1.. SPDX-License-Identifier: GPL-2.0 2 3.. _kfuncs-header-label: 4 5============================= 6BPF Kernel Functions (kfuncs) 7============================= 8 91. Introduction 10=============== 11 12BPF Kernel Functions or more commonly known as kfuncs are functions in the Linux 13kernel which are exposed for use by BPF programs. Unlike normal BPF helpers, 14kfuncs do not have a stable interface and can change from one kernel release to 15another. Hence, BPF programs need to be updated in response to changes in the 16kernel. See :ref:`BPF_kfunc_lifecycle_expectations` for more information. 17 182. Defining a kfunc 19=================== 20 21There are two ways to expose a kernel function to BPF programs, either make an 22existing function in the kernel visible, or add a new wrapper for BPF. In both 23cases, care must be taken that BPF program can only call such function in a 24valid context. To enforce this, visibility of a kfunc can be per program type. 25 26If you are not creating a BPF wrapper for existing kernel function, skip ahead 27to :ref:`BPF_kfunc_nodef`. 28 292.1 Creating a wrapper kfunc 30---------------------------- 31 32When defining a wrapper kfunc, the wrapper function should have extern linkage. 33This prevents the compiler from optimizing away dead code, as this wrapper kfunc 34is not invoked anywhere in the kernel itself. It is not necessary to provide a 35prototype in a header for the wrapper kfunc. 36 37An example is given below:: 38 39 /* Disables missing prototype warnings */ 40 __bpf_kfunc_start_defs(); 41 42 __bpf_kfunc struct task_struct *bpf_find_get_task_by_vpid(pid_t nr) 43 { 44 return find_get_task_by_vpid(nr); 45 } 46 47 __bpf_kfunc_end_defs(); 48 49A wrapper kfunc is often needed when we need to annotate parameters of the 50kfunc. Otherwise one may directly make the kfunc visible to the BPF program by 51registering it with the BPF subsystem. See :ref:`BPF_kfunc_nodef`. 52 532.2 kfunc Parameters 54-------------------- 55 56All kfuncs now require trusted arguments by default. This means that all 57pointer arguments must be valid, and all pointers to BTF objects must be 58passed in their unmodified form (at a zero offset, and without having been 59obtained from walking another pointer, with exceptions described below). 60 61There are two types of pointers to kernel objects which are considered "trusted": 62 631. Pointers which are passed as tracepoint or struct_ops callback arguments. 642. Pointers which were returned from a KF_ACQUIRE kfunc. 65 66Pointers to non-BTF objects (e.g. scalar pointers) may also be passed to 67kfuncs, and may have a non-zero offset. 68 69The definition of "valid" pointers is subject to change at any time, and has 70absolutely no ABI stability guarantees. 71 72As mentioned above, a nested pointer obtained from walking a trusted pointer is 73no longer trusted, with one exception. If a struct type has a field that is 74guaranteed to be valid (trusted or rcu, as in KF_RCU description below) as long 75as its parent pointer is valid, the following macros can be used to express 76that to the verifier: 77 78* ``BTF_TYPE_SAFE_TRUSTED`` 79* ``BTF_TYPE_SAFE_RCU`` 80* ``BTF_TYPE_SAFE_RCU_OR_NULL`` 81 82For example, 83 84.. code-block:: c 85 86 BTF_TYPE_SAFE_TRUSTED(struct socket) { 87 struct sock *sk; 88 }; 89 90or 91 92.. code-block:: c 93 94 BTF_TYPE_SAFE_RCU(struct task_struct) { 95 const cpumask_t *cpus_ptr; 96 struct css_set __rcu *cgroups; 97 struct task_struct __rcu *real_parent; 98 struct task_struct *group_leader; 99 }; 100 101In other words, you must: 102 1031. Wrap the valid pointer type in a ``BTF_TYPE_SAFE_*`` macro. 104 1052. Specify the type and name of the valid nested field. This field must match 106 the field in the original type definition exactly. 107 108A new type declared by a ``BTF_TYPE_SAFE_*`` macro also needs to be emitted so 109that it appears in BTF. For example, ``BTF_TYPE_SAFE_TRUSTED(struct socket)`` 110is emitted in the ``type_is_trusted()`` function as follows: 111 112.. code-block:: c 113 114 BTF_TYPE_EMIT(BTF_TYPE_SAFE_TRUSTED(struct socket)); 115 1162.3 Annotating kfunc parameters 117------------------------------- 118 119Similar to BPF helpers, there is sometime need for additional context required 120by the verifier to make the usage of kernel functions safer and more useful. 121Hence, we can annotate a parameter by suffixing the name of the argument of the 122kfunc with a __tag, where tag may be one of the supported annotations. 123 1242.3.1 __sz Annotation 125--------------------- 126 127This annotation is used to indicate a memory and size pair in the argument list. 128An example is given below:: 129 130 __bpf_kfunc void bpf_memzero(void *mem, int mem__sz) 131 { 132 ... 133 } 134 135Here, the verifier will treat first argument as a PTR_TO_MEM, and second 136argument as its size. By default, without __sz annotation, the size of the type 137of the pointer is used. Without __sz annotation, a kfunc cannot accept a void 138pointer. 139 1402.3.2 __k Annotation 141-------------------- 142 143This annotation is only understood for scalar arguments, where it indicates that 144the verifier must check the scalar argument to be a known constant, which does 145not indicate a size parameter, and the value of the constant is relevant to the 146safety of the program. 147 148An example is given below:: 149 150 __bpf_kfunc void *bpf_obj_new(u32 local_type_id__k, ...) 151 { 152 ... 153 } 154 155Here, bpf_obj_new uses local_type_id argument to find out the size of that type 156ID in program's BTF and return a sized pointer to it. Each type ID will have a 157distinct size, hence it is crucial to treat each such call as distinct when 158values don't match during verifier state pruning checks. 159 160Hence, whenever a constant scalar argument is accepted by a kfunc which is not a 161size parameter, and the value of the constant matters for program safety, __k 162suffix should be used. 163 1642.3.3 __uninit Annotation 165------------------------- 166 167This annotation is used to indicate that the argument will be treated as 168uninitialized. 169 170An example is given below:: 171 172 __bpf_kfunc int bpf_dynptr_from_skb(..., struct bpf_dynptr_kern *ptr__uninit) 173 { 174 ... 175 } 176 177Here, the dynptr will be treated as an uninitialized dynptr. Without this 178annotation, the verifier will reject the program if the dynptr passed in is 179not initialized. 180 1812.3.4 __nullable Annotation 182--------------------------- 183 184This annotation is used to indicate that the pointer argument may be NULL. 185The verifier will allow passing NULL for such arguments. 186 187An example is given below:: 188 189 __bpf_kfunc void bpf_task_release(struct task_struct *task__nullable) 190 { 191 ... 192 } 193 194Here, the task pointer may be NULL. The kfunc is responsible for checking if 195the pointer is NULL before dereferencing it. 196 197The __nullable annotation can be combined with other annotations. For example, 198when used with __sz or __szk annotations for memory and size pairs, the 199verifier will skip size validation when a NULL pointer is passed, but will 200still process the size argument to extract constant size information when 201needed:: 202 203 __bpf_kfunc void *bpf_dynptr_slice(..., void *buffer__nullable, 204 u32 buffer__szk) 205 206Here, the buffer may be NULL. If the buffer is not NULL, it must be at least 207buffer__szk bytes in size. The kfunc is responsible for checking if the buffer 208is NULL before using it. 209 2102.3.5 __str Annotation 211---------------------------- 212This annotation is used to indicate that the argument is a constant string. 213 214An example is given below:: 215 216 __bpf_kfunc bpf_get_file_xattr(..., const char *name__str, ...) 217 { 218 ... 219 } 220 221In this case, ``bpf_get_file_xattr()`` can be called as:: 222 223 bpf_get_file_xattr(..., "xattr_name", ...); 224 225Or:: 226 227 const char name[] = "xattr_name"; /* This need to be global */ 228 int BPF_PROG(...) 229 { 230 ... 231 bpf_get_file_xattr(..., name, ...); 232 ... 233 } 234 2352.3.6 __prog Annotation 236--------------------------- 237This annotation is used to indicate that the argument needs to be fixed up to 238the bpf_prog_aux of the caller BPF program. Any value passed into this argument 239is ignored, and rewritten by the verifier. 240 241An example is given below:: 242 243 __bpf_kfunc int bpf_wq_set_callback_impl(struct bpf_wq *wq, 244 int (callback_fn)(void *map, int *key, void *value), 245 unsigned int flags, 246 void *aux__prog) 247 { 248 struct bpf_prog_aux *aux = aux__prog; 249 ... 250 } 251 252.. _BPF_kfunc_nodef: 253 2542.4 Using an existing kernel function 255------------------------------------- 256 257When an existing function in the kernel is fit for consumption by BPF programs, 258it can be directly registered with the BPF subsystem. However, care must still 259be taken to review the context in which it will be invoked by the BPF program 260and whether it is safe to do so. 261 2622.5 Annotating kfuncs 263--------------------- 264 265In addition to kfuncs' arguments, verifier may need more information about the 266type of kfunc(s) being registered with the BPF subsystem. To do so, we define 267flags on a set of kfuncs as follows:: 268 269 BTF_KFUNCS_START(bpf_task_set) 270 BTF_ID_FLAGS(func, bpf_get_task_pid, KF_ACQUIRE | KF_RET_NULL) 271 BTF_ID_FLAGS(func, bpf_put_pid, KF_RELEASE) 272 BTF_KFUNCS_END(bpf_task_set) 273 274This set encodes the BTF ID of each kfunc listed above, and encodes the flags 275along with it. Ofcourse, it is also allowed to specify no flags. 276 277kfunc definitions should also always be annotated with the ``__bpf_kfunc`` 278macro. This prevents issues such as the compiler inlining the kfunc if it's a 279static kernel function, or the function being elided in an LTO build as it's 280not used in the rest of the kernel. Developers should not manually add 281annotations to their kfunc to prevent these issues. If an annotation is 282required to prevent such an issue with your kfunc, it is a bug and should be 283added to the definition of the macro so that other kfuncs are similarly 284protected. An example is given below:: 285 286 __bpf_kfunc struct task_struct *bpf_get_task_pid(s32 pid) 287 { 288 ... 289 } 290 2912.5.1 KF_ACQUIRE flag 292--------------------- 293 294The KF_ACQUIRE flag is used to indicate that the kfunc returns a pointer to a 295refcounted object. The verifier will then ensure that the pointer to the object 296is eventually released using a release kfunc, or transferred to a map using a 297referenced kptr (by invoking bpf_kptr_xchg). If not, the verifier fails the 298loading of the BPF program until no lingering references remain in all possible 299explored states of the program. 300 3012.5.2 KF_RET_NULL flag 302---------------------- 303 304The KF_RET_NULL flag is used to indicate that the pointer returned by the kfunc 305may be NULL. Hence, it forces the user to do a NULL check on the pointer 306returned from the kfunc before making use of it (dereferencing or passing to 307another helper). This flag is often used in pairing with KF_ACQUIRE flag, but 308both are orthogonal to each other. 309 3102.5.3 KF_RELEASE flag 311--------------------- 312 313The KF_RELEASE flag is used to indicate that the kfunc releases the pointer 314passed in to it. There can be only one referenced pointer that can be passed 315in. All copies of the pointer being released are invalidated as a result of 316invoking kfunc with this flag. 317 3182.5.4 KF_SLEEPABLE flag 319----------------------- 320 321The KF_SLEEPABLE flag is used for kfuncs that may sleep. Such kfuncs can only 322be called by sleepable BPF programs (BPF_F_SLEEPABLE). 323 3242.5.5 KF_DESTRUCTIVE flag 325-------------------------- 326 327The KF_DESTRUCTIVE flag is used to indicate functions calling which is 328destructive to the system. For example such a call can result in system 329rebooting or panicking. Due to this additional restrictions apply to these 330calls. At the moment they only require CAP_SYS_BOOT capability, but more can be 331added later. 332 3332.5.6 KF_RCU flag 334----------------- 335 336The KF_RCU flag allows kfuncs to opt out of the default trusted args 337requirement and accept RCU pointers with weaker guarantees. The kfuncs marked 338with KF_RCU expect either PTR_TRUSTED or MEM_RCU arguments. The verifier 339guarantees that the objects are valid and there is no use-after-free. The 340pointers are not NULL, but the object's refcount could have reached zero. The 341kfuncs need to consider doing refcnt != 0 check, especially when returning a 342KF_ACQUIRE pointer. Note as well that a KF_ACQUIRE kfunc that is KF_RCU should 343very likely also be KF_RET_NULL. 344 3452.5.7 KF_RCU_PROTECTED flag 346--------------------------- 347 348The KF_RCU_PROTECTED flag is used to indicate that the kfunc must be invoked in 349an RCU critical section. This is assumed by default in non-sleepable programs, 350and must be explicitly ensured by calling ``bpf_rcu_read_lock`` for sleepable 351ones. 352 353If the kfunc returns a pointer value, this flag also enforces that the returned 354pointer is RCU protected, and can only be used while the RCU critical section is 355active. 356 357The flag is distinct from the ``KF_RCU`` flag, which only ensures that its 358arguments are at least RCU protected pointers. This may transitively imply that 359RCU protection is ensured, but it does not work in cases of kfuncs which require 360RCU protection but do not take RCU protected arguments. 361 362.. _KF_deprecated_flag: 363 3642.5.8 KF_DEPRECATED flag 365------------------------ 366 367The KF_DEPRECATED flag is used for kfuncs which are scheduled to be 368changed or removed in a subsequent kernel release. A kfunc that is 369marked with KF_DEPRECATED should also have any relevant information 370captured in its kernel doc. Such information typically includes the 371kfunc's expected remaining lifespan, a recommendation for new 372functionality that can replace it if any is available, and possibly a 373rationale for why it is being removed. 374 375Note that while on some occasions, a KF_DEPRECATED kfunc may continue to be 376supported and have its KF_DEPRECATED flag removed, it is likely to be far more 377difficult to remove a KF_DEPRECATED flag after it's been added than it is to 378prevent it from being added in the first place. As described in 379:ref:`BPF_kfunc_lifecycle_expectations`, users that rely on specific kfuncs are 380encouraged to make their use-cases known as early as possible, and participate 381in upstream discussions regarding whether to keep, change, deprecate, or remove 382those kfuncs if and when such discussions occur. 383 3842.6 Registering the kfuncs 385-------------------------- 386 387Once the kfunc is prepared for use, the final step to making it visible is 388registering it with the BPF subsystem. Registration is done per BPF program 389type. An example is shown below:: 390 391 BTF_KFUNCS_START(bpf_task_set) 392 BTF_ID_FLAGS(func, bpf_get_task_pid, KF_ACQUIRE | KF_RET_NULL) 393 BTF_ID_FLAGS(func, bpf_put_pid, KF_RELEASE) 394 BTF_KFUNCS_END(bpf_task_set) 395 396 static const struct btf_kfunc_id_set bpf_task_kfunc_set = { 397 .owner = THIS_MODULE, 398 .set = &bpf_task_set, 399 }; 400 401 static int init_subsystem(void) 402 { 403 return register_btf_kfunc_id_set(BPF_PROG_TYPE_TRACING, &bpf_task_kfunc_set); 404 } 405 late_initcall(init_subsystem); 406 4072.7 Specifying no-cast aliases with ___init 408-------------------------------------------- 409 410The verifier will always enforce that the BTF type of a pointer passed to a 411kfunc by a BPF program, matches the type of pointer specified in the kfunc 412definition. The verifier, does, however, allow types that are equivalent 413according to the C standard to be passed to the same kfunc arg, even if their 414BTF_IDs differ. 415 416For example, for the following type definition: 417 418.. code-block:: c 419 420 struct bpf_cpumask { 421 cpumask_t cpumask; 422 refcount_t usage; 423 }; 424 425The verifier would allow a ``struct bpf_cpumask *`` to be passed to a kfunc 426taking a ``cpumask_t *`` (which is a typedef of ``struct cpumask *``). For 427instance, both ``struct cpumask *`` and ``struct bpf_cpmuask *`` can be passed 428to bpf_cpumask_test_cpu(). 429 430In some cases, this type-aliasing behavior is not desired. ``struct 431nf_conn___init`` is one such example: 432 433.. code-block:: c 434 435 struct nf_conn___init { 436 struct nf_conn ct; 437 }; 438 439The C standard would consider these types to be equivalent, but it would not 440always be safe to pass either type to a trusted kfunc. ``struct 441nf_conn___init`` represents an allocated ``struct nf_conn`` object that has 442*not yet been initialized*, so it would therefore be unsafe to pass a ``struct 443nf_conn___init *`` to a kfunc that's expecting a fully initialized ``struct 444nf_conn *`` (e.g. ``bpf_ct_change_timeout()``). 445 446In order to accommodate such requirements, the verifier will enforce strict 447PTR_TO_BTF_ID type matching if two types have the exact same name, with one 448being suffixed with ``___init``. 449 450.. _BPF_kfunc_lifecycle_expectations: 451 4523. kfunc lifecycle expectations 453=============================== 454 455kfuncs provide a kernel <-> kernel API, and thus are not bound by any of the 456strict stability restrictions associated with kernel <-> user UAPIs. This means 457they can be thought of as similar to EXPORT_SYMBOL_GPL, and can therefore be 458modified or removed by a maintainer of the subsystem they're defined in when 459it's deemed necessary. 460 461Like any other change to the kernel, maintainers will not change or remove a 462kfunc without having a reasonable justification. Whether or not they'll choose 463to change a kfunc will ultimately depend on a variety of factors, such as how 464widely used the kfunc is, how long the kfunc has been in the kernel, whether an 465alternative kfunc exists, what the norm is in terms of stability for the 466subsystem in question, and of course what the technical cost is of continuing 467to support the kfunc. 468 469There are several implications of this: 470 471a) kfuncs that are widely used or have been in the kernel for a long time will 472 be more difficult to justify being changed or removed by a maintainer. In 473 other words, kfuncs that are known to have a lot of users and provide 474 significant value provide stronger incentives for maintainers to invest the 475 time and complexity in supporting them. It is therefore important for 476 developers that are using kfuncs in their BPF programs to communicate and 477 explain how and why those kfuncs are being used, and to participate in 478 discussions regarding those kfuncs when they occur upstream. 479 480b) Unlike regular kernel symbols marked with EXPORT_SYMBOL_GPL, BPF programs 481 that call kfuncs are generally not part of the kernel tree. This means that 482 refactoring cannot typically change callers in-place when a kfunc changes, 483 as is done for e.g. an upstreamed driver being updated in place when a 484 kernel symbol is changed. 485 486 Unlike with regular kernel symbols, this is expected behavior for BPF 487 symbols, and out-of-tree BPF programs that use kfuncs should be considered 488 relevant to discussions and decisions around modifying and removing those 489 kfuncs. The BPF community will take an active role in participating in 490 upstream discussions when necessary to ensure that the perspectives of such 491 users are taken into account. 492 493c) A kfunc will never have any hard stability guarantees. BPF APIs cannot and 494 will not ever hard-block a change in the kernel purely for stability 495 reasons. That being said, kfuncs are features that are meant to solve 496 problems and provide value to users. The decision of whether to change or 497 remove a kfunc is a multivariate technical decision that is made on a 498 case-by-case basis, and which is informed by data points such as those 499 mentioned above. It is expected that a kfunc being removed or changed with 500 no warning will not be a common occurrence or take place without sound 501 justification, but it is a possibility that must be accepted if one is to 502 use kfuncs. 503 5043.1 kfunc deprecation 505--------------------- 506 507As described above, while sometimes a maintainer may find that a kfunc must be 508changed or removed immediately to accommodate some changes in their subsystem, 509usually kfuncs will be able to accommodate a longer and more measured 510deprecation process. For example, if a new kfunc comes along which provides 511superior functionality to an existing kfunc, the existing kfunc may be 512deprecated for some period of time to allow users to migrate their BPF programs 513to use the new one. Or, if a kfunc has no known users, a decision may be made 514to remove the kfunc (without providing an alternative API) after some 515deprecation period so as to provide users with a window to notify the kfunc 516maintainer if it turns out that the kfunc is actually being used. 517 518It's expected that the common case will be that kfuncs will go through a 519deprecation period rather than being changed or removed without warning. As 520described in :ref:`KF_deprecated_flag`, the kfunc framework provides the 521KF_DEPRECATED flag to kfunc developers to signal to users that a kfunc has been 522deprecated. Once a kfunc has been marked with KF_DEPRECATED, the following 523procedure is followed for removal: 524 5251. Any relevant information for deprecated kfuncs is documented in the kfunc's 526 kernel docs. This documentation will typically include the kfunc's expected 527 remaining lifespan, a recommendation for new functionality that can replace 528 the usage of the deprecated function (or an explanation as to why no such 529 replacement exists), etc. 530 5312. The deprecated kfunc is kept in the kernel for some period of time after it 532 was first marked as deprecated. This time period will be chosen on a 533 case-by-case basis, and will typically depend on how widespread the use of 534 the kfunc is, how long it has been in the kernel, and how hard it is to move 535 to alternatives. This deprecation time period is "best effort", and as 536 described :ref:`above<BPF_kfunc_lifecycle_expectations>`, circumstances may 537 sometimes dictate that the kfunc be removed before the full intended 538 deprecation period has elapsed. 539 5403. After the deprecation period the kfunc will be removed. At this point, BPF 541 programs calling the kfunc will be rejected by the verifier. 542 5434. Core kfuncs 544============== 545 546The BPF subsystem provides a number of "core" kfuncs that are potentially 547applicable to a wide variety of different possible use cases and programs. 548Those kfuncs are documented here. 549 5504.1 struct task_struct * kfuncs 551------------------------------- 552 553There are a number of kfuncs that allow ``struct task_struct *`` objects to be 554used as kptrs: 555 556.. kernel-doc:: kernel/bpf/helpers.c 557 :identifiers: bpf_task_acquire bpf_task_release 558 559These kfuncs are useful when you want to acquire or release a reference to a 560``struct task_struct *`` that was passed as e.g. a tracepoint arg, or a 561struct_ops callback arg. For example: 562 563.. code-block:: c 564 565 /** 566 * A trivial example tracepoint program that shows how to 567 * acquire and release a struct task_struct * pointer. 568 */ 569 SEC("tp_btf/task_newtask") 570 int BPF_PROG(task_acquire_release_example, struct task_struct *task, u64 clone_flags) 571 { 572 struct task_struct *acquired; 573 574 acquired = bpf_task_acquire(task); 575 if (acquired) 576 /* 577 * In a typical program you'd do something like store 578 * the task in a map, and the map will automatically 579 * release it later. Here, we release it manually. 580 */ 581 bpf_task_release(acquired); 582 return 0; 583 } 584 585 586References acquired on ``struct task_struct *`` objects are RCU protected. 587Therefore, when in an RCU read region, you can obtain a pointer to a task 588embedded in a map value without having to acquire a reference: 589 590.. code-block:: c 591 592 #define private(name) SEC(".data." #name) __hidden __attribute__((aligned(8))) 593 private(TASK) static struct task_struct *global; 594 595 /** 596 * A trivial example showing how to access a task stored 597 * in a map using RCU. 598 */ 599 SEC("tp_btf/task_newtask") 600 int BPF_PROG(task_rcu_read_example, struct task_struct *task, u64 clone_flags) 601 { 602 struct task_struct *local_copy; 603 604 bpf_rcu_read_lock(); 605 local_copy = global; 606 if (local_copy) 607 /* 608 * We could also pass local_copy to kfuncs or helper functions here, 609 * as we're guaranteed that local_copy will be valid until we exit 610 * the RCU read region below. 611 */ 612 bpf_printk("Global task %s is valid", local_copy->comm); 613 else 614 bpf_printk("No global task found"); 615 bpf_rcu_read_unlock(); 616 617 /* At this point we can no longer reference local_copy. */ 618 619 return 0; 620 } 621 622---- 623 624A BPF program can also look up a task from a pid. This can be useful if the 625caller doesn't have a trusted pointer to a ``struct task_struct *`` object that 626it can acquire a reference on with bpf_task_acquire(). 627 628.. kernel-doc:: kernel/bpf/helpers.c 629 :identifiers: bpf_task_from_pid 630 631Here is an example of it being used: 632 633.. code-block:: c 634 635 SEC("tp_btf/task_newtask") 636 int BPF_PROG(task_get_pid_example, struct task_struct *task, u64 clone_flags) 637 { 638 struct task_struct *lookup; 639 640 lookup = bpf_task_from_pid(task->pid); 641 if (!lookup) 642 /* A task should always be found, as %task is a tracepoint arg. */ 643 return -ENOENT; 644 645 if (lookup->pid != task->pid) { 646 /* bpf_task_from_pid() looks up the task via its 647 * globally-unique pid from the init_pid_ns. Thus, 648 * the pid of the lookup task should always be the 649 * same as the input task. 650 */ 651 bpf_task_release(lookup); 652 return -EINVAL; 653 } 654 655 /* bpf_task_from_pid() returns an acquired reference, 656 * so it must be dropped before returning from the 657 * tracepoint handler. 658 */ 659 bpf_task_release(lookup); 660 return 0; 661 } 662 6634.2 struct cgroup * kfuncs 664-------------------------- 665 666``struct cgroup *`` objects also have acquire and release functions: 667 668.. kernel-doc:: kernel/bpf/helpers.c 669 :identifiers: bpf_cgroup_acquire bpf_cgroup_release 670 671These kfuncs are used in exactly the same manner as bpf_task_acquire() and 672bpf_task_release() respectively, so we won't provide examples for them. 673 674---- 675 676Other kfuncs available for interacting with ``struct cgroup *`` objects are 677bpf_cgroup_ancestor() and bpf_cgroup_from_id(), allowing callers to access 678the ancestor of a cgroup and find a cgroup by its ID, respectively. Both 679return a cgroup kptr. 680 681.. kernel-doc:: kernel/bpf/helpers.c 682 :identifiers: bpf_cgroup_ancestor 683 684.. kernel-doc:: kernel/bpf/helpers.c 685 :identifiers: bpf_cgroup_from_id 686 687Eventually, BPF should be updated to allow this to happen with a normal memory 688load in the program itself. This is currently not possible without more work in 689the verifier. bpf_cgroup_ancestor() can be used as follows: 690 691.. code-block:: c 692 693 /** 694 * Simple tracepoint example that illustrates how a cgroup's 695 * ancestor can be accessed using bpf_cgroup_ancestor(). 696 */ 697 SEC("tp_btf/cgroup_mkdir") 698 int BPF_PROG(cgrp_ancestor_example, struct cgroup *cgrp, const char *path) 699 { 700 struct cgroup *parent; 701 702 /* The parent cgroup resides at the level before the current cgroup's level. */ 703 parent = bpf_cgroup_ancestor(cgrp, cgrp->level - 1); 704 if (!parent) 705 return -ENOENT; 706 707 bpf_printk("Parent id is %d", parent->self.id); 708 709 /* Return the parent cgroup that was acquired above. */ 710 bpf_cgroup_release(parent); 711 return 0; 712 } 713 7144.3 struct cpumask * kfuncs 715--------------------------- 716 717BPF provides a set of kfuncs that can be used to query, allocate, mutate, and 718destroy struct cpumask * objects. Please refer to :ref:`cpumasks-header-label` 719for more details. 720