** *PROG* *ATTACH_TYPE* [*MAP*]
| **bpftool** **prog detach** *PROG* *ATTACH_TYPE* [*MAP*]
| **bpftool** **prog tracelog**
| **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*]
| **bpftool** **prog profile** *PROG* [**duration** *DURATION*] *METRICs*
| **bpftool** **prog help**
|
| *MAP* := { **id** *MAP_ID* | **pinned** *FILE* }
| *PROG* := { **id** *PROG_ID* | **pinned** *FILE* | **tag** *PROG_TAG* | **name** *PROG_NAME* }
| *TYPE* := {
| **socket** | **kprobe** | **kretprobe** | **classifier** | **action** |
| **tracepoint** | **raw_tracepoint** | **xdp** | **perf_event** | **cgroup/skb** |
| **cgroup/sock** | **cgroup/dev** | **lwt_in** | **lwt_out** | **lwt_xmit** |
| **lwt_seg6local** | **sockops** | **sk_skb** | **sk_msg** | **lirc_mode2** |
| **cgroup/bind4** | **cgroup/bind6** | **cgroup/post_bind4** | **cgroup/post_bind6** |
| **cgroup/connect4** | **cgroup/connect6** | **cgroup/getpeername4** | **cgroup/getpeername6** |
| **cgroup/getsockname4** | **cgroup/getsockname6** | **cgroup/sendmsg4** | **cgroup/sendmsg6** |
| **cgroup/recvmsg4** | **cgroup/recvmsg6** | **cgroup/sysctl** |
| **cgroup/getsockopt** | **cgroup/setsockopt** | **cgroup/sock_release** |
| **struct_ops** | **fentry** | **fexit** | **freplace** | **sk_lookup**
| }
| *ATTACH_TYPE* := {
| **msg_verdict** | **skb_verdict** | **stream_verdict** | **stream_parser** | **flow_dissector**
| }
| *METRICs* := {
| **cycles** | **instructions** | **l1d_loads** | **llc_misses** |
| **itlb_misses** | **dtlb_misses**
| }
DESCRIPTION
===========
**bpftool prog { show | list }** [*PROG*]
Show information about loaded programs. If *PROG* is
specified show information only about given programs,
otherwise list all programs currently loaded on the system.
In case of **tag** or **name**, *PROG* may match several
programs which will all be shown.
Output will start with program ID followed by program type and
zero or more named attributes (depending on kernel version).
Since Linux 5.1 the kernel can collect statistics on BPF
programs (such as the total time spent running the program,
and the number of times it was run). If available, bpftool
shows such statistics. However, the kernel does not collect
them by defaults, as it slightly impacts performance on each
program run. Activation or deactivation of the feature is
performed via the **kernel.bpf_stats_enabled** sysctl knob.
Since Linux 5.8 bpftool is able to discover information about
processes that hold open file descriptors (FDs) against BPF
programs. On such kernels bpftool will automatically emit this
information as well.
**bpftool prog dump xlated** *PROG* [{ **file** *FILE* | **opcodes** | **visual** | **linum** }]
Dump eBPF instructions of the programs from the kernel. By
default, eBPF will be disassembled and printed to standard
output in human-readable format. In this case, **opcodes**
controls if raw opcodes should be printed as well.
In case of **tag** or **name**, *PROG* may match several
programs which will all be dumped. However, if **file** or
**visual** is specified, *PROG* must match a single program.
If **file** is specified, the binary image will instead be
written to *FILE*.
If **visual** is specified, control flow graph (CFG) will be
built instead, and eBPF instructions will be presented with
CFG in DOT format, on standard output.
If the programs have line_info available, the source line will
be displayed by default. If **linum** is specified,
the filename, line number and line column will also be
displayed on top of the source line.
**bpftool prog dump jited** *PROG* [{ **file** *FILE* | **opcodes** | **linum** }]
Dump jited image (host machine code) of the program.
If *FILE* is specified image will be written to a file,
otherwise it will be disassembled and printed to stdout.
*PROG* must match a single program when **file** is specified.
**opcodes** controls if raw opcodes will be printed.
If the prog has line_info available, the source line will
be displayed by default. If **linum** is specified,
the filename, line number and line column will also be
displayed on top of the source line.
**bpftool prog pin** *PROG* *FILE*
Pin program *PROG* as *FILE*.
Note: *FILE* must be located in *bpffs* mount. It must not
contain a dot character ('.'), which is reserved for future
extensions of *bpffs*.
**bpftool prog { load | loadall }** *OBJ* *PATH* [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** *NAME*] [**pinmaps** *MAP_DIR*]
Load bpf program(s) from binary *OBJ* and pin as *PATH*.
**bpftool prog load** pins only the first program from the
*OBJ* as *PATH*. **bpftool prog loadall** pins all programs
from the *OBJ* under *PATH* directory.
**type** is optional, if not specified program type will be
inferred from section names.
By default bpftool will create new maps as declared in the ELF
object being loaded. **map** parameter allows for the reuse
of existing maps. It can be specified multiple times, each
time for a different map. *IDX* refers to index of the map
to be replaced in the ELF file counting from 0, while *NAME*
allows to replace a map by name. *MAP* specifies the map to
use, referring to it by **id** or through a **pinned** file.
If **dev** *NAME* is specified program will be loaded onto
given networking device (offload).
Optional **pinmaps** argument can be provided to pin all
maps under *MAP_DIR* directory.
Note: *PATH* must be located in *bpffs* mount. It must not
contain a dot character ('.'), which is reserved for future
extensions of *bpffs*.
**bpftool prog attach** *PROG* *ATTACH_TYPE* [*MAP*]
Attach bpf program *PROG* (with type specified by
*ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP*
parameter, with the exception of *flow_dissector* which is
attached to current networking name space.
**bpftool prog detach** *PROG* *ATTACH_TYPE* [*MAP*]
Detach bpf program *PROG* (with type specified by
*ATTACH_TYPE*). Most *ATTACH_TYPEs* require a *MAP*
parameter, with the exception of *flow_dissector* which is
detached from the current networking name space.
**bpftool prog tracelog**
Dump the trace pipe of the system to the console (stdout).
Hit <Ctrl+C> to stop printing. BPF programs can write to this
trace pipe at runtime with the **bpf_trace_printk**\ () helper.
This should be used only for debugging purposes. For
streaming data from BPF programs to user space, one can use
perf events (see also **bpftool-map**\ (8)).
**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*]
Run BPF program *PROG* in the kernel testing infrastructure
for BPF, meaning that the program works on the data and
context provided by the user, and not on actual packets or
monitored functions etc. Return value and duration for the
test run are printed out to the console.
Input data is read from the *FILE* passed with **data_in**.
If this *FILE* is "**-**", input data is read from standard
input. Input context, if any, is read from *FILE* passed with
**ctx_in**. Again, "**-**" can be used to read from standard
input, but only if standard input is not already in use for
input data. If a *FILE* is passed with **data_out**, output
data is written to that file. Similarly, output context is
written to the *FILE* passed with **ctx_out**. For both
output flows, "**-**" can be used to print to the standard
output (as plain text, or JSON if relevant option was
passed). If output keywords are omitted, output data and
context are discarded. Keywords **data_size_out** and
**ctx_size_out** are used to pass the size (in bytes) for the
output buffers to the kernel, although the default of 32 kB
should be more than enough for most cases.
Keyword **repeat** is used to indicate the number of
consecutive runs to perform. Note that output data and
context printed to files correspond to the last of those
runs. The duration printed out at the end of the runs is an
average over all runs performed by the command.
Not all program types support test run. Among those which do,
not all of them can take the **ctx_in**/**ctx_out**
arguments. bpftool does not perform checks on program types.
**bpftool prog profile** *PROG* [**duration** *DURATION*] *METRICs*
Profile *METRICs* for bpf program *PROG* for *DURATION*
seconds or until user hits <Ctrl+C>. *DURATION* is optional.
If *DURATION* is not specified, the profiling will run up to
**UINT_MAX** seconds.
**bpftool prog help**
Print short help message.
OPTIONS
=======
.. include:: common_options.rst
-f, --bpffs
When showing BPF programs, show file names of pinned
programs.
-m, --mapcompat
Allow loading maps with unknown map definitions.
-n, --nomount
Do not automatically attempt to mount any virtual file system
(such as tracefs or BPF virtual file system) when necessary.
-L, --use-loader
Load program as a "loader" program. This is useful to debug
the generation of such programs. When this option is in
use, bpftool attempts to load the programs from the object
file into the kernel, but does not pin them (therefore, the
*PATH* must not be provided).
When combined with the **-d**\ \|\ **--debug** option,
additional debug messages are generated, and the execution
of the loader program will use the **bpf_trace_printk**\ ()
helper to log each step of loading BTF, creating the maps,
and loading the programs (see **bpftool prog tracelog** as
a way to dump those messages).
EXAMPLES
========
**# bpftool prog show**
::
10: xdp name some_prog tag 005a3d2123620c8b gpl run_time_ns 81632 run_cnt 10
loaded_at 2017-09-29T20:11:00+0000 uid 0
xlated 528B jited 370B memlock 4096B map_ids 10
pids systemd(1)
**# bpftool --json --pretty prog show**
::
[{
"id": 10,
"type": "xdp",
"tag": "005a3d2123620c8b",
"gpl_compatible": true,
"run_time_ns": 81632,
"run_cnt": 10,
"loaded_at": 1506715860,
"uid": 0,
"bytes_xlated": 528,
"jited": true,
"bytes_jited": 370,
"bytes_memlock": 4096,
"map_ids": [10
],
"pids": [{
"pid": 1,
"comm": "systemd"
}
]
}
]
|
| **# bpftool prog dump xlated id 10 file /tmp/t**
| **$ ls -l /tmp/t**
::
-rw------- 1 root root 560 Jul 22 01:42 /tmp/t
**# bpftool prog dump jited tag 005a3d2123620c8b**
::
0: push %rbp
1: mov %rsp,%rbp
2: sub $0x228,%rsp
3: sub $0x28,%rbp
4: mov %rbx,0x0(%rbp)
|
| **# mount -t bpf none /sys/fs/bpf/**
| **# bpftool prog pin id 10 /sys/fs/bpf/prog**
| **# bpftool prog load ./my_prog.o /sys/fs/bpf/prog2**
| **# ls -l /sys/fs/bpf/**
::
-rw------- 1 root root 0 Jul 22 01:43 prog
-rw------- 1 root root 0 Jul 22 01:44 prog2
**# bpftool prog dump jited pinned /sys/fs/bpf/prog opcodes**
::
0: push %rbp
55
1: mov %rsp,%rbp
48 89 e5
4: sub $0x228,%rsp
48 81 ec 28 02 00 00
b: sub $0x28,%rbp
48 83 ed 28
f: mov %rbx,0x0(%rbp)
48 89 5d 00
|
| **# bpftool prog load xdp1_kern.o /sys/fs/bpf/xdp1 type xdp map name rxcnt id 7**
| **# bpftool prog show pinned /sys/fs/bpf/xdp1**
::
9: xdp name xdp_prog1 tag 539ec6ce11b52f98 gpl
loaded_at 2018-06-25T16:17:31-0700 uid 0
xlated 488B jited 336B memlock 4096B map_ids 7
**# rm /sys/fs/bpf/xdp1**
|
| **# bpftool prog profile id 337 duration 10 cycles instructions llc_misses**
::
51397 run_cnt
40176203 cycles (83.05%)
42518139 instructions # 1.06 insns per cycle (83.39%)
123 llc_misses # 2.89 LLC misses per million insns (83.15%)
|
| Output below is for the trace logs.
| Run in separate terminals:
| **# bpftool prog tracelog**
| **# bpftool prog load -L -d file.o**
::
bpftool-620059 [004] d... 2634685.517903: bpf_trace_printk: btf_load size 665 r=5
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=6
bpftool-620059 [004] d... 2634685.517997: bpf_trace_printk: prog_load sample insn_cnt 13 r=7
bpftool-620059 [004] d... 2634685.517999: bpf_trace_printk: close(5) = 0
Reference in a new issue View git blame Copy permalink