5 files changed, 409 insertions, 60 deletions

diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c
index 633c8c701ff6..58bdf5fd24cc 100644
--- a/kernel/bpf/arraymap.c
+++ b/kernel/bpf/arraymap.c
@@ -586,10 +586,17 @@ int bpf_fd_array_map_update_elem(struct bpf_map *map, struct file *map_file,
 	if (IS_ERR(new_ptr))
 		return PTR_ERR(new_ptr);
 
-	old_ptr = xchg(array->ptrs + index, new_ptr);
+	if (map->ops->map_poke_run) {
+		mutex_lock(&array->aux->poke_mutex);
+		old_ptr = xchg(array->ptrs + index, new_ptr);
+		map->ops->map_poke_run(map, index, old_ptr, new_ptr);
+		mutex_unlock(&array->aux->poke_mutex);
+	} else {
+		old_ptr = xchg(array->ptrs + index, new_ptr);
+	}
+
 	if (old_ptr)
 		map->ops->map_fd_put_ptr(old_ptr);
-
 	return 0;
 }
 
@@ -602,7 +609,15 @@ static int fd_array_map_delete_elem(struct bpf_map *map, void *key)
 	if (index >= array->map.max_entries)
 		return -E2BIG;
 
-	old_ptr = xchg(array->ptrs + index, NULL);
+	if (map->ops->map_poke_run) {
+		mutex_lock(&array->aux->poke_mutex);
+		old_ptr = xchg(array->ptrs + index, NULL);
+		map->ops->map_poke_run(map, index, old_ptr, NULL);
+		mutex_unlock(&array->aux->poke_mutex);
+	} else {
+		old_ptr = xchg(array->ptrs + index, NULL);
+	}
+
 	if (old_ptr) {
 		map->ops->map_fd_put_ptr(old_ptr);
 		return 0;
@@ -671,17 +686,205 @@ static void prog_array_map_seq_show_elem(struct bpf_map *map, void *key,
 	rcu_read_unlock();
 }
 
+struct prog_poke_elem {
+	struct list_head list;
+	struct bpf_prog_aux *aux;
+};
+
+static int prog_array_map_poke_track(struct bpf_map *map,
+				     struct bpf_prog_aux *prog_aux)
+{
+	struct prog_poke_elem *elem;
+	struct bpf_array_aux *aux;
+	int ret = 0;
+
+	aux = container_of(map, struct bpf_array, map)->aux;
+	mutex_lock(&aux->poke_mutex);
+	list_for_each_entry(elem, &aux->poke_progs, list) {
+		if (elem->aux == prog_aux)
+			goto out;
+	}
+
+	elem = kmalloc(sizeof(*elem), GFP_KERNEL);
+	if (!elem) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	INIT_LIST_HEAD(&elem->list);
+	/* We must track the program's aux info at this point in time
+	 * since the program pointer itself may not be stable yet, see
+	 * also comment in prog_array_map_poke_run().
+	 */
+	elem->aux = prog_aux;
+
+	list_add_tail(&elem->list, &aux->poke_progs);
+out:
+	mutex_unlock(&aux->poke_mutex);
+	return ret;
+}
+
+static void prog_array_map_poke_untrack(struct bpf_map *map,
+					struct bpf_prog_aux *prog_aux)
+{
+	struct prog_poke_elem *elem, *tmp;
+	struct bpf_array_aux *aux;
+
+	aux = container_of(map, struct bpf_array, map)->aux;
+	mutex_lock(&aux->poke_mutex);
+	list_for_each_entry_safe(elem, tmp, &aux->poke_progs, list) {
+		if (elem->aux == prog_aux) {
+			list_del_init(&elem->list);
+			kfree(elem);
+			break;
+		}
+	}
+	mutex_unlock(&aux->poke_mutex);
+}
+
+static void prog_array_map_poke_run(struct bpf_map *map, u32 key,
+				    struct bpf_prog *old,
+				    struct bpf_prog *new)
+{
+	enum bpf_text_poke_type type;
+	struct prog_poke_elem *elem;
+	struct bpf_array_aux *aux;
+
+	if (!old && new)
+		type = BPF_MOD_NOP_TO_JUMP;
+	else if (old && !new)
+		type = BPF_MOD_JUMP_TO_NOP;
+	else if (old && new)
+		type = BPF_MOD_JUMP_TO_JUMP;
+	else
+		return;
+
+	aux = container_of(map, struct bpf_array, map)->aux;
+	WARN_ON_ONCE(!mutex_is_locked(&aux->poke_mutex));
+
+	list_for_each_entry(elem, &aux->poke_progs, list) {
+		struct bpf_jit_poke_descriptor *poke;
+		int i, ret;
+
+		for (i = 0; i < elem->aux->size_poke_tab; i++) {
+			poke = &elem->aux->poke_tab[i];
+
+			/* Few things to be aware of:
+			 *
+			 * 1) We can only ever access aux in this context, but
+			 *    not aux->prog since it might not be stable yet and
+			 *    there could be danger of use after free otherwise.
+			 * 2) Initially when we start tracking aux, the program
+			 *    is not JITed yet and also does not have a kallsyms
+			 *    entry. We skip these as poke->ip_stable is not
+			 *    active yet. The JIT will do the final fixup before
+			 *    setting it stable. The various poke->ip_stable are
+			 *    successively activated, so tail call updates can
+			 *    arrive from here while JIT is still finishing its
+			 *    final fixup for non-activated poke entries.
+			 * 3) On program teardown, the program's kallsym entry gets
+			 *    removed out of RCU callback, but we can only untrack
+			 *    from sleepable context, therefore bpf_arch_text_poke()
+			 *    might not see that this is in BPF text section and
+			 *    bails out with -EINVAL. As these are unreachable since
+			 *    RCU grace period already passed, we simply skip them.
+			 * 4) Also programs reaching refcount of zero while patching
+			 *    is in progress is okay since we're protected under
+			 *    poke_mutex and untrack the programs before the JIT
+			 *    buffer is freed. When we're still in the middle of
+			 *    patching and suddenly kallsyms entry of the program
+			 *    gets evicted, we just skip the rest which is fine due
+			 *    to point 3).
+			 * 5) Any other error happening below from bpf_arch_text_poke()
+			 *    is a unexpected bug.
+			 */
+			if (!READ_ONCE(poke->ip_stable))
+				continue;
+			if (poke->reason != BPF_POKE_REASON_TAIL_CALL)
+				continue;
+			if (poke->tail_call.map != map ||
+			    poke->tail_call.key != key)
+				continue;
+
+			ret = bpf_arch_text_poke(poke->ip, type,
+						 old ? (u8 *)old->bpf_func +
+						 poke->adj_off : NULL,
+						 new ? (u8 *)new->bpf_func +
+						 poke->adj_off : NULL);
+			BUG_ON(ret < 0 && ret != -EINVAL);
+		}
+	}
+}
+
+static void prog_array_map_clear_deferred(struct work_struct *work)
+{
+	struct bpf_map *map = container_of(work, struct bpf_array_aux,
+					   work)->map;
+	bpf_fd_array_map_clear(map);
+	bpf_map_put(map);
+}
+
+static void prog_array_map_clear(struct bpf_map *map)
+{
+	struct bpf_array_aux *aux = container_of(map, struct bpf_array,
+						 map)->aux;
+	bpf_map_inc(map);
+	schedule_work(&aux->work);
+}
+
+static struct bpf_map *prog_array_map_alloc(union bpf_attr *attr)
+{
+	struct bpf_array_aux *aux;
+	struct bpf_map *map;
+
+	aux = kzalloc(sizeof(*aux), GFP_KERNEL);
+	if (!aux)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_WORK(&aux->work, prog_array_map_clear_deferred);
+	INIT_LIST_HEAD(&aux->poke_progs);
+	mutex_init(&aux->poke_mutex);
+
+	map = array_map_alloc(attr);
+	if (IS_ERR(map)) {
+		kfree(aux);
+		return map;
+	}
+
+	container_of(map, struct bpf_array, map)->aux = aux;
+	aux->map = map;
+
+	return map;
+}
+
+static void prog_array_map_free(struct bpf_map *map)
+{
+	struct prog_poke_elem *elem, *tmp;
+	struct bpf_array_aux *aux;
+
+	aux = container_of(map, struct bpf_array, map)->aux;
+	list_for_each_entry_safe(elem, tmp, &aux->poke_progs, list) {
+		list_del_init(&elem->list);
+		kfree(elem);
+	}
+	kfree(aux);
+	fd_array_map_free(map);
+}
+
 const struct bpf_map_ops prog_array_map_ops = {
 	.map_alloc_check = fd_array_map_alloc_check,
-	.map_alloc = array_map_alloc,
-	.map_free = fd_array_map_free,
+	.map_alloc = prog_array_map_alloc,
+	.map_free = prog_array_map_free,
+	.map_poke_track = prog_array_map_poke_track,
+	.map_poke_untrack = prog_array_map_poke_untrack,
+	.map_poke_run = prog_array_map_poke_run,
 	.map_get_next_key = array_map_get_next_key,
 	.map_lookup_elem = fd_array_map_lookup_elem,
 	.map_delete_elem = fd_array_map_delete_elem,
 	.map_fd_get_ptr = prog_fd_array_get_ptr,
 	.map_fd_put_ptr = prog_fd_array_put_ptr,
 	.map_fd_sys_lookup_elem = prog_fd_array_sys_lookup_elem,
-	.map_release_uref = bpf_fd_array_map_clear,
+	.map_release_uref = prog_array_map_clear,
 	.map_seq_show_elem = prog_array_map_seq_show_elem,
 };
 
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index b5945c3aaa8e..49e32acad7d8 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -256,6 +256,7 @@ void __bpf_prog_free(struct bpf_prog *fp)
 {
 	if (fp->aux) {
 		free_percpu(fp->aux->stats);
+		kfree(fp->aux->poke_tab);
 		kfree(fp->aux);
 	}
 	vfree(fp);
@@ -756,6 +757,39 @@ int bpf_get_kallsym(unsigned int symnum, unsigned long *value, char *type,
 	return ret;
 }
 
+int bpf_jit_add_poke_descriptor(struct bpf_prog *prog,
+				struct bpf_jit_poke_descriptor *poke)
+{
+	struct bpf_jit_poke_descriptor *tab = prog->aux->poke_tab;
+	static const u32 poke_tab_max = 1024;
+	u32 slot = prog->aux->size_poke_tab;
+	u32 size = slot + 1;
+
+	if (size > poke_tab_max)
+		return -ENOSPC;
+	if (poke->ip || poke->ip_stable || poke->adj_off)
+		return -EINVAL;
+
+	switch (poke->reason) {
+	case BPF_POKE_REASON_TAIL_CALL:
+		if (!poke->tail_call.map)
+			return -EINVAL;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	tab = krealloc(tab, size * sizeof(*poke), GFP_KERNEL);
+	if (!tab)
+		return -ENOMEM;
+
+	memcpy(&tab[slot], poke, sizeof(*poke));
+	prog->aux->size_poke_tab = size;
+	prog->aux->poke_tab = tab;
+
+	return slot;
+}
+
 static atomic_long_t bpf_jit_current;
 
 /* Can be overridden by an arch's JIT compiler if it has a custom,
@@ -1691,18 +1725,17 @@ bool bpf_prog_array_compatible(struct bpf_array *array,
 	if (fp->kprobe_override)
 		return false;
 
-	if (!array->owner_prog_type) {
+	if (!array->aux->type) {
 		/* There's no owner yet where we could check for
 		 * compatibility.
 		 */
-		array->owner_prog_type = fp->type;
-		array->owner_jited = fp->jited;
-
+		array->aux->type  = fp->type;
+		array->aux->jited = fp->jited;
 		return true;
 	}
 
-	return array->owner_prog_type == fp->type &&
-	       array->owner_jited == fp->jited;
+	return array->aux->type  == fp->type &&
+	       array->aux->jited == fp->jited;
 }
 
 static int bpf_check_tail_call(const struct bpf_prog *fp)
@@ -2003,12 +2036,40 @@ int bpf_prog_array_copy_info(struct bpf_prog_array *array,
 								     : 0;
 }
 
+static void bpf_free_cgroup_storage(struct bpf_prog_aux *aux)
+{
+	enum bpf_cgroup_storage_type stype;
+
+	for_each_cgroup_storage_type(stype) {
+		if (!aux->cgroup_storage[stype])
+			continue;
+		bpf_cgroup_storage_release(aux->prog,
+					   aux->cgroup_storage[stype]);
+	}
+}
+
+static void bpf_free_used_maps(struct bpf_prog_aux *aux)
+{
+	struct bpf_map *map;
+	int i;
+
+	bpf_free_cgroup_storage(aux);
+	for (i = 0; i < aux->used_map_cnt; i++) {
+		map = aux->used_maps[i];
+		if (map->ops->map_poke_untrack)
+			map->ops->map_poke_untrack(map, aux);
+		bpf_map_put(map);
+	}
+	kfree(aux->used_maps);
+}
+
 static void bpf_prog_free_deferred(struct work_struct *work)
 {
 	struct bpf_prog_aux *aux;
 	int i;
 
 	aux = container_of(work, struct bpf_prog_aux, work);
+	bpf_free_used_maps(aux);
 	if (bpf_prog_is_dev_bound(aux))
 		bpf_prog_offload_destroy(aux->prog);
 #ifdef CONFIG_PERF_EVENTS
diff --git a/kernel/bpf/map_in_map.c b/kernel/bpf/map_in_map.c
index 4cbe987be35b..5e9366b33f0f 100644
--- a/kernel/bpf/map_in_map.c
+++ b/kernel/bpf/map_in_map.c
@@ -17,9 +17,8 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd)
 	if (IS_ERR(inner_map))
 		return inner_map;
 
-	/* prog_array->owner_prog_type and owner_jited
-	 * is a runtime binding.  Doing static check alone
-	 * in the verifier is not enough.
+	/* prog_array->aux->{type,jited} is a runtime binding.
+	 * Doing static check alone in the verifier is not enough.
 	 */
 	if (inner_map->map_type == BPF_MAP_TYPE_PROG_ARRAY ||
 	    inner_map->map_type == BPF_MAP_TYPE_CGROUP_STORAGE ||
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 4ae52eb05f41..e3461ec59570 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -25,12 +25,13 @@
 #include <linux/nospec.h>
 #include <uapi/linux/btf.h>
 
-#define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PROG_ARRAY || \
-			   (map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
-			   (map)->map_type == BPF_MAP_TYPE_CGROUP_ARRAY || \
-			   (map)->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS)
+#define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
+			  (map)->map_type == BPF_MAP_TYPE_CGROUP_ARRAY || \
+			  (map)->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS)
+#define IS_FD_PROG_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PROG_ARRAY)
 #define IS_FD_HASH(map) ((map)->map_type == BPF_MAP_TYPE_HASH_OF_MAPS)
-#define IS_FD_MAP(map) (IS_FD_ARRAY(map) || IS_FD_HASH(map))
+#define IS_FD_MAP(map) (IS_FD_ARRAY(map) || IS_FD_PROG_ARRAY(map) || \
+			IS_FD_HASH(map))
 
 #define BPF_OBJ_FLAG_MASK   (BPF_F_RDONLY | BPF_F_WRONLY)
 
@@ -389,13 +390,12 @@ static void bpf_map_show_fdinfo(struct seq_file *m, struct file *filp)
 {
 	const struct bpf_map *map = filp->private_data;
 	const struct bpf_array *array;
-	u32 owner_prog_type = 0;
-	u32 owner_jited = 0;
+	u32 type = 0, jited = 0;
 
 	if (map->map_type == BPF_MAP_TYPE_PROG_ARRAY) {
 		array = container_of(map, struct bpf_array, map);
-		owner_prog_type = array->owner_prog_type;
-		owner_jited = array->owner_jited;
+		type  = array->aux->type;
+		jited = array->aux->jited;
 	}
 
 	seq_printf(m,
@@ -415,12 +415,9 @@ static void bpf_map_show_fdinfo(struct seq_file *m, struct file *filp)
 		   map->memory.pages * 1ULL << PAGE_SHIFT,
 		   map->id,
 		   READ_ONCE(map->frozen));
-
-	if (owner_prog_type) {
-		seq_printf(m, "owner_prog_type:\t%u\n",
-			   owner_prog_type);
-		seq_printf(m, "owner_jited:\t%u\n",
-			   owner_jited);
+	if (type) {
+		seq_printf(m, "owner_prog_type:\t%u\n", type);
+		seq_printf(m, "owner_jited:\t%u\n", jited);
 	}
 }
 #endif
@@ -881,7 +878,7 @@ static int map_lookup_elem(union bpf_attr *attr)
 		err = bpf_percpu_cgroup_storage_copy(map, key, value);
 	} else if (map->map_type == BPF_MAP_TYPE_STACK_TRACE) {
 		err = bpf_stackmap_copy(map, key, value);
-	} else if (IS_FD_ARRAY(map)) {
+	} else if (IS_FD_ARRAY(map) || IS_FD_PROG_ARRAY(map)) {
 		err = bpf_fd_array_map_lookup_elem(map, key, value);
 	} else if (IS_FD_HASH(map)) {
 		err = bpf_fd_htab_map_lookup_elem(map, key, value);
@@ -1008,6 +1005,10 @@ static int map_update_elem(union bpf_attr *attr)
 		   map->map_type == BPF_MAP_TYPE_SOCKMAP) {
 		err = map->ops->map_update_elem(map, key, value, attr->flags);
 		goto out;
+	} else if (IS_FD_PROG_ARRAY(map)) {
+		err = bpf_fd_array_map_update_elem(map, f.file, key, value,
+						   attr->flags);
+		goto out;
 	}
 
 	/* must increment bpf_prog_active to avoid kprobe+bpf triggering from
@@ -1090,6 +1091,9 @@ static int map_delete_elem(union bpf_attr *attr)
 	if (bpf_map_is_dev_bound(map)) {
 		err = bpf_map_offload_delete_elem(map, key);
 		goto out;
+	} else if (IS_FD_PROG_ARRAY(map)) {
+		err = map->ops->map_delete_elem(map, key);
+		goto out;
 	}
 
 	preempt_disable();
@@ -1302,25 +1306,6 @@ static int find_prog_type(enum bpf_prog_type type, struct bpf_prog *prog)
 	return 0;
 }
 
-/* drop refcnt on maps used by eBPF program and free auxilary data */
-static void free_used_maps(struct bpf_prog_aux *aux)
-{
-	enum bpf_cgroup_storage_type stype;
-	int i;
-
-	for_each_cgroup_storage_type(stype) {
-		if (!aux->cgroup_storage[stype])
-			continue;
-		bpf_cgroup_storage_release(aux->prog,
-					   aux->cgroup_storage[stype]);
-	}
-
-	for (i = 0; i < aux->used_map_cnt; i++)
-		bpf_map_put(aux->used_maps[i]);
-
-	kfree(aux->used_maps);
-}
-
 int __bpf_prog_charge(struct user_struct *user, u32 pages)
 {
 	unsigned long memlock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
@@ -1415,7 +1400,6 @@ static void __bpf_prog_put_rcu(struct rcu_head *rcu)
 
 	kvfree(aux->func_info);
 	kfree(aux->func_info_aux);
-	free_used_maps(aux);
 	bpf_prog_uncharge_memlock(aux->prog);
 	security_bpf_prog_free(aux);
 	bpf_prog_free(aux->prog);
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index fc85714428c7..a0482e1c4a77 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -171,6 +171,9 @@ struct bpf_verifier_stack_elem {
 #define BPF_COMPLEXITY_LIMIT_JMP_SEQ	8192
 #define BPF_COMPLEXITY_LIMIT_STATES	64
 
+#define BPF_MAP_KEY_POISON	(1ULL << 63)
+#define BPF_MAP_KEY_SEEN	(1ULL << 62)
+
 #define BPF_MAP_PTR_UNPRIV	1UL
 #define BPF_MAP_PTR_POISON	((void *)((0xeB9FUL << 1) +	\
 					  POISON_POINTER_DELTA))
@@ -178,12 +181,12 @@ struct bpf_verifier_stack_elem {
 
 static bool bpf_map_ptr_poisoned(const struct bpf_insn_aux_data *aux)
 {
-	return BPF_MAP_PTR(aux->map_state) == BPF_MAP_PTR_POISON;
+	return BPF_MAP_PTR(aux->map_ptr_state) == BPF_MAP_PTR_POISON;
 }
 
 static bool bpf_map_ptr_unpriv(const struct bpf_insn_aux_data *aux)
 {
-	return aux->map_state & BPF_MAP_PTR_UNPRIV;
+	return aux->map_ptr_state & BPF_MAP_PTR_UNPRIV;
 }
 
 static void bpf_map_ptr_store(struct bpf_insn_aux_data *aux,
@@ -191,8 +194,31 @@ static void bpf_map_ptr_store(struct bpf_insn_aux_data *aux,
 {
 	BUILD_BUG_ON((unsigned long)BPF_MAP_PTR_POISON & BPF_MAP_PTR_UNPRIV);
 	unpriv |= bpf_map_ptr_unpriv(aux);
-	aux->map_state = (unsigned long)map |
-			 (unpriv ? BPF_MAP_PTR_UNPRIV : 0UL);
+	aux->map_ptr_state = (unsigned long)map |
+			     (unpriv ? BPF_MAP_PTR_UNPRIV : 0UL);
+}
+
+static bool bpf_map_key_poisoned(const struct bpf_insn_aux_data *aux)
+{
+	return aux->map_key_state & BPF_MAP_KEY_POISON;
+}
+
+static bool bpf_map_key_unseen(const struct bpf_insn_aux_data *aux)
+{
+	return !(aux->map_key_state & BPF_MAP_KEY_SEEN);
+}
+
+static u64 bpf_map_key_immediate(const struct bpf_insn_aux_data *aux)
+{
+	return aux->map_key_state & ~(BPF_MAP_KEY_SEEN | BPF_MAP_KEY_POISON);
+}
+
+static void bpf_map_key_store(struct bpf_insn_aux_data *aux, u64 state)
+{
+	bool poisoned = bpf_map_key_poisoned(aux);
+
+	aux->map_key_state = state | BPF_MAP_KEY_SEEN |
+			     (poisoned ? BPF_MAP_KEY_POISON : 0ULL);
 }
 
 struct bpf_call_arg_meta {
@@ -4090,15 +4116,49 @@ record_func_map(struct bpf_verifier_env *env, struct bpf_call_arg_meta *meta,
 		return -EACCES;
 	}
 
-	if (!BPF_MAP_PTR(aux->map_state))
+	if (!BPF_MAP_PTR(aux->map_ptr_state))
 		bpf_map_ptr_store(aux, meta->map_ptr,
 				  meta->map_ptr->unpriv_array);
-	else if (BPF_MAP_PTR(aux->map_state) != meta->map_ptr)
+	else if (BPF_MAP_PTR(aux->map_ptr_state) != meta->map_ptr)
 		bpf_map_ptr_store(aux, BPF_MAP_PTR_POISON,
 				  meta->map_ptr->unpriv_array);
 	return 0;
 }
 
+static int
+record_func_key(struct bpf_verifier_env *env, struct bpf_call_arg_meta *meta,
+		int func_id, int insn_idx)
+{
+	struct bpf_insn_aux_data *aux = &env->insn_aux_data[insn_idx];
+	struct bpf_reg_state *regs = cur_regs(env), *reg;
+	struct bpf_map *map = meta->map_ptr;
+	struct tnum range;
+	u64 val;
+
+	if (func_id != BPF_FUNC_tail_call)
+		return 0;
+	if (!map || map->map_type != BPF_MAP_TYPE_PROG_ARRAY) {
+		verbose(env, "kernel subsystem misconfigured verifier\n");
+		return -EINVAL;
+	}
+
+	range = tnum_range(0, map->max_entries - 1);
+	reg = &regs[BPF_REG_3];
+
+	if (!register_is_const(reg) || !tnum_in(range, reg->var_off)) {
+		bpf_map_key_store(aux, BPF_MAP_KEY_POISON);
+		return 0;
+	}
+
+	val = reg->var_off.value;
+	if (bpf_map_key_unseen(aux))
+		bpf_map_key_store(aux, val);
+	else if (!bpf_map_key_poisoned(aux) &&
+		  bpf_map_key_immediate(aux) != val)
+		bpf_map_key_store(aux, BPF_MAP_KEY_POISON);
+	return 0;
+}
+
 static int check_reference_leak(struct bpf_verifier_env *env)
 {
 	struct bpf_func_state *state = cur_func(env);
@@ -4173,6 +4233,10 @@ static int check_helper_call(struct bpf_verifier_env *env, int func_id, int insn
 	if (err)
 		return err;
 
+	err = record_func_key(env, &meta, func_id, insn_idx);
+	if (err)
+		return err;
+
 	/* Mark slots with STACK_MISC in case of raw mode, stack offset
 	 * is inferred from register state.
 	 */
@@ -9065,6 +9129,7 @@ static int fixup_call_args(struct bpf_verifier_env *env)
 static int fixup_bpf_calls(struct bpf_verifier_env *env)
 {
 	struct bpf_prog *prog = env->prog;
+	bool expect_blinding = bpf_jit_blinding_enabled(prog);
 	struct bpf_insn *insn = prog->insnsi;
 	const struct bpf_func_proto *fn;
 	const int insn_cnt = prog->len;
@@ -9073,7 +9138,7 @@ static int fixup_bpf_calls(struct bpf_verifier_env *env)
 	struct bpf_insn insn_buf[16];
 	struct bpf_prog *new_prog;
 	struct bpf_map *map_ptr;
-	int i, cnt, delta = 0;
+	int i, ret, cnt, delta = 0;
 
 	for (i = 0; i < insn_cnt; i++, insn++) {
 		if (insn->code == (BPF_ALU64 | BPF_MOD | BPF_X) ||
@@ -9217,6 +9282,26 @@ static int fixup_bpf_calls(struct bpf_verifier_env *env)
 			insn->code = BPF_JMP | BPF_TAIL_CALL;
 
 			aux = &env->insn_aux_data[i + delta];
+			if (prog->jit_requested && !expect_blinding &&
+			    !bpf_map_key_poisoned(aux) &&
+			    !bpf_map_ptr_poisoned(aux) &&
+			    !bpf_map_ptr_unpriv(aux)) {
+				struct bpf_jit_poke_descriptor desc = {
+					.reason = BPF_POKE_REASON_TAIL_CALL,
+					.tail_call.map = BPF_MAP_PTR(aux->map_ptr_state),
+					.tail_call.key = bpf_map_key_immediate(aux),
+				};
+
+				ret = bpf_jit_add_poke_descriptor(prog, &desc);
+				if (ret < 0) {
+					verbose(env, "adding tail call poke descriptor failed\n");
+					return ret;
+				}
+
+				insn->imm = ret + 1;
+				continue;
+			}
+
 			if (!bpf_map_ptr_unpriv(aux))
 				continue;
 
@@ -9231,7 +9316,7 @@ static int fixup_bpf_calls(struct bpf_verifier_env *env)
 				return -EINVAL;
 			}
 
-			map_ptr = BPF_MAP_PTR(aux->map_state);
+			map_ptr = BPF_MAP_PTR(aux->map_ptr_state);
 			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,
@@ -9265,7 +9350,7 @@ static int fixup_bpf_calls(struct bpf_verifier_env *env)
 			if (bpf_map_ptr_poisoned(aux))
 				goto patch_call_imm;
 
-			map_ptr = BPF_MAP_PTR(aux->map_state);
+			map_ptr = BPF_MAP_PTR(aux->map_ptr_state);
 			ops = map_ptr->ops;
 			if (insn->imm == BPF_FUNC_map_lookup_elem &&
 			    ops->map_gen_lookup) {
@@ -9345,6 +9430,23 @@ patch_call_imm:
 		insn->imm = fn->func - __bpf_call_base;
 	}
 
+	/* Since poke tab is now finalized, publish aux to tracker. */
+	for (i = 0; i < prog->aux->size_poke_tab; i++) {
+		map_ptr = prog->aux->poke_tab[i].tail_call.map;
+		if (!map_ptr->ops->map_poke_track ||
+		    !map_ptr->ops->map_poke_untrack ||
+		    !map_ptr->ops->map_poke_run) {
+			verbose(env, "bpf verifier is misconfigured\n");
+			return -EINVAL;
+		}
+
+		ret = map_ptr->ops->map_poke_track(map_ptr, prog->aux);
+		if (ret < 0) {
+			verbose(env, "tracking tail call prog failed\n");
+			return ret;
+		}
+	}
+
 	return 0;
 }