brintos

brintos / linux-shallow public Read only

0
0
Text · 15.5 KiB · ca860fd Raw
471 lines · plain
1.. SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)2 3================4bpftool-gen5================6-------------------------------------------------------------------------------7tool for BPF code-generation8-------------------------------------------------------------------------------9 10:Manual section: 811 12.. include:: substitutions.rst13 14SYNOPSIS15========16 17**bpftool** [*OPTIONS*] **gen** *COMMAND*18 19*OPTIONS* := { |COMMON_OPTIONS| | { **-L** | **--use-loader** } }20 21*COMMAND* := { **object** | **skeleton** | **help** }22 23GEN COMMANDS24=============25 26| **bpftool** **gen object** *OUTPUT_FILE* *INPUT_FILE* [*INPUT_FILE*...]27| **bpftool** **gen skeleton** *FILE* [**name** *OBJECT_NAME*]28| **bpftool** **gen subskeleton** *FILE* [**name** *OBJECT_NAME*]29| **bpftool** **gen min_core_btf** *INPUT* *OUTPUT* *OBJECT* [*OBJECT*...]30| **bpftool** **gen help**31 32DESCRIPTION33===========34bpftool gen object *OUTPUT_FILE* *INPUT_FILE* [*INPUT_FILE*...]35    Statically link (combine) together one or more *INPUT_FILE*'s into a single36    resulting *OUTPUT_FILE*. All the files involved are BPF ELF object files.37 38    The rules of BPF static linking are mostly the same as for user-space39    object files, but in addition to combining data and instruction sections,40    .BTF and .BTF.ext (if present in any of the input files) data are combined41    together. .BTF data is deduplicated, so all the common types across42    *INPUT_FILE*'s will only be represented once in the resulting BTF43    information.44 45    BPF static linking allows to partition BPF source code into individually46    compiled files that are then linked into a single resulting BPF object47    file, which can be used to generated BPF skeleton (with **gen skeleton**48    command) or passed directly into **libbpf** (using **bpf_object__open()**49    family of APIs).50 51bpftool gen skeleton *FILE*52    Generate BPF skeleton C header file for a given *FILE*.53 54    BPF skeleton is an alternative interface to existing libbpf APIs for55    working with BPF objects. Skeleton code is intended to significantly56    shorten and simplify code to load and work with BPF programs from userspace57    side. Generated code is tailored to specific input BPF object *FILE*,58    reflecting its structure by listing out available maps, program, variables,59    etc. Skeleton eliminates the need to lookup mentioned components by name.60    Instead, if skeleton instantiation succeeds, they are populated in skeleton61    structure as valid libbpf types (e.g., **struct bpf_map** pointer) and can62    be passed to existing generic libbpf APIs.63 64    In addition to simple and reliable access to maps and programs, skeleton65    provides a storage for BPF links (**struct bpf_link**) for each BPF program66    within BPF object. When requested, supported BPF programs will be67    automatically attached and resulting BPF links stored for further use by68    user in pre-allocated fields in skeleton struct. For BPF programs that69    can't be automatically attached by libbpf, user can attach them manually,70    but store resulting BPF link in per-program link field. All such set up71    links will be automatically destroyed on BPF skeleton destruction. This72    eliminates the need for users to manage links manually and rely on libbpf73    support to detach programs and free up resources.74 75    Another facility provided by BPF skeleton is an interface to global76    variables of all supported kinds: mutable, read-only, as well as extern77    ones. This interface allows to pre-setup initial values of variables before78    BPF object is loaded and verified by kernel. For non-read-only variables,79    the same interface can be used to fetch values of global variables on80    userspace side, even if they are modified by BPF code.81 82    During skeleton generation, contents of source BPF object *FILE* is83    embedded within generated code and is thus not necessary to keep around.84    This ensures skeleton and BPF object file are matching 1-to-1 and always85    stay in sync. Generated code is dual-licensed under LGPL-2.1 and86    BSD-2-Clause licenses.87 88    It is a design goal and guarantee that skeleton interfaces are89    interoperable with generic libbpf APIs. User should always be able to use90    skeleton API to create and load BPF object, and later use libbpf APIs to91    keep working with specific maps, programs, etc.92 93    As part of skeleton, few custom functions are generated. Each of them is94    prefixed with object name. Object name can either be derived from object95    file name, i.e., if BPF object file name is **example.o**, BPF object name96    will be **example**. Object name can be also specified explicitly through97    **name** *OBJECT_NAME* parameter. The following custom functions are98    provided (assuming **example** as the object name):99 100    - **example__open** and **example__open_opts**.101      These functions are used to instantiate skeleton. It corresponds to102      libbpf's **bpf_object__open**\ () API. **_opts** variants accepts extra103      **bpf_object_open_opts** options.104 105    - **example__load**.106      This function creates maps, loads and verifies BPF programs, initializes107      global data maps. It corresponds to libbpf's **bpf_object__load**\ ()108      API.109 110    - **example__open_and_load** combines **example__open** and111      **example__load** invocations in one commonly used operation.112 113    - **example__attach** and **example__detach**.114      This pair of functions allow to attach and detach, correspondingly,115      already loaded BPF object. Only BPF programs of types supported by libbpf116      for auto-attachment will be auto-attached and their corresponding BPF117      links instantiated. For other BPF programs, user can manually create a118      BPF link and assign it to corresponding fields in skeleton struct.119      **example__detach** will detach both links created automatically, as well120      as those populated by user manually.121 122    - **example__destroy**.123      Detach and unload BPF programs, free up all the resources used by124      skeleton and BPF object.125 126    If BPF object has global variables, corresponding structs with memory127    layout corresponding to global data data section layout will be created.128    Currently supported ones are: *.data*, *.bss*, *.rodata*, and *.kconfig*129    structs/data sections. These data sections/structs can be used to set up130    initial values of variables, if set before **example__load**. Afterwards,131    if target kernel supports memory-mapped BPF arrays, same structs can be132    used to fetch and update (non-read-only) data from userspace, with same133    simplicity as for BPF side.134 135bpftool gen subskeleton *FILE*136    Generate BPF subskeleton C header file for a given *FILE*.137 138    Subskeletons are similar to skeletons, except they do not own the139    corresponding maps, programs, or global variables. They require that the140    object file used to generate them is already loaded into a *bpf_object* by141    some other means.142 143    This functionality is useful when a library is included into a larger BPF144    program. A subskeleton for the library would have access to all objects and145    globals defined in it, without having to know about the larger program.146 147    Consequently, there are only two functions defined for subskeletons:148 149    - **example__open(bpf_object\*)**.150      Instantiates a subskeleton from an already opened (but not necessarily151      loaded) **bpf_object**.152 153    - **example__destroy()**.154      Frees the storage for the subskeleton but *does not* unload any BPF155      programs or maps.156 157bpftool gen min_core_btf *INPUT* *OUTPUT* *OBJECT* [*OBJECT*...]158    Generate a minimum BTF file as *OUTPUT*, derived from a given *INPUT* BTF159    file, containing all needed BTF types so one, or more, given eBPF objects160    CO-RE relocations may be satisfied.161 162    When kernels aren't compiled with CONFIG_DEBUG_INFO_BTF, libbpf, when163    loading an eBPF object, has to rely on external BTF files to be able to164    calculate CO-RE relocations.165 166    Usually, an external BTF file is built from existing kernel DWARF data167    using pahole. It contains all the types used by its respective kernel image168    and, because of that, is big.169 170    The min_core_btf feature builds smaller BTF files, customized to one or171    multiple eBPF objects, so they can be distributed together with an eBPF172    CO-RE based application, turning the application portable to different173    kernel versions.174 175    Check examples below for more information on how to use it.176 177bpftool gen help178    Print short help message.179 180OPTIONS181=======182.. include:: common_options.rst183 184-L, --use-loader185    For skeletons, generate a "light" skeleton (also known as "loader"186    skeleton). A light skeleton contains a loader eBPF program. It does not use187    the majority of the libbpf infrastructure, and does not need libelf.188 189EXAMPLES190========191**$ cat example1.bpf.c**192 193::194 195  #include <stdbool.h>196  #include <linux/ptrace.h>197  #include <linux/bpf.h>198  #include <bpf/bpf_helpers.h>199 200  const volatile int param1 = 42;201  bool global_flag = true;202  struct { int x; } data = {};203 204  SEC("raw_tp/sys_enter")205  int handle_sys_enter(struct pt_regs *ctx)206  {207  	static long my_static_var;208  	if (global_flag)209  		my_static_var++;210  	else211  		data.x += param1;212  	return 0;213  }214 215**$ cat example2.bpf.c**216 217::218 219  #include <linux/ptrace.h>220  #include <linux/bpf.h>221  #include <bpf/bpf_helpers.h>222 223  struct {224  	__uint(type, BPF_MAP_TYPE_HASH);225  	__uint(max_entries, 128);226  	__type(key, int);227  	__type(value, long);228  } my_map SEC(".maps");229 230  SEC("raw_tp/sys_exit")231  int handle_sys_exit(struct pt_regs *ctx)232  {233  	int zero = 0;234  	bpf_map_lookup_elem(&my_map, &zero);235  	return 0;236  }237 238**$ cat example3.bpf.c**239 240::241 242  #include <linux/ptrace.h>243  #include <linux/bpf.h>244  #include <bpf/bpf_helpers.h>245  /* This header file is provided by the bpf_testmod module. */246  #include "bpf_testmod.h"247 248  int test_2_result = 0;249 250  /* bpf_Testmod.ko calls this function, passing a "4"251   * and testmod_map->data.252   */253  SEC("struct_ops/test_2")254  void BPF_PROG(test_2, int a, int b)255  {256	test_2_result = a + b;257  }258 259  SEC(".struct_ops")260  struct bpf_testmod_ops testmod_map = {261	.test_2 = (void *)test_2,262	.data = 0x1,263  };264 265This is example BPF application with three BPF programs and a mix of BPF266maps and global variables. Source code is split across three source code267files.268 269**$ clang --target=bpf -g example1.bpf.c -o example1.bpf.o**270 271**$ clang --target=bpf -g example2.bpf.c -o example2.bpf.o**272 273**$ clang --target=bpf -g example3.bpf.c -o example3.bpf.o**274 275**$ bpftool gen object example.bpf.o example1.bpf.o example2.bpf.o example3.bpf.o**276 277This set of commands compiles *example1.bpf.c*, *example2.bpf.c* and278*example3.bpf.c* individually and then statically links respective object279files into the final BPF ELF object file *example.bpf.o*.280 281**$ bpftool gen skeleton example.bpf.o name example | tee example.skel.h**282 283::284 285  /* SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) */286 287  /* THIS FILE IS AUTOGENERATED! */288  #ifndef __EXAMPLE_SKEL_H__289  #define __EXAMPLE_SKEL_H__290 291  #include <stdlib.h>292  #include <bpf/libbpf.h>293 294  struct example {295  	struct bpf_object_skeleton *skeleton;296  	struct bpf_object *obj;297  	struct {298  		struct bpf_map *rodata;299  		struct bpf_map *data;300  		struct bpf_map *bss;301  		struct bpf_map *my_map;302		struct bpf_map *testmod_map;303  	} maps;304	struct {305		struct example__testmod_map__bpf_testmod_ops {306			const struct bpf_program *test_1;307			const struct bpf_program *test_2;308			int data;309		} *testmod_map;310	} struct_ops;311  	struct {312  		struct bpf_program *handle_sys_enter;313  		struct bpf_program *handle_sys_exit;314  	} progs;315  	struct {316  		struct bpf_link *handle_sys_enter;317  		struct bpf_link *handle_sys_exit;318  	} links;319  	struct example__bss {320  		struct {321  			int x;322  		} data;323		int test_2_result;324  	} *bss;325  	struct example__data {326  		_Bool global_flag;327  		long int handle_sys_enter_my_static_var;328  	} *data;329  	struct example__rodata {330  		int param1;331  	} *rodata;332  };333 334  static void example__destroy(struct example *obj);335  static inline struct example *example__open_opts(336                const struct bpf_object_open_opts *opts);337  static inline struct example *example__open();338  static inline int example__load(struct example *obj);339  static inline struct example *example__open_and_load();340  static inline int example__attach(struct example *obj);341  static inline void example__detach(struct example *obj);342 343  #endif /* __EXAMPLE_SKEL_H__ */344 345**$ cat example.c**346 347::348 349  #include "example.skel.h"350 351  int main()352  {353  	struct example *skel;354  	int err = 0;355 356  	skel = example__open();357  	if (!skel)358  		goto cleanup;359 360  	skel->rodata->param1 = 128;361 362	/* Change the value through the pointer of shadow type */363	skel->struct_ops.testmod_map->data = 13;364 365  	err = example__load(skel);366  	if (err)367  		goto cleanup;368 369	/* The result of the function test_2() */370	printf("test_2_result: %d\n", skel->bss->test_2_result);371 372  	err = example__attach(skel);373  	if (err)374  		goto cleanup;375 376  	/* all libbpf APIs are usable */377  	printf("my_map name: %s\n", bpf_map__name(skel->maps.my_map));378  	printf("sys_enter prog FD: %d\n",379  	       bpf_program__fd(skel->progs.handle_sys_enter));380 381  	/* detach and re-attach sys_exit program */382  	bpf_link__destroy(skel->links.handle_sys_exit);383  	skel->links.handle_sys_exit =384  		bpf_program__attach(skel->progs.handle_sys_exit);385 386  	printf("my_static_var: %ld\n",387  	       skel->bss->handle_sys_enter_my_static_var);388 389  cleanup:390  	example__destroy(skel);391  	return err;392  }393 394**# ./example**395 396::397 398  test_2_result: 17399  my_map name: my_map400  sys_enter prog FD: 8401  my_static_var: 7402 403This is a stripped-out version of skeleton generated for above example code.404 405min_core_btf406------------407 408**$ bpftool btf dump file 5.4.0-example.btf format raw**409 410::411 412  [1] INT 'long unsigned int' size=8 bits_offset=0 nr_bits=64 encoding=(none)413  [2] CONST '(anon)' type_id=1414  [3] VOLATILE '(anon)' type_id=1415  [4] ARRAY '(anon)' type_id=1 index_type_id=21 nr_elems=2416  [5] PTR '(anon)' type_id=8417  [6] CONST '(anon)' type_id=5418  [7] INT 'char' size=1 bits_offset=0 nr_bits=8 encoding=(none)419  [8] CONST '(anon)' type_id=7420  [9] INT 'unsigned int' size=4 bits_offset=0 nr_bits=32 encoding=(none)421  <long output>422 423**$ bpftool btf dump file one.bpf.o format raw**424 425::426 427  [1] PTR '(anon)' type_id=2428  [2] STRUCT 'trace_event_raw_sys_enter' size=64 vlen=4429        'ent' type_id=3 bits_offset=0430        'id' type_id=7 bits_offset=64431        'args' type_id=9 bits_offset=128432        '__data' type_id=12 bits_offset=512433  [3] STRUCT 'trace_entry' size=8 vlen=4434        'type' type_id=4 bits_offset=0435        'flags' type_id=5 bits_offset=16436        'preempt_count' type_id=5 bits_offset=24437  <long output>438 439**$ bpftool gen min_core_btf 5.4.0-example.btf 5.4.0-smaller.btf one.bpf.o**440 441**$ bpftool btf dump file 5.4.0-smaller.btf format raw**442 443::444 445  [1] TYPEDEF 'pid_t' type_id=6446  [2] STRUCT 'trace_event_raw_sys_enter' size=64 vlen=1447        'args' type_id=4 bits_offset=128448  [3] STRUCT 'task_struct' size=9216 vlen=2449        'pid' type_id=1 bits_offset=17920450        'real_parent' type_id=7 bits_offset=18048451  [4] ARRAY '(anon)' type_id=5 index_type_id=8 nr_elems=6452  [5] INT 'long unsigned int' size=8 bits_offset=0 nr_bits=64 encoding=(none)453  [6] TYPEDEF '__kernel_pid_t' type_id=8454  [7] PTR '(anon)' type_id=3455  [8] INT 'int' size=4 bits_offset=0 nr_bits=32 encoding=SIGNED456  <end>457 458Now, the "5.4.0-smaller.btf" file may be used by libbpf as an external BTF file459when loading the "one.bpf.o" object into the "5.4.0-example" kernel. Note that460the generated BTF file won't allow other eBPF objects to be loaded, just the461ones given to min_core_btf.462 463::464 465  LIBBPF_OPTS(bpf_object_open_opts, opts, .btf_custom_path = "5.4.0-smaller.btf");466  struct bpf_object *obj;467 468  obj = bpf_object__open_file("one.bpf.o", &opts);469 470  ...471