[PATCH 4.9 53/96] bpf: prevent out-of-bounds speculation

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

4.9-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>
[ Backported to 4.9 - gregkh ]
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 include/linux/bpf.h          |    2 ++
 include/linux/bpf_verifier.h |    5 ++++-
 kernel/bpf/arraymap.c        |   31 ++++++++++++++++++++++---------
 kernel/bpf/verifier.c        |   42 +++++++++++++++++++++++++++++++++++++++---
 4 files changed, 67 insertions(+), 13 deletions(-)

--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -43,6 +43,7 @@ struct bpf_map {
 	u32 max_entries;
 	u32 map_flags;
 	u32 pages;
+	bool unpriv_array;
 	struct user_struct *user;
 	const struct bpf_map_ops *ops;
 	struct work_struct work;
@@ -189,6 +190,7 @@ 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/include/linux/bpf_verifier.h
+++ b/include/linux/bpf_verifier.h
@@ -67,7 +67,10 @@ struct bpf_verifier_state_list {
 };
 
 struct bpf_insn_aux_data {
-	enum bpf_reg_type ptr_type;	/* pointer type for load/store insns */
+	union {
+		enum bpf_reg_type ptr_type;     /* pointer type for load/store insns */
+		struct bpf_map *map_ptr;        /* pointer for call insn into lookup_elem */
+	};
 	bool seen; /* this insn was processed by the verifier */
 };
 
--- a/kernel/bpf/arraymap.c
+++ b/kernel/bpf/arraymap.c
@@ -46,9 +46,10 @@ static int bpf_array_alloc_percpu(struct
 static struct bpf_map *array_map_alloc(union bpf_attr *attr)
 {
 	bool percpu = attr->map_type == BPF_MAP_TYPE_PERCPU_ARRAY;
+	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 ||
@@ -63,11 +64,20 @@ 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)
@@ -77,6 +87,8 @@ static struct bpf_map *array_map_alloc(u
 	array = bpf_map_area_alloc(array_size);
 	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;
@@ -110,7 +122,7 @@ 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);
 }
 
 /* Called from eBPF program */
@@ -122,7 +134,7 @@ 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)
@@ -142,7 +154,7 @@ 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;
@@ -190,10 +202,11 @@ 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;
 }
@@ -227,7 +240,7 @@ 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;
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -1187,7 +1187,7 @@ static void clear_all_pkt_pointers(struc
 	}
 }
 
-static int check_call(struct bpf_verifier_env *env, int func_id)
+static int check_call(struct bpf_verifier_env *env, int func_id, int insn_idx)
 {
 	struct bpf_verifier_state *state = &env->cur_state;
 	const struct bpf_func_proto *fn = NULL;
@@ -1238,6 +1238,13 @@ 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;
@@ -3019,7 +3026,7 @@ static int do_check(struct bpf_verifier_
 					return -EINVAL;
 				}
 
-				err = check_call(env, insn->imm);
+				err = check_call(env, insn->imm, insn_idx);
 				if (err)
 					return err;
 
@@ -3372,7 +3379,11 @@ static int fixup_bpf_calls(struct bpf_ve
 	struct bpf_insn *insn = prog->insnsi;
 	const struct bpf_func_proto *fn;
 	const int insn_cnt = prog->len;
-	int i;
+	struct bpf_insn insn_buf[16];
+	struct bpf_prog *new_prog;
+	struct bpf_map *map_ptr;
+	int i, cnt, delta = 0;
+
 
 	for (i = 0; i < insn_cnt; i++, insn++) {
 		if (insn->code != (BPF_JMP | BPF_CALL))
@@ -3390,6 +3401,31 @@ static int fixup_bpf_calls(struct bpf_ve
  			 */
 			insn->imm = 0;
 			insn->code |= BPF_X;
+
+			/* 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->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;
 		}