diff options
Diffstat (limited to 'tools/lib')
| -rw-r--r-- | tools/lib/bpf/Build | 2 | ||||
| -rw-r--r-- | tools/lib/bpf/Makefile | 3 | ||||
| -rw-r--r-- | tools/lib/bpf/bpf.c | 127 | ||||
| -rw-r--r-- | tools/lib/bpf/bpf.h | 23 | ||||
| -rw-r--r-- | tools/lib/bpf/btf.c | 373 | ||||
| -rw-r--r-- | tools/lib/bpf/btf.h | 22 | ||||
| -rw-r--r-- | tools/lib/bpf/libbpf.c | 289 | ||||
| -rw-r--r-- | tools/lib/bpf/libbpf.h | 69 |
8 files changed, 840 insertions, 68 deletions
diff --git a/tools/lib/bpf/Build b/tools/lib/bpf/Build index 64c679d67109..6070e655042d 100644 --- a/tools/lib/bpf/Build +++ b/tools/lib/bpf/Build @@ -1 +1 @@ -libbpf-y := libbpf.o bpf.o nlattr.o +libbpf-y := libbpf.o bpf.o nlattr.o btf.o diff --git a/tools/lib/bpf/Makefile b/tools/lib/bpf/Makefile index e6d5f8d1477f..5390e7725e43 100644 --- a/tools/lib/bpf/Makefile +++ b/tools/lib/bpf/Makefile @@ -69,7 +69,7 @@ FEATURE_USER = .libbpf FEATURE_TESTS = libelf libelf-getphdrnum libelf-mmap bpf FEATURE_DISPLAY = libelf bpf -INCLUDES = -I. -I$(srctree)/tools/include -I$(srctree)/tools/arch/$(ARCH)/include/uapi -I$(srctree)/tools/include/uapi +INCLUDES = -I. -I$(srctree)/tools/include -I$(srctree)/tools/arch/$(ARCH)/include/uapi -I$(srctree)/tools/include/uapi -I$(srctree)/tools/perf FEATURE_CHECK_CFLAGS-bpf = $(INCLUDES) check_feat := 1 @@ -189,6 +189,7 @@ install_headers: $(call QUIET_INSTALL, headers) \ $(call do_install,bpf.h,$(prefix)/include/bpf,644); \ $(call do_install,libbpf.h,$(prefix)/include/bpf,644); + $(call do_install,btf.h,$(prefix)/include/bpf,644); install: install_lib diff --git a/tools/lib/bpf/bpf.c b/tools/lib/bpf/bpf.c index acbb3f8b3bec..9ddc89dae962 100644 --- a/tools/lib/bpf/bpf.c +++ b/tools/lib/bpf/bpf.c @@ -73,43 +73,77 @@ static inline int sys_bpf(enum bpf_cmd cmd, union bpf_attr *attr, return syscall(__NR_bpf, cmd, attr, size); } -int bpf_create_map_node(enum bpf_map_type map_type, const char *name, - int key_size, int value_size, int max_entries, - __u32 map_flags, int node) +int bpf_create_map_xattr(const struct bpf_create_map_attr *create_attr) { - __u32 name_len = name ? strlen(name) : 0; + __u32 name_len = create_attr->name ? strlen(create_attr->name) : 0; union bpf_attr attr; memset(&attr, '\0', sizeof(attr)); - attr.map_type = map_type; - attr.key_size = key_size; - attr.value_size = value_size; - attr.max_entries = max_entries; - attr.map_flags = map_flags; - memcpy(attr.map_name, name, min(name_len, BPF_OBJ_NAME_LEN - 1)); + attr.map_type = create_attr->map_type; + attr.key_size = create_attr->key_size; + attr.value_size = create_attr->value_size; + attr.max_entries = create_attr->max_entries; + attr.map_flags = create_attr->map_flags; + memcpy(attr.map_name, create_attr->name, + min(name_len, BPF_OBJ_NAME_LEN - 1)); + attr.numa_node = create_attr->numa_node; + attr.btf_fd = create_attr->btf_fd; + attr.btf_key_type_id = create_attr->btf_key_type_id; + attr.btf_value_type_id = create_attr->btf_value_type_id; + attr.map_ifindex = create_attr->map_ifindex; + + return sys_bpf(BPF_MAP_CREATE, &attr, sizeof(attr)); +} +int bpf_create_map_node(enum bpf_map_type map_type, const char *name, + int key_size, int value_size, int max_entries, + __u32 map_flags, int node) +{ + struct bpf_create_map_attr map_attr = {}; + + map_attr.name = name; + map_attr.map_type = map_type; + map_attr.map_flags = map_flags; + map_attr.key_size = key_size; + map_attr.value_size = value_size; + map_attr.max_entries = max_entries; if (node >= 0) { - attr.map_flags |= BPF_F_NUMA_NODE; - attr.numa_node = node; + map_attr.numa_node = node; + map_attr.map_flags |= BPF_F_NUMA_NODE; } - return sys_bpf(BPF_MAP_CREATE, &attr, sizeof(attr)); + return bpf_create_map_xattr(&map_attr); } int bpf_create_map(enum bpf_map_type map_type, int key_size, int value_size, int max_entries, __u32 map_flags) { - return bpf_create_map_node(map_type, NULL, key_size, value_size, - max_entries, map_flags, -1); + struct bpf_create_map_attr map_attr = {}; + + map_attr.map_type = map_type; + map_attr.map_flags = map_flags; + map_attr.key_size = key_size; + map_attr.value_size = value_size; + map_attr.max_entries = max_entries; + + return bpf_create_map_xattr(&map_attr); } int bpf_create_map_name(enum bpf_map_type map_type, const char *name, int key_size, int value_size, int max_entries, __u32 map_flags) { - return bpf_create_map_node(map_type, name, key_size, value_size, - max_entries, map_flags, -1); + struct bpf_create_map_attr map_attr = {}; + + map_attr.name = name; + map_attr.map_type = map_type; + map_attr.map_flags = map_flags; + map_attr.key_size = key_size; + map_attr.value_size = value_size; + map_attr.max_entries = max_entries; + + return bpf_create_map_xattr(&map_attr); } int bpf_create_map_in_map_node(enum bpf_map_type map_type, const char *name, @@ -168,6 +202,7 @@ int bpf_load_program_xattr(const struct bpf_load_program_attr *load_attr, attr.log_size = 0; attr.log_level = 0; attr.kern_version = load_attr->kern_version; + attr.prog_ifindex = load_attr->prog_ifindex; memcpy(attr.prog_name, load_attr->name, min(name_len, BPF_OBJ_NAME_LEN - 1)); @@ -425,6 +460,16 @@ int bpf_map_get_fd_by_id(__u32 id) return sys_bpf(BPF_MAP_GET_FD_BY_ID, &attr, sizeof(attr)); } +int bpf_btf_get_fd_by_id(__u32 id) +{ + union bpf_attr attr; + + bzero(&attr, sizeof(attr)); + attr.btf_id = id; + + return sys_bpf(BPF_BTF_GET_FD_BY_ID, &attr, sizeof(attr)); +} + int bpf_obj_get_info_by_fd(int prog_fd, void *info, __u32 *info_len) { union bpf_attr attr; @@ -573,3 +618,51 @@ cleanup: close(sock); return ret; } + +int bpf_load_btf(void *btf, __u32 btf_size, char *log_buf, __u32 log_buf_size, + bool do_log) +{ + union bpf_attr attr = {}; + int fd; + + attr.btf = ptr_to_u64(btf); + attr.btf_size = btf_size; + +retry: + if (do_log && log_buf && log_buf_size) { + attr.btf_log_level = 1; + attr.btf_log_size = log_buf_size; + attr.btf_log_buf = ptr_to_u64(log_buf); + } + + fd = sys_bpf(BPF_BTF_LOAD, &attr, sizeof(attr)); + if (fd == -1 && !do_log && log_buf && log_buf_size) { + do_log = true; + goto retry; + } + + return fd; +} + +int bpf_task_fd_query(int pid, int fd, __u32 flags, char *buf, __u32 *buf_len, + __u32 *prog_id, __u32 *fd_type, __u64 *probe_offset, + __u64 *probe_addr) +{ + union bpf_attr attr = {}; + int err; + + attr.task_fd_query.pid = pid; + attr.task_fd_query.fd = fd; + attr.task_fd_query.flags = flags; + attr.task_fd_query.buf = ptr_to_u64(buf); + attr.task_fd_query.buf_len = *buf_len; + + err = sys_bpf(BPF_TASK_FD_QUERY, &attr, sizeof(attr)); + *buf_len = attr.task_fd_query.buf_len; + *prog_id = attr.task_fd_query.prog_id; + *fd_type = attr.task_fd_query.fd_type; + *probe_offset = attr.task_fd_query.probe_offset; + *probe_addr = attr.task_fd_query.probe_addr; + + return err; +} diff --git a/tools/lib/bpf/bpf.h b/tools/lib/bpf/bpf.h index 39f6a0d64a3b..0639a30a457d 100644 --- a/tools/lib/bpf/bpf.h +++ b/tools/lib/bpf/bpf.h @@ -24,8 +24,24 @@ #define __BPF_BPF_H #include <linux/bpf.h> +#include <stdbool.h> #include <stddef.h> +struct bpf_create_map_attr { + const char *name; + enum bpf_map_type map_type; + __u32 map_flags; + __u32 key_size; + __u32 value_size; + __u32 max_entries; + __u32 numa_node; + __u32 btf_fd; + __u32 btf_key_type_id; + __u32 btf_value_type_id; + __u32 map_ifindex; +}; + +int bpf_create_map_xattr(const struct bpf_create_map_attr *create_attr); int bpf_create_map_node(enum bpf_map_type map_type, const char *name, int key_size, int value_size, int max_entries, __u32 map_flags, int node); @@ -49,6 +65,7 @@ struct bpf_load_program_attr { size_t insns_cnt; const char *license; __u32 kern_version; + __u32 prog_ifindex; }; /* Recommend log buffer size */ @@ -83,8 +100,14 @@ int bpf_prog_get_next_id(__u32 start_id, __u32 *next_id); int bpf_map_get_next_id(__u32 start_id, __u32 *next_id); int bpf_prog_get_fd_by_id(__u32 id); int bpf_map_get_fd_by_id(__u32 id); +int bpf_btf_get_fd_by_id(__u32 id); int bpf_obj_get_info_by_fd(int prog_fd, void *info, __u32 *info_len); int bpf_prog_query(int target_fd, enum bpf_attach_type type, __u32 query_flags, __u32 *attach_flags, __u32 *prog_ids, __u32 *prog_cnt); int bpf_raw_tracepoint_open(const char *name, int prog_fd); +int bpf_load_btf(void *btf, __u32 btf_size, char *log_buf, __u32 log_buf_size, + bool do_log); +int bpf_task_fd_query(int pid, int fd, __u32 flags, char *buf, __u32 *buf_len, + __u32 *prog_id, __u32 *fd_type, __u64 *probe_offset, + __u64 *probe_addr); #endif diff --git a/tools/lib/bpf/btf.c b/tools/lib/bpf/btf.c new file mode 100644 index 000000000000..8c54a4b6f187 --- /dev/null +++ b/tools/lib/bpf/btf.c @@ -0,0 +1,373 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2018 Facebook */ + +#include <stdlib.h> +#include <stdint.h> +#include <string.h> +#include <unistd.h> +#include <errno.h> +#include <linux/err.h> +#include <linux/btf.h> +#include "btf.h" +#include "bpf.h" + +#define elog(fmt, ...) { if (err_log) err_log(fmt, ##__VA_ARGS__); } +#define max(a, b) ((a) > (b) ? (a) : (b)) +#define min(a, b) ((a) < (b) ? (a) : (b)) + +#define BTF_MAX_NR_TYPES 65535 + +static struct btf_type btf_void; + +struct btf { + union { + struct btf_header *hdr; + void *data; + }; + struct btf_type **types; + const char *strings; + void *nohdr_data; + uint32_t nr_types; + uint32_t types_size; + uint32_t data_size; + int fd; +}; + +static const char *btf_name_by_offset(const struct btf *btf, uint32_t offset) +{ + if (offset < btf->hdr->str_len) + return &btf->strings[offset]; + else + return NULL; +} + +static int btf_add_type(struct btf *btf, struct btf_type *t) +{ + if (btf->types_size - btf->nr_types < 2) { + struct btf_type **new_types; + u32 expand_by, new_size; + + if (btf->types_size == BTF_MAX_NR_TYPES) + return -E2BIG; + + expand_by = max(btf->types_size >> 2, 16); + new_size = min(BTF_MAX_NR_TYPES, btf->types_size + expand_by); + + new_types = realloc(btf->types, sizeof(*new_types) * new_size); + if (!new_types) + return -ENOMEM; + + if (btf->nr_types == 0) + new_types[0] = &btf_void; + + btf->types = new_types; + btf->types_size = new_size; + } + + btf->types[++(btf->nr_types)] = t; + + return 0; +} + +static int btf_parse_hdr(struct btf *btf, btf_print_fn_t err_log) +{ + const struct btf_header *hdr = btf->hdr; + u32 meta_left; + + if (btf->data_size < sizeof(struct btf_header)) { + elog("BTF header not found\n"); + return -EINVAL; + } + + if (hdr->magic != BTF_MAGIC) { + elog("Invalid BTF magic:%x\n", hdr->magic); + return -EINVAL; + } + + if (hdr->version != BTF_VERSION) { + elog("Unsupported BTF version:%u\n", hdr->version); + return -ENOTSUP; + } + + if (hdr->flags) { + elog("Unsupported BTF flags:%x\n", hdr->flags); + return -ENOTSUP; + } + + meta_left = btf->data_size - sizeof(*hdr); + if (!meta_left) { + elog("BTF has no data\n"); + return -EINVAL; + } + + if (meta_left < hdr->type_off) { + elog("Invalid BTF type section offset:%u\n", hdr->type_off); + return -EINVAL; + } + + if (meta_left < hdr->str_off) { + elog("Invalid BTF string section offset:%u\n", hdr->str_off); + return -EINVAL; + } + + if (hdr->type_off >= hdr->str_off) { + elog("BTF type section offset >= string section offset. No type?\n"); + return -EINVAL; + } + + if (hdr->type_off & 0x02) { + elog("BTF type section is not aligned to 4 bytes\n"); + return -EINVAL; + } + + btf->nohdr_data = btf->hdr + 1; + + return 0; +} + +static int btf_parse_str_sec(struct btf *btf, btf_print_fn_t err_log) +{ + const struct btf_header *hdr = btf->hdr; + const char *start = btf->nohdr_data + hdr->str_off; + const char *end = start + btf->hdr->str_len; + + if (!hdr->str_len || hdr->str_len - 1 > BTF_MAX_NAME_OFFSET || + start[0] || end[-1]) { + elog("Invalid BTF string section\n"); + return -EINVAL; + } + + btf->strings = start; + + return 0; +} + +static int btf_parse_type_sec(struct btf *btf, btf_print_fn_t err_log) +{ + struct btf_header *hdr = btf->hdr; + void *nohdr_data = btf->nohdr_data; + void *next_type = nohdr_data + hdr->type_off; + void *end_type = nohdr_data + hdr->str_off; + + while (next_type < end_type) { + struct btf_type *t = next_type; + uint16_t vlen = BTF_INFO_VLEN(t->info); + int err; + + next_type += sizeof(*t); + switch (BTF_INFO_KIND(t->info)) { + case BTF_KIND_INT: + next_type += sizeof(int); + break; + case BTF_KIND_ARRAY: + next_type += sizeof(struct btf_array); + break; + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: + next_type += vlen * sizeof(struct btf_member); + break; + case BTF_KIND_ENUM: + next_type += vlen * sizeof(struct btf_enum); + break; + case BTF_KIND_TYPEDEF: + case BTF_KIND_PTR: + case BTF_KIND_FWD: + case BTF_KIND_VOLATILE: + case BTF_KIND_CONST: + case BTF_KIND_RESTRICT: + break; + default: + elog("Unsupported BTF_KIND:%u\n", + BTF_INFO_KIND(t->info)); + return -EINVAL; + } + + err = btf_add_type(btf, t); + if (err) + return err; + } + + return 0; +} + +static const struct btf_type *btf_type_by_id(const struct btf *btf, + uint32_t type_id) +{ + if (type_id > btf->nr_types) + return NULL; + + return btf->types[type_id]; +} + +static bool btf_type_is_void(const struct btf_type *t) +{ + return t == &btf_void || BTF_INFO_KIND(t->info) == BTF_KIND_FWD; +} + +static bool btf_type_is_void_or_null(const struct btf_type *t) +{ + return !t || btf_type_is_void(t); +} + +static int64_t btf_type_size(const struct btf_type *t) +{ + switch (BTF_INFO_KIND(t->info)) { + case BTF_KIND_INT: + case BTF_KIND_STRUCT: + case BTF_KIND_UNION: + case BTF_KIND_ENUM: + return t->size; + case BTF_KIND_PTR: + return sizeof(void *); + default: + return -EINVAL; + } +} + +#define MAX_RESOLVE_DEPTH 32 + +int64_t btf__resolve_size(const struct btf *btf, uint32_t type_id) +{ + const struct btf_array *array; + const struct btf_type *t; + uint32_t nelems = 1; + int64_t size = -1; + int i; + + t = btf_type_by_id(btf, type_id); + for (i = 0; i < MAX_RESOLVE_DEPTH && !btf_type_is_void_or_null(t); + i++) { + size = btf_type_size(t); + if (size >= 0) + break; + + switch (BTF_INFO_KIND(t->info)) { + case BTF_KIND_TYPEDEF: + case BTF_KIND_VOLATILE: + case BTF_KIND_CONST: + case BTF_KIND_RESTRICT: + type_id = t->type; + break; + case BTF_KIND_ARRAY: + array = (const struct btf_array *)(t + 1); + if (nelems && array->nelems > UINT32_MAX / nelems) + return -E2BIG; + nelems *= array->nelems; + type_id = array->type; + break; + default: + return -EINVAL; + } + + t = btf_type_by_id(btf, type_id); + } + + if (size < 0) + return -EINVAL; + + if (nelems && size > UINT32_MAX / nelems) + return -E2BIG; + + return nelems * size; +} + +int32_t btf__find_by_name(const struct btf *btf, const char *type_name) +{ + uint32_t i; + + if (!strcmp(type_name, "void")) + return 0; + + for (i = 1; i <= btf->nr_types; i++) { + const struct btf_type *t = btf->types[i]; + const char *name = btf_name_by_offset(btf, t->name_off); + + if (name && !strcmp(type_name, name)) + return i; + } + + return -ENOENT; +} + +void btf__free(struct btf *btf) +{ + if (!btf) + return; + + if (btf->fd != -1) + close(btf->fd); + + free(btf->data); + free(btf->types); + free(btf); +} + +struct btf *btf__new(uint8_t *data, uint32_t size, + btf_print_fn_t err_log) +{ + uint32_t log_buf_size = 0; + char *log_buf = NULL; + struct btf *btf; + int err; + + btf = calloc(1, sizeof(struct btf)); + if (!btf) + return ERR_PTR(-ENOMEM); + + btf->fd = -1; + + if (err_log) { + log_buf = malloc(BPF_LOG_BUF_SIZE); + if (!log_buf) { + err = -ENOMEM; + goto done; + } + *log_buf = 0; + log_buf_size = BPF_LOG_BUF_SIZE; + } + + btf->data = malloc(size); + if (!btf->data) { + err = -ENOMEM; + goto done; + } + + memcpy(btf->data, data, size); + btf->data_size = size; + + btf->fd = bpf_load_btf(btf->data, btf->data_size, + log_buf, log_buf_size, false); + + if (btf->fd == -1) { + err = -errno; + elog("Error loading BTF: %s(%d)\n", strerror(errno), errno); + if (log_buf && *log_buf) + elog("%s\n", log_buf); + goto done; + } + + err = btf_parse_hdr(btf, err_log); + if (err) + goto done; + + err = btf_parse_str_sec(btf, err_log); + if (err) + goto done; + + err = btf_parse_type_sec(btf, err_log); + +done: + free(log_buf); + + if (err) { + btf__free(btf); + return ERR_PTR(err); + } + + return btf; +} + +int btf__fd(const struct btf *btf) +{ + return btf->fd; +} diff --git a/tools/lib/bpf/btf.h b/tools/lib/bpf/btf.h new file mode 100644 index 000000000000..74bb344035bb --- /dev/null +++ b/tools/lib/bpf/btf.h @@ -0,0 +1,22 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2018 Facebook */ + +#ifndef __BPF_BTF_H +#define __BPF_BTF_H + +#include <stdint.h> + +#define BTF_ELF_SEC ".BTF" + +struct btf; + +typedef int (*btf_print_fn_t)(const char *, ...) + __attribute__((format(printf, 1, 2))); + +void btf__free(struct btf *btf); +struct btf *btf__new(uint8_t *data, uint32_t size, btf_print_fn_t err_log); +int32_t btf__find_by_name(const struct btf *btf, const char *type_name); +int64_t btf__resolve_size(const struct btf *btf, uint32_t type_id); +int btf__fd(const struct btf *btf); + +#endif diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c index 0f9f06df49bc..a1e96b5de5ff 100644 --- a/tools/lib/bpf/libbpf.c +++ b/tools/lib/bpf/libbpf.c @@ -31,6 +31,7 @@ #include <unistd.h> #include <fcntl.h> #include <errno.h> +#include <perf-sys.h> #include <asm/unistd.h> #include <linux/err.h> #include <linux/kernel.h> @@ -45,6 +46,7 @@ #include "libbpf.h" #include "bpf.h" +#include "btf.h" #ifndef EM_BPF #define EM_BPF 247 @@ -176,6 +178,7 @@ struct bpf_program { /* Index in elf obj file, for relocation use. */ int idx; char *name; + int prog_ifindex; char *section_name; struct bpf_insn *insns; size_t insns_cnt, main_prog_cnt; @@ -211,7 +214,10 @@ struct bpf_map { int fd; char *name; size_t offset; + int map_ifindex; struct bpf_map_def def; + uint32_t btf_key_type_id; + uint32_t btf_value_type_id; void *priv; bpf_map_clear_priv_t clear_priv; }; @@ -256,6 +262,8 @@ struct bpf_object { */ struct list_head list; + struct btf *btf; + void *priv; bpf_object_clear_priv_t clear_priv; @@ -819,7 +827,15 @@ static int bpf_object__elf_collect(struct bpf_object *obj) data->d_size); else if (strcmp(name, "maps") == 0) obj->efile.maps_shndx = idx; - else if (sh.sh_type == SHT_SYMTAB) { + else if (strcmp(name, BTF_ELF_SEC) == 0) { + obj->btf = btf__new(data->d_buf, data->d_size, + __pr_debug); + if (IS_ERR(obj->btf)) { + pr_warning("Error loading ELF section %s: %ld. Ignored and continue.\n", + BTF_ELF_SEC, PTR_ERR(obj->btf)); + obj->btf = NULL; + } + } else if (sh.sh_type == SHT_SYMTAB) { if (obj->efile.symbols) { pr_warning("bpf: multiple SYMTAB in %s\n", obj->path); @@ -996,33 +1012,127 @@ bpf_program__collect_reloc(struct bpf_program *prog, GElf_Shdr *shdr, return 0; } +static int bpf_map_find_btf_info(struct bpf_map *map, const struct btf *btf) +{ + struct bpf_map_def *def = &map->def; + const size_t max_name = 256; + int64_t key_size, value_size; + int32_t key_id, value_id; + char name[max_name]; + + /* Find key type by name from BTF */ + if (snprintf(name, max_name, "%s_key", map->name) == max_name) { + pr_warning("map:%s length of BTF key_type:%s_key is too long\n", + map->name, map->name); + return -EINVAL; + } + + key_id = btf__find_by_name(btf, name); + if (key_id < 0) { + pr_debug("map:%s key_type:%s cannot be found in BTF\n", + map->name, name); + return key_id; + } + + key_size = btf__resolve_size(btf, key_id); + if (key_size < 0) { + pr_warning("map:%s key_type:%s cannot get the BTF type_size\n", + map->name, name); + return key_size; + } + + if (def->key_size != key_size) { + pr_warning("map:%s key_type:%s has BTF type_size:%u != key_size:%u\n", + map->name, name, (unsigned int)key_size, def->key_size); + return -EINVAL; + } + + /* Find value type from BTF */ + if (snprintf(name, max_name, "%s_value", map->name) == max_name) { + pr_warning("map:%s length of BTF value_type:%s_value is too long\n", + map->name, map->name); + return -EINVAL; + } + + value_id = btf__find_by_name(btf, name); + if (value_id < 0) { + pr_debug("map:%s value_type:%s cannot be found in BTF\n", + map->name, name); + return value_id; + } + + value_size = btf__resolve_size(btf, value_id); + if (value_size < 0) { + pr_warning("map:%s value_type:%s cannot get the BTF type_size\n", + map->name, name); + return value_size; + } + + if (def->value_size != value_size) { + pr_warning("map:%s value_type:%s has BTF type_size:%u != value_size:%u\n", + map->name, name, (unsigned int)value_size, def->value_size); + return -EINVAL; + } + + map->btf_key_type_id = key_id; + map->btf_value_type_id = value_id; + + return 0; +} + static int bpf_object__create_maps(struct bpf_object *obj) { + struct bpf_create_map_attr create_attr = {}; unsigned int i; + int err; for (i = 0; i < obj->nr_maps; i++) { - struct bpf_map_def *def = &obj->maps[i].def; - int *pfd = &obj->maps[i].fd; - - *pfd = bpf_create_map_name(def->type, - obj->maps[i].name, - def->key_size, - def->value_size, - def->max_entries, - def->map_flags); + struct bpf_map *map = &obj->maps[i]; + struct bpf_map_def *def = &map->def; + int *pfd = &map->fd; + + create_attr.name = map->name; + create_attr.map_ifindex = map->map_ifindex; + create_attr.map_type = def->type; + create_attr.map_flags = def->map_flags; + create_attr.key_size = def->key_size; + create_attr.value_size = def->value_size; + create_attr.max_entries = def->max_entries; + create_attr.btf_fd = 0; + create_attr.btf_key_type_id = 0; + create_attr.btf_value_type_id = 0; + + if (obj->btf && !bpf_map_find_btf_info(map, obj->btf)) { + create_attr.btf_fd = btf__fd(obj->btf); + create_attr.btf_key_type_id = map->btf_key_type_id; + create_attr.btf_value_type_id = map->btf_value_type_id; + } + + *pfd = bpf_create_map_xattr(&create_attr); + if (*pfd < 0 && create_attr.btf_key_type_id) { + pr_warning("Error in bpf_create_map_xattr(%s):%s(%d). Retrying without BTF.\n", + map->name, strerror(errno), errno); + create_attr.btf_fd = 0; + create_attr.btf_key_type_id = 0; + create_attr.btf_value_type_id = 0; + map->btf_key_type_id = 0; + map->btf_value_type_id = 0; + *pfd = bpf_create_map_xattr(&create_attr); + } + if (*pfd < 0) { size_t j; - int err = *pfd; + err = *pfd; pr_warning("failed to create map (name: '%s'): %s\n", - obj->maps[i].name, + map->name, strerror(errno)); for (j = 0; j < i; j++) zclose(obj->maps[j].fd); return err; } - pr_debug("create map %s: fd=%d\n", obj->maps[i].name, *pfd); + pr_debug("create map %s: fd=%d\n", map->name, *pfd); } return 0; @@ -1166,7 +1276,7 @@ static int bpf_object__collect_reloc(struct bpf_object *obj) static int load_program(enum bpf_prog_type type, enum bpf_attach_type expected_attach_type, const char *name, struct bpf_insn *insns, int insns_cnt, - char *license, u32 kern_version, int *pfd) + char *license, u32 kern_version, int *pfd, int prog_ifindex) { struct bpf_load_program_attr load_attr; char *log_buf; @@ -1180,6 +1290,7 @@ load_program(enum bpf_prog_type type, enum bpf_attach_type expected_attach_type, load_attr.insns_cnt = insns_cnt; load_attr.license = license; load_attr.kern_version = kern_version; + load_attr.prog_ifindex = prog_ifindex; if (!load_attr.insns || !load_attr.insns_cnt) return -EINVAL; @@ -1261,7 +1372,8 @@ bpf_program__load(struct bpf_program *prog, } err = load_program(prog->type, prog->expected_attach_type, prog->name, prog->insns, prog->insns_cnt, - license, kern_version, &fd); + license, kern_version, &fd, + prog->prog_ifindex); if (!err) prog->instances.fds[0] = fd; goto out; @@ -1292,7 +1404,8 @@ bpf_program__load(struct bpf_program *prog, err = load_program(prog->type, prog->expected_attach_type, prog->name, result.new_insn_ptr, result.new_insn_cnt, - license, kern_version, &fd); + license, kern_version, &fd, + prog->prog_ifindex); if (err) { pr_warning("Loading the %dth instance of program '%s' failed\n", @@ -1331,9 +1444,39 @@ bpf_object__load_progs(struct bpf_object *obj) return 0; } -static int bpf_object__validate(struct bpf_object *obj) +static bool bpf_prog_type__needs_kver(enum bpf_prog_type type) +{ + switch (type) { + case BPF_PROG_TYPE_SOCKET_FILTER: + case BPF_PROG_TYPE_SCHED_CLS: + case BPF_PROG_TYPE_SCHED_ACT: + case BPF_PROG_TYPE_XDP: + case BPF_PROG_TYPE_CGROUP_SKB: + case BPF_PROG_TYPE_CGROUP_SOCK: + case BPF_PROG_TYPE_LWT_IN: + case BPF_PROG_TYPE_LWT_OUT: + case BPF_PROG_TYPE_LWT_XMIT: + case BPF_PROG_TYPE_LWT_SEG6LOCAL: + case BPF_PROG_TYPE_SOCK_OPS: + case BPF_PROG_TYPE_SK_SKB: + case BPF_PROG_TYPE_CGROUP_DEVICE: + case BPF_PROG_TYPE_SK_MSG: + case BPF_PROG_TYPE_CGROUP_SOCK_ADDR: + case BPF_PROG_TYPE_LIRC_MODE2: + return false; + case BPF_PROG_TYPE_UNSPEC: + case BPF_PROG_TYPE_KPROBE: + case BPF_PROG_TYPE_TRACEPOINT: + case BPF_PROG_TYPE_PERF_EVENT: + case BPF_PROG_TYPE_RAW_TRACEPOINT: + default: + return true; + } +} + +static int bpf_object__validate(struct bpf_object *obj, bool needs_kver) { - if (obj->kern_version == 0) { + if (needs_kver && obj->kern_version == 0) { pr_warning("%s doesn't provide kernel version\n", obj->path); return -LIBBPF_ERRNO__KVERSION; @@ -1342,7 +1485,8 @@ static int bpf_object__validate(struct bpf_object *obj) } static struct bpf_object * -__bpf_object__open(const char *path, void *obj_buf, size_t obj_buf_sz) +__bpf_object__open(const char *path, void *obj_buf, size_t obj_buf_sz, + bool needs_kver) { struct bpf_object *obj; int err; @@ -1360,7 +1504,7 @@ __bpf_object__open(const char *path, void *obj_buf, size_t obj_buf_sz) CHECK_ERR(bpf_object__check_endianness(obj), err, out); CHECK_ERR(bpf_object__elf_collect(obj), err, out); CHECK_ERR(bpf_object__collect_reloc(obj), err, out); - CHECK_ERR(bpf_object__validate(obj), err, out); + CHECK_ERR(bpf_object__validate(obj, needs_kver), err, out); bpf_object__elf_finish(obj); return obj; @@ -1377,7 +1521,7 @@ struct bpf_object *bpf_object__open(const char *path) pr_debug("loading %s\n", path); - return __bpf_object__open(path, NULL, 0); + return __bpf_object__open(path, NULL, 0, true); } struct bpf_object *bpf_object__open_buffer(void *obj_buf, @@ -1400,7 +1544,7 @@ struct bpf_object *bpf_object__open_buffer(void *obj_buf, pr_debug("loading object '%s' from buffer\n", name); - return __bpf_object__open(name, obj_buf, obj_buf_sz); + return __bpf_object__open(name, obj_buf, obj_buf_sz, true); } int bpf_object__unload(struct bpf_object *obj) @@ -1641,6 +1785,7 @@ void bpf_object__close(struct bpf_object *obj) bpf_object__elf_finish(obj); bpf_object__unload(obj); + btf__free(obj->btf); for (i = 0; i < obj->nr_maps; i++) { zfree(&obj->maps[i].name); @@ -1692,6 +1837,11 @@ unsigned int bpf_object__kversion(struct bpf_object *obj) return obj ? obj->kern_version : 0; } +int bpf_object__btf_fd(const struct bpf_object *obj) +{ + return obj->btf ? btf__fd(obj->btf) : -1; +} + int bpf_object__set_priv(struct bpf_object *obj, void *priv, bpf_object_clear_priv_t clear_priv) { @@ -1845,11 +1995,12 @@ BPF_PROG_TYPE_FNS(kprobe, BPF_PROG_TYPE_KPROBE); BPF_PROG_TYPE_FNS(sched_cls, BPF_PROG_TYPE_SCHED_CLS); BPF_PROG_TYPE_FNS(sched_act, BPF_PROG_TYPE_SCHED_ACT); BPF_PROG_TYPE_FNS(tracepoint, BPF_PROG_TYPE_TRACEPOINT); +BPF_PROG_TYPE_FNS(raw_tracepoint, BPF_PROG_TYPE_RAW_TRACEPOINT); BPF_PROG_TYPE_FNS(xdp, BPF_PROG_TYPE_XDP); BPF_PROG_TYPE_FNS(perf_event, BPF_PROG_TYPE_PERF_EVENT); -static void bpf_program__set_expected_attach_type(struct bpf_program *prog, - enum bpf_attach_type type) +void bpf_program__set_expected_attach_type(struct bpf_program *prog, + enum bpf_attach_type type) { prog->expected_attach_type = type; } @@ -1859,6 +2010,9 @@ static void bpf_program__set_expected_attach_type(struct bpf_program *prog, #define BPF_PROG_SEC(string, ptype) BPF_PROG_SEC_FULL(string, ptype, 0) +#define BPF_S_PROG_SEC(string, ptype) \ + BPF_PROG_SEC_FULL(string, BPF_PROG_TYPE_CGROUP_SOCK, ptype) + #define BPF_SA_PROG_SEC(string, ptype) \ BPF_PROG_SEC_FULL(string, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, ptype) @@ -1874,6 +2028,7 @@ static const struct { BPF_PROG_SEC("classifier", BPF_PROG_TYPE_SCHED_CLS), BPF_PROG_SEC("action", BPF_PROG_TYPE_SCHED_ACT), BPF_PROG_SEC("tracepoint/", BPF_PROG_TYPE_TRACEPOINT), + BPF_PROG_SEC("raw_tracepoint/", BPF_PROG_TYPE_RAW_TRACEPOINT), BPF_PROG_SEC("xdp", BPF_PROG_TYPE_XDP), BPF_PROG_SEC("perf_event", BPF_PROG_TYPE_PERF_EVENT), BPF_PROG_SEC("cgroup/skb", BPF_PROG_TYPE_CGROUP_SKB), @@ -1889,10 +2044,15 @@ static const struct { BPF_SA_PROG_SEC("cgroup/bind6", BPF_CGROUP_INET6_BIND), BPF_SA_PROG_SEC("cgroup/connect4", BPF_CGROUP_INET4_CONNECT), BPF_SA_PROG_SEC("cgroup/connect6", BPF_CGROUP_INET6_CONNECT), + BPF_SA_PROG_SEC("cgroup/sendmsg4", BPF_CGROUP_UDP4_SENDMSG), + BPF_SA_PROG_SEC("cgroup/sendmsg6", BPF_CGROUP_UDP6_SENDMSG), + BPF_S_PROG_SEC("cgroup/post_bind4", BPF_CGROUP_INET4_POST_BIND), + BPF_S_PROG_SEC("cgroup/post_bind6", BPF_CGROUP_INET6_POST_BIND), }; #undef BPF_PROG_SEC #undef BPF_PROG_SEC_FULL +#undef BPF_S_PROG_SEC #undef BPF_SA_PROG_SEC static int bpf_program__identify_section(struct bpf_program *prog) @@ -1929,6 +2089,16 @@ const char *bpf_map__name(struct bpf_map *map) return map ? map->name : NULL; } +uint32_t bpf_map__btf_key_type_id(const struct bpf_map *map) +{ + return map ? map->btf_key_type_id : 0; +} + +uint32_t bpf_map__btf_value_type_id(const struct bpf_map *map) +{ + return map ? map->btf_value_type_id : 0; +} + int bpf_map__set_priv(struct bpf_map *map, void *priv, bpf_map_clear_priv_t clear_priv) { @@ -2028,13 +2198,17 @@ int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr, enum bpf_attach_type expected_attach_type; enum bpf_prog_type prog_type; struct bpf_object *obj; + struct bpf_map *map; int section_idx; int err; if (!attr) return -EINVAL; + if (!attr->file) + return -EINVAL; - obj = bpf_object__open(attr->file); + obj = __bpf_object__open(attr->file, NULL, 0, + bpf_prog_type__needs_kver(attr->prog_type)); if (IS_ERR_OR_NULL(obj)) return -ENOENT; @@ -2044,6 +2218,7 @@ int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr, * section name. */ prog_type = attr->prog_type; + prog->prog_ifindex = attr->ifindex; expected_attach_type = attr->expected_attach_type; if (prog_type == BPF_PROG_TYPE_UNSPEC) { section_idx = bpf_program__identify_section(prog); @@ -2064,6 +2239,10 @@ int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr, first_prog = prog; } + bpf_map__for_each(map, obj) { + map->map_ifindex = attr->ifindex; + } + if (!first_prog) { pr_warning("object file doesn't contain bpf program\n"); bpf_object__close(obj); @@ -2080,3 +2259,63 @@ int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr, *prog_fd = bpf_program__fd(first_prog); return 0; } + +enum bpf_perf_event_ret +bpf_perf_event_read_simple(void *mem, unsigned long size, + unsigned long page_size, void **buf, size_t *buf_len, + bpf_perf_event_print_t fn, void *priv) +{ + volatile struct perf_event_mmap_page *header = mem; + __u64 data_tail = header->data_tail; + __u64 data_head = header->data_head; + void *base, *begin, *end; + int ret; + + asm volatile("" ::: "memory"); /* in real code it should be smp_rmb() */ + if (data_head == data_tail) + return LIBBPF_PERF_EVENT_CONT; + + base = ((char *)header) + page_size; + + begin = base + data_tail % size; + end = base + data_head % size; + + while (begin != end) { + struct perf_event_header *ehdr; + + ehdr = begin; + if (begin + ehdr->size > base + size) { + long len = base + size - begin; + + if (*buf_len < ehdr->size) { + free(*buf); + *buf = malloc(ehdr->size); + if (!*buf) { + ret = LIBBPF_PERF_EVENT_ERROR; + break; + } + *buf_len = ehdr->size; + } + + memcpy(*buf, begin, len); + memcpy(*buf + len, base, ehdr->size - len); + ehdr = (void *)*buf; + begin = base + ehdr->size - len; + } else if (begin + ehdr->size == base + size) { + begin = base; + } else { + begin += ehdr->size; + } + + ret = fn(ehdr, priv); + if (ret != LIBBPF_PERF_EVENT_CONT) + break; + + data_tail += ehdr->size; + } + + __sync_synchronize(); /* smp_mb() */ + header->data_tail = data_tail; + + return ret; +} diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h index a3a62a583f27..09976531aa74 100644 --- a/tools/lib/bpf/libbpf.h +++ b/tools/lib/bpf/libbpf.h @@ -52,8 +52,8 @@ enum libbpf_errno { int libbpf_strerror(int err, char *buf, size_t size); /* - * In include/linux/compiler-gcc.h, __printf is defined. However - * it should be better if libbpf.h doesn't depend on Linux header file. + * __printf is defined in include/linux/compiler-gcc.h. However, + * it would be better if libbpf.h didn't depend on Linux header files. * So instead of __printf, here we use gcc attribute directly. */ typedef int (*libbpf_print_fn_t)(const char *, ...) @@ -78,6 +78,7 @@ int bpf_object__load(struct bpf_object *obj); int bpf_object__unload(struct bpf_object *obj); const char *bpf_object__name(struct bpf_object *obj); unsigned int bpf_object__kversion(struct bpf_object *obj); +int bpf_object__btf_fd(const struct bpf_object *obj); struct bpf_object *bpf_object__next(struct bpf_object *prev); #define bpf_object__for_each_safe(pos, tmp) \ @@ -91,7 +92,7 @@ int bpf_object__set_priv(struct bpf_object *obj, void *priv, bpf_object_clear_priv_t clear_priv); void *bpf_object__priv(struct bpf_object *prog); -/* Accessors of bpf_program. */ +/* Accessors of bpf_program */ struct bpf_program; struct bpf_program *bpf_program__next(struct bpf_program *prog, struct bpf_object *obj); @@ -120,28 +121,28 @@ struct bpf_insn; /* * Libbpf allows callers to adjust BPF programs before being loaded - * into kernel. One program in an object file can be transform into - * multiple variants to be attached to different code. + * into kernel. One program in an object file can be transformed into + * multiple variants to be attached to different hooks. * * bpf_program_prep_t, bpf_program__set_prep and bpf_program__nth_fd - * are APIs for this propose. + * form an API for this purpose. * * - bpf_program_prep_t: - * It defines 'preprocessor', which is a caller defined function + * Defines a 'preprocessor', which is a caller defined function * passed to libbpf through bpf_program__set_prep(), and will be * called before program is loaded. The processor should adjust - * the program one time for each instances according to the number + * the program one time for each instance according to the instance id * passed to it. * * - bpf_program__set_prep: - * Attachs a preprocessor to a BPF program. The number of instances - * whould be created is also passed through this function. + * Attaches a preprocessor to a BPF program. The number of instances + * that should be created is also passed through this function. * * - bpf_program__nth_fd: - * After the program is loaded, get resuling fds from bpf program for - * each instances. + * After the program is loaded, get resulting FD of a given instance + * of the BPF program. * - * If bpf_program__set_prep() is not used, the program whould be loaded + * If bpf_program__set_prep() is not used, the program would be loaded * without adjustment during bpf_object__load(). The program has only * one instance. In this case bpf_program__fd(prog) is equal to * bpf_program__nth_fd(prog, 0). @@ -155,7 +156,7 @@ struct bpf_prog_prep_result { struct bpf_insn *new_insn_ptr; int new_insn_cnt; - /* If not NULL, result fd is set to it */ + /* If not NULL, result FD is written to it. */ int *pfd; }; @@ -168,8 +169,8 @@ struct bpf_prog_prep_result { * - res: Output parameter, result of transformation. * * Return value: - * - Zero: pre-processing success. - * - Non-zero: pre-processing, stop loading. + * - Zero: pre-processing success. + * - Non-zero: pre-processing error, stop loading. */ typedef int (*bpf_program_prep_t)(struct bpf_program *prog, int n, struct bpf_insn *insns, int insns_cnt, @@ -181,19 +182,23 @@ int bpf_program__set_prep(struct bpf_program *prog, int nr_instance, int bpf_program__nth_fd(struct bpf_program *prog, int n); /* - * Adjust type of bpf program. Default is kprobe. + * Adjust type of BPF program. Default is kprobe. */ int bpf_program__set_socket_filter(struct bpf_program *prog); int bpf_program__set_tracepoint(struct bpf_program *prog); +int bpf_program__set_raw_tracepoint(struct bpf_program *prog); int bpf_program__set_kprobe(struct bpf_program *prog); int bpf_program__set_sched_cls(struct bpf_program *prog); int bpf_program__set_sched_act(struct bpf_program *prog); int bpf_program__set_xdp(struct bpf_program *prog); int bpf_program__set_perf_event(struct bpf_program *prog); void bpf_program__set_type(struct bpf_program *prog, enum bpf_prog_type type); +void bpf_program__set_expected_attach_type(struct bpf_program *prog, + enum bpf_attach_type type); bool bpf_program__is_socket_filter(struct bpf_program *prog); bool bpf_program__is_tracepoint(struct bpf_program *prog); +bool bpf_program__is_raw_tracepoint(struct bpf_program *prog); bool bpf_program__is_kprobe(struct bpf_program *prog); bool bpf_program__is_sched_cls(struct bpf_program *prog); bool bpf_program__is_sched_act(struct bpf_program *prog); @@ -201,10 +206,10 @@ bool bpf_program__is_xdp(struct bpf_program *prog); bool bpf_program__is_perf_event(struct bpf_program *prog); /* - * We don't need __attribute__((packed)) now since it is - * unnecessary for 'bpf_map_def' because they are all aligned. - * In addition, using it will trigger -Wpacked warning message, - * and will be treated as an error due to -Werror. + * No need for __attribute__((packed)), all members of 'bpf_map_def' + * are all aligned. In addition, using __attribute__((packed)) + * would trigger a -Wpacked warning message, and lead to an error + * if -Werror is set. */ struct bpf_map_def { unsigned int type; @@ -215,8 +220,8 @@ struct bpf_map_def { }; /* - * There is another 'struct bpf_map' in include/linux/map.h. However, - * it is not a uapi header so no need to consider name clash. + * The 'struct bpf_map' in include/linux/bpf.h is internal to the kernel, + * so no need to worry about a name clash. */ struct bpf_map; struct bpf_map * @@ -224,7 +229,7 @@ bpf_object__find_map_by_name(struct bpf_object *obj, const char *name); /* * Get bpf_map through the offset of corresponding struct bpf_map_def - * in the bpf object file. + * in the BPF object file. */ struct bpf_map * bpf_object__find_map_by_offset(struct bpf_object *obj, size_t offset); @@ -239,6 +244,8 @@ bpf_map__next(struct bpf_map *map, struct bpf_object *obj); int bpf_map__fd(struct bpf_map *map); const struct bpf_map_def *bpf_map__def(struct bpf_map *map); const char *bpf_map__name(struct bpf_map *map); +uint32_t bpf_map__btf_key_type_id(const struct bpf_map *map); +uint32_t bpf_map__btf_value_type_id(const struct bpf_map *map); typedef void (*bpf_map_clear_priv_t)(struct bpf_map *, void *); int bpf_map__set_priv(struct bpf_map *map, void *priv, @@ -252,6 +259,7 @@ struct bpf_prog_load_attr { const char *file; enum bpf_prog_type prog_type; enum bpf_attach_type expected_attach_type; + int ifindex; }; int bpf_prog_load_xattr(const struct bpf_prog_load_attr *attr, @@ -260,4 +268,17 @@ int bpf_prog_load(const char *file, enum bpf_prog_type type, struct bpf_object **pobj, int *prog_fd); int bpf_set_link_xdp_fd(int ifindex, int fd, __u32 flags); + +enum bpf_perf_event_ret { + LIBBPF_PERF_EVENT_DONE = 0, + LIBBPF_PERF_EVENT_ERROR = -1, + LIBBPF_PERF_EVENT_CONT = -2, +}; + +typedef enum bpf_perf_event_ret (*bpf_perf_event_print_t)(void *event, + void *priv); +int bpf_perf_event_read_simple(void *mem, unsigned long size, + unsigned long page_size, + void **buf, size_t *buf_len, + bpf_perf_event_print_t fn, void *priv); #endif |