4.14-stable review patch. If anyone has any objections, please let me know.
------------------
From: Alexei Starovoitov ast@kernel.org
commit b2157399cc9898260d6031c5bfe45fe137c1fbe7 upstream.
Under speculation, CPUs may mis-predict branches in bounds checks. Thus, memory accesses under a bounds check may be speculated even if the bounds check fails, providing a primitive for building a side channel.
To avoid leaking kernel data round up array-based maps and mask the index after bounds check, so speculated load with out of bounds index will load either valid value from the array or zero from the padded area.
Unconditionally mask index for all array types even when max_entries are not rounded to power of 2 for root user. When map is created by unpriv user generate a sequence of bpf insns that includes AND operation to make sure that JITed code includes the same 'index & index_mask' operation.
If prog_array map is created by unpriv user replace bpf_tail_call(ctx, map, index); with if (index >= max_entries) { index &= map->index_mask; bpf_tail_call(ctx, map, index); } (along with roundup to power 2) to prevent out-of-bounds speculation. There is secondary redundant 'if (index >= max_entries)' in the interpreter and in all JITs, but they can be optimized later if necessary.
Other array-like maps (cpumap, devmap, sockmap, perf_event_array, cgroup_array) cannot be used by unpriv, so no changes there.
That fixes bpf side of "Variant 1: bounds check bypass (CVE-2017-5753)" on all architectures with and without JIT.
v2->v3: Daniel noticed that attack potentially can be crafted via syscall commands without loading the program, so add masking to those paths as well.
Signed-off-by: Alexei Starovoitov ast@kernel.org Acked-by: John Fastabend john.fastabend@gmail.com Signed-off-by: Daniel Borkmann daniel@iogearbox.net Cc: Jiri Slaby jslaby@suse.cz Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
--- include/linux/bpf.h | 2 ++ kernel/bpf/arraymap.c | 47 ++++++++++++++++++++++++++++++++++++----------- kernel/bpf/verifier.c | 36 ++++++++++++++++++++++++++++++++++++ 3 files changed, 74 insertions(+), 11 deletions(-)
--- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -51,6 +51,7 @@ struct bpf_map { u32 pages; u32 id; int numa_node; + bool unpriv_array; struct user_struct *user; const struct bpf_map_ops *ops; struct work_struct work; @@ -195,6 +196,7 @@ struct bpf_prog_aux { struct bpf_array { struct bpf_map map; u32 elem_size; + u32 index_mask; /* 'ownership' of prog_array is claimed by the first program that * is going to use this map or by the first program which FD is stored * in the map to make sure that all callers and callees have the same --- a/kernel/bpf/arraymap.c +++ b/kernel/bpf/arraymap.c @@ -50,9 +50,10 @@ static struct bpf_map *array_map_alloc(u { bool percpu = attr->map_type == BPF_MAP_TYPE_PERCPU_ARRAY; int numa_node = bpf_map_attr_numa_node(attr); + u32 elem_size, index_mask, max_entries; + bool unpriv = !capable(CAP_SYS_ADMIN); struct bpf_array *array; u64 array_size; - u32 elem_size;
/* check sanity of attributes */ if (attr->max_entries == 0 || attr->key_size != 4 || @@ -68,11 +69,20 @@ static struct bpf_map *array_map_alloc(u
elem_size = round_up(attr->value_size, 8);
+ max_entries = attr->max_entries; + index_mask = roundup_pow_of_two(max_entries) - 1; + + if (unpriv) + /* round up array size to nearest power of 2, + * since cpu will speculate within index_mask limits + */ + max_entries = index_mask + 1; + array_size = sizeof(*array); if (percpu) - array_size += (u64) attr->max_entries * sizeof(void *); + array_size += (u64) max_entries * sizeof(void *); else - array_size += (u64) attr->max_entries * elem_size; + array_size += (u64) max_entries * elem_size;
/* make sure there is no u32 overflow later in round_up() */ if (array_size >= U32_MAX - PAGE_SIZE) @@ -82,6 +92,8 @@ static struct bpf_map *array_map_alloc(u array = bpf_map_area_alloc(array_size, numa_node); if (!array) return ERR_PTR(-ENOMEM); + array->index_mask = index_mask; + array->map.unpriv_array = unpriv;
/* copy mandatory map attributes */ array->map.map_type = attr->map_type; @@ -117,12 +129,13 @@ static void *array_map_lookup_elem(struc if (unlikely(index >= array->map.max_entries)) return NULL;
- return array->value + array->elem_size * index; + return array->value + array->elem_size * (index & array->index_mask); }
/* emit BPF instructions equivalent to C code of array_map_lookup_elem() */ static u32 array_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf) { + struct bpf_array *array = container_of(map, struct bpf_array, map); struct bpf_insn *insn = insn_buf; u32 elem_size = round_up(map->value_size, 8); const int ret = BPF_REG_0; @@ -131,7 +144,12 @@ static u32 array_map_gen_lookup(struct b
*insn++ = BPF_ALU64_IMM(BPF_ADD, map_ptr, offsetof(struct bpf_array, value)); *insn++ = BPF_LDX_MEM(BPF_W, ret, index, 0); - *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 3); + if (map->unpriv_array) { + *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 4); + *insn++ = BPF_ALU32_IMM(BPF_AND, ret, array->index_mask); + } else { + *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 3); + }
if (is_power_of_2(elem_size)) { *insn++ = BPF_ALU64_IMM(BPF_LSH, ret, ilog2(elem_size)); @@ -153,7 +171,7 @@ static void *percpu_array_map_lookup_ele if (unlikely(index >= array->map.max_entries)) return NULL;
- return this_cpu_ptr(array->pptrs[index]); + return this_cpu_ptr(array->pptrs[index & array->index_mask]); }
int bpf_percpu_array_copy(struct bpf_map *map, void *key, void *value) @@ -173,7 +191,7 @@ int bpf_percpu_array_copy(struct bpf_map */ size = round_up(map->value_size, 8); rcu_read_lock(); - pptr = array->pptrs[index]; + pptr = array->pptrs[index & array->index_mask]; for_each_possible_cpu(cpu) { bpf_long_memcpy(value + off, per_cpu_ptr(pptr, cpu), size); off += size; @@ -221,10 +239,11 @@ static int array_map_update_elem(struct return -EEXIST;
if (array->map.map_type == BPF_MAP_TYPE_PERCPU_ARRAY) - memcpy(this_cpu_ptr(array->pptrs[index]), + memcpy(this_cpu_ptr(array->pptrs[index & array->index_mask]), value, map->value_size); else - memcpy(array->value + array->elem_size * index, + memcpy(array->value + + array->elem_size * (index & array->index_mask), value, map->value_size); return 0; } @@ -258,7 +277,7 @@ int bpf_percpu_array_update(struct bpf_m */ size = round_up(map->value_size, 8); rcu_read_lock(); - pptr = array->pptrs[index]; + pptr = array->pptrs[index & array->index_mask]; for_each_possible_cpu(cpu) { bpf_long_memcpy(per_cpu_ptr(pptr, cpu), value + off, size); off += size; @@ -609,6 +628,7 @@ static void *array_of_map_lookup_elem(st static u32 array_of_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf) { + struct bpf_array *array = container_of(map, struct bpf_array, map); u32 elem_size = round_up(map->value_size, 8); struct bpf_insn *insn = insn_buf; const int ret = BPF_REG_0; @@ -617,7 +637,12 @@ static u32 array_of_map_gen_lookup(struc
*insn++ = BPF_ALU64_IMM(BPF_ADD, map_ptr, offsetof(struct bpf_array, value)); *insn++ = BPF_LDX_MEM(BPF_W, ret, index, 0); - *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 5); + if (map->unpriv_array) { + *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 6); + *insn++ = BPF_ALU32_IMM(BPF_AND, ret, array->index_mask); + } else { + *insn++ = BPF_JMP_IMM(BPF_JGE, ret, map->max_entries, 5); + } if (is_power_of_2(elem_size)) *insn++ = BPF_ALU64_IMM(BPF_LSH, ret, ilog2(elem_size)); else --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -1701,6 +1701,13 @@ static int check_call(struct bpf_verifie err = check_func_arg(env, BPF_REG_2, fn->arg2_type, &meta); if (err) return err; + if (func_id == BPF_FUNC_tail_call) { + if (meta.map_ptr == NULL) { + verbose("verifier bug\n"); + return -EINVAL; + } + env->insn_aux_data[insn_idx].map_ptr = meta.map_ptr; + } err = check_func_arg(env, BPF_REG_3, fn->arg3_type, &meta); if (err) return err; @@ -4315,6 +4322,35 @@ static int fixup_bpf_calls(struct bpf_ve */ insn->imm = 0; insn->code = BPF_JMP | BPF_TAIL_CALL; + + /* instead of changing every JIT dealing with tail_call + * emit two extra insns: + * if (index >= max_entries) goto out; + * index &= array->index_mask; + * to avoid out-of-bounds cpu speculation + */ + map_ptr = env->insn_aux_data[i + delta].map_ptr; + if (map_ptr == BPF_MAP_PTR_POISON) { + verbose("tail_call obusing map_ptr\n"); + return -EINVAL; + } + if (!map_ptr->unpriv_array) + continue; + insn_buf[0] = BPF_JMP_IMM(BPF_JGE, BPF_REG_3, + map_ptr->max_entries, 2); + insn_buf[1] = BPF_ALU32_IMM(BPF_AND, BPF_REG_3, + container_of(map_ptr, + struct bpf_array, + map)->index_mask); + insn_buf[2] = *insn; + cnt = 3; + new_prog = bpf_patch_insn_data(env, i + delta, insn_buf, cnt); + if (!new_prog) + return -ENOMEM; + + delta += cnt - 1; + env->prog = prog = new_prog; + insn = new_prog->insnsi + i + delta; continue; }