From: Joe Burton <jevburton@xxxxxxxxxx> Add two functions that fentry/fexit/fmod_ret programs can attach to: bpf_map_trace_update_elem bpf_map_trace_delete_elem These functions have the same arguments as bpf_map_{update,delete}_elem. Invoke these functions from the following map types: BPF_MAP_TYPE_ARRAY BPF_MAP_TYPE_PERCPU_ARRAY BPF_MAP_TYPE_HASH BPF_MAP_TYPE_PERCPU_HASH BPF_MAP_TYPE_LRU_HASH BPF_MAP_TYPE_LRU_PERCPU_HASH The only guarantee about these functions is that they are invoked before the corresponding action occurs. Other conditions may prevent the corresponding action from occurring after the function is invoked. Signed-off-by: Joe Burton <jevburton@xxxxxxxxxx> --- kernel/bpf/Makefile | 2 +- kernel/bpf/arraymap.c | 4 +++- kernel/bpf/hashtab.c | 20 +++++++++++++++++++- kernel/bpf/map_trace.c | 17 +++++++++++++++++ kernel/bpf/map_trace.h | 19 +++++++++++++++++++ 5 files changed, 59 insertions(+), 3 deletions(-) create mode 100644 kernel/bpf/map_trace.c create mode 100644 kernel/bpf/map_trace.h diff --git a/kernel/bpf/Makefile b/kernel/bpf/Makefile index c1a9be6a4b9f..0cf38dab339a 100644 --- a/kernel/bpf/Makefile +++ b/kernel/bpf/Makefile @@ -9,7 +9,7 @@ CFLAGS_core.o += $(call cc-disable-warning, override-init) $(cflags-nogcse-yy) obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o bpf_iter.o map_iter.o task_iter.o prog_iter.o obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o bloom_filter.o obj-$(CONFIG_BPF_SYSCALL) += local_storage.o queue_stack_maps.o ringbuf.o -obj-$(CONFIG_BPF_SYSCALL) += bpf_local_storage.o bpf_task_storage.o +obj-$(CONFIG_BPF_SYSCALL) += bpf_local_storage.o bpf_task_storage.o map_trace.o obj-${CONFIG_BPF_LSM} += bpf_inode_storage.o obj-$(CONFIG_BPF_SYSCALL) += disasm.o obj-$(CONFIG_BPF_JIT) += trampoline.o diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c index c7a5be3bf8be..e9e7bd27ffad 100644 --- a/kernel/bpf/arraymap.c +++ b/kernel/bpf/arraymap.c @@ -13,6 +13,7 @@ #include <linux/rcupdate_trace.h> #include "map_in_map.h" +#include "map_trace.h" #define ARRAY_CREATE_FLAG_MASK \ (BPF_F_NUMA_NODE | BPF_F_MMAPABLE | BPF_F_ACCESS_MASK | \ @@ -329,7 +330,8 @@ static int array_map_update_elem(struct bpf_map *map, void *key, void *value, copy_map_value(map, val, value); check_and_free_timer_in_array(array, val); } - return 0; + + return bpf_map_trace_update_elem(map, key, value, map_flags); } int bpf_percpu_array_update(struct bpf_map *map, void *key, void *value, diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c index d29af9988f37..8fb19ed707e8 100644 --- a/kernel/bpf/hashtab.c +++ b/kernel/bpf/hashtab.c @@ -13,6 +13,7 @@ #include "percpu_freelist.h" #include "bpf_lru_list.h" #include "map_in_map.h" +#include "map_trace.h" #define HTAB_CREATE_FLAG_MASK \ (BPF_F_NO_PREALLOC | BPF_F_NO_COMMON_LRU | BPF_F_NUMA_NODE | \ @@ -1055,7 +1056,8 @@ static int htab_map_update_elem(struct bpf_map *map, void *key, void *value, copy_map_value_locked(map, l_old->key + round_up(key_size, 8), value, false); - return 0; + return bpf_map_trace_update_elem(map, key, value, + map_flags); } /* fall through, grab the bucket lock and lookup again. * 99.9% chance that the element won't be found, @@ -1109,6 +1111,8 @@ static int htab_map_update_elem(struct bpf_map *map, void *key, void *value, ret = 0; err: htab_unlock_bucket(htab, b, hash, flags); + if (!ret) + ret = bpf_map_trace_update_elem(map, key, value, map_flags); return ret; } @@ -1133,6 +1137,10 @@ static int htab_lru_map_update_elem(struct bpf_map *map, void *key, void *value, /* unknown flags */ return -EINVAL; + ret = bpf_map_trace_update_elem(map, key, value, map_flags); + if (unlikely(ret)) + return ret; + WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_trace_held() && !rcu_read_lock_bh_held()); @@ -1182,6 +1190,8 @@ static int htab_lru_map_update_elem(struct bpf_map *map, void *key, void *value, else if (l_old) htab_lru_push_free(htab, l_old); + if (!ret) + ret = bpf_map_trace_update_elem(map, key, value, map_flags); return ret; } @@ -1237,6 +1247,8 @@ static int __htab_percpu_map_update_elem(struct bpf_map *map, void *key, ret = 0; err: htab_unlock_bucket(htab, b, hash, flags); + if (!ret) + ret = bpf_map_trace_update_elem(map, key, value, map_flags); return ret; } @@ -1304,6 +1316,8 @@ static int __htab_lru_percpu_map_update_elem(struct bpf_map *map, void *key, htab_unlock_bucket(htab, b, hash, flags); if (l_new) bpf_lru_push_free(&htab->lru, &l_new->lru_node); + if (!ret) + ret = bpf_map_trace_update_elem(map, key, value, map_flags); return ret; } @@ -1354,6 +1368,8 @@ static int htab_map_delete_elem(struct bpf_map *map, void *key) } htab_unlock_bucket(htab, b, hash, flags); + if (!ret) + ret = bpf_map_trace_delete_elem(map, key); return ret; } @@ -1390,6 +1406,8 @@ static int htab_lru_map_delete_elem(struct bpf_map *map, void *key) htab_unlock_bucket(htab, b, hash, flags); if (l) htab_lru_push_free(htab, l); + if (!ret) + ret = bpf_map_trace_delete_elem(map, key); return ret; } diff --git a/kernel/bpf/map_trace.c b/kernel/bpf/map_trace.c new file mode 100644 index 000000000000..336848e83daf --- /dev/null +++ b/kernel/bpf/map_trace.c @@ -0,0 +1,17 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright (c) 2022 Google */ +#include "map_trace.h" + +noinline int bpf_map_trace_update_elem(struct bpf_map *map, void *key, + void *value, u64 map_flags) +{ + return 0; +} +ALLOW_ERROR_INJECTION(bpf_map_trace_update_elem, ERRNO); + +noinline int bpf_map_trace_delete_elem(struct bpf_map *map, void *key) +{ + return 0; +} +ALLOW_ERROR_INJECTION(bpf_map_trace_delete_elem, ERRNO); + diff --git a/kernel/bpf/map_trace.h b/kernel/bpf/map_trace.h new file mode 100644 index 000000000000..ae943af9e2a5 --- /dev/null +++ b/kernel/bpf/map_trace.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* Copyright (c) 2022 Google */ +#ifndef __BPF_MAP_TRACE_H_ +#define __BPF_MAP_TRACE_H_ + +#include <linux/bpf.h> + +/* + * Map tracing hooks. They are called from some, but not all, bpf map types. + * For those map types which call them, the only guarantee is that they are + * called after the corresponding action (bpf_map_update_elem, etc.) takes + * effect. + */ +int bpf_map_trace_update_elem(struct bpf_map *map, void *key, + void *value, u64 map_flags); + +int bpf_map_trace_delete_elem(struct bpf_map *map, void *key); + +#endif // __BPF_MAP_TRACE_H_ -- 2.34.1.448.ga2b2bfdf31-goog