mirror of
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git
synced 2024-10-03 07:38:10 +00:00
e542f2c4cd
The need to increase RLIMIT_MEMLOCK to do anything useful with BPF is one of the first extremely frustrating gotchas that all new BPF users go through and in some cases have to learn it a very hard way. Luckily, starting with upstream Linux kernel version 5.11, BPF subsystem dropped the dependency on memlock and uses memcg-based memory accounting instead. Unfortunately, detecting memcg-based BPF memory accounting is far from trivial (as can be evidenced by this patch), so in practice most BPF applications still do unconditional RLIMIT_MEMLOCK increase. As we move towards libbpf 1.0, it would be good to allow users to forget about RLIMIT_MEMLOCK vs memcg and let libbpf do the sensible adjustment automatically. This patch paves the way forward in this matter. Libbpf will do feature detection of memcg-based accounting, and if detected, will do nothing. But if the kernel is too old, just like BCC, libbpf will automatically increase RLIMIT_MEMLOCK on behalf of user application ([0]). As this is technically a breaking change, during the transition period applications have to opt into libbpf 1.0 mode by setting LIBBPF_STRICT_AUTO_RLIMIT_MEMLOCK bit when calling libbpf_set_strict_mode(). Libbpf allows to control the exact amount of set RLIMIT_MEMLOCK limit with libbpf_set_memlock_rlim_max() API. Passing 0 will make libbpf do nothing with RLIMIT_MEMLOCK. libbpf_set_memlock_rlim_max() has to be called before the first bpf_prog_load(), bpf_btf_load(), or bpf_object__load() call, otherwise it has no effect and will return -EBUSY. [0] Closes: https://github.com/libbpf/libbpf/issues/369 Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://lore.kernel.org/bpf/20211214195904.1785155-2-andrii@kernel.org
88 lines
2.8 KiB
C
88 lines
2.8 KiB
C
/* SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) */
|
|
|
|
/*
|
|
* Libbpf legacy APIs (either discouraged or deprecated, as mentioned in [0])
|
|
*
|
|
* [0] https://docs.google.com/document/d/1UyjTZuPFWiPFyKk1tV5an11_iaRuec6U-ZESZ54nNTY
|
|
*
|
|
* Copyright (C) 2021 Facebook
|
|
*/
|
|
#ifndef __LIBBPF_LEGACY_BPF_H
|
|
#define __LIBBPF_LEGACY_BPF_H
|
|
|
|
#include <linux/bpf.h>
|
|
#include <stdbool.h>
|
|
#include <stddef.h>
|
|
#include <stdint.h>
|
|
#include "libbpf_common.h"
|
|
|
|
#ifdef __cplusplus
|
|
extern "C" {
|
|
#endif
|
|
|
|
enum libbpf_strict_mode {
|
|
/* Turn on all supported strict features of libbpf to simulate libbpf
|
|
* v1.0 behavior.
|
|
* This will be the default behavior in libbpf v1.0.
|
|
*/
|
|
LIBBPF_STRICT_ALL = 0xffffffff,
|
|
|
|
/*
|
|
* Disable any libbpf 1.0 behaviors. This is the default before libbpf
|
|
* v1.0. It won't be supported anymore in v1.0, please update your
|
|
* code so that it handles LIBBPF_STRICT_ALL mode before libbpf v1.0.
|
|
*/
|
|
LIBBPF_STRICT_NONE = 0x00,
|
|
/*
|
|
* Return NULL pointers on error, not ERR_PTR(err).
|
|
* Additionally, libbpf also always sets errno to corresponding Exx
|
|
* (positive) error code.
|
|
*/
|
|
LIBBPF_STRICT_CLEAN_PTRS = 0x01,
|
|
/*
|
|
* Return actual error codes from low-level APIs directly, not just -1.
|
|
* Additionally, libbpf also always sets errno to corresponding Exx
|
|
* (positive) error code.
|
|
*/
|
|
LIBBPF_STRICT_DIRECT_ERRS = 0x02,
|
|
/*
|
|
* Enforce strict BPF program section (SEC()) names.
|
|
* E.g., while prefiously SEC("xdp_whatever") or SEC("perf_event_blah") were
|
|
* allowed, with LIBBPF_STRICT_SEC_PREFIX this will become
|
|
* unrecognized by libbpf and would have to be just SEC("xdp") and
|
|
* SEC("xdp") and SEC("perf_event").
|
|
*
|
|
* Note, in this mode the program pin path will be based on the
|
|
* function name instead of section name.
|
|
*/
|
|
LIBBPF_STRICT_SEC_NAME = 0x04,
|
|
/*
|
|
* Disable the global 'bpf_objects_list'. Maintaining this list adds
|
|
* a race condition to bpf_object__open() and bpf_object__close().
|
|
* Clients can maintain it on their own if it is valuable for them.
|
|
*/
|
|
LIBBPF_STRICT_NO_OBJECT_LIST = 0x08,
|
|
/*
|
|
* Automatically bump RLIMIT_MEMLOCK using setrlimit() before the
|
|
* first BPF program or map creation operation. This is done only if
|
|
* kernel is too old to support memcg-based memory accounting for BPF
|
|
* subsystem. By default, RLIMIT_MEMLOCK limit is set to RLIM_INFINITY,
|
|
* but it can be overriden with libbpf_set_memlock_rlim_max() API.
|
|
* Note that libbpf_set_memlock_rlim_max() needs to be called before
|
|
* the very first bpf_prog_load(), bpf_map_create() or bpf_object__load()
|
|
* operation.
|
|
*/
|
|
LIBBPF_STRICT_AUTO_RLIMIT_MEMLOCK = 0x10,
|
|
|
|
__LIBBPF_STRICT_LAST,
|
|
};
|
|
|
|
LIBBPF_API int libbpf_set_strict_mode(enum libbpf_strict_mode mode);
|
|
|
|
#define DECLARE_LIBBPF_OPTS LIBBPF_OPTS
|
|
|
|
#ifdef __cplusplus
|
|
} /* extern "C" */
|
|
#endif
|
|
|
|
#endif /* __LIBBPF_LEGACY_BPF_H */
|