]> git.saurik.com Git - apple/xnu.git/blobdiff - bsd/net/necp.c
xnu-6153.61.1.tar.gz
[apple/xnu.git] / bsd / net / necp.c
index 513cea4aead8943e76fc98b846ba65f1ced59095..2133f31f3601376d03be95e0c2f2eedc550860c0 100644 (file)
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013-2018 Apple Inc. All rights reserved.
+ * Copyright (c) 2013-2019 Apple Inc. All rights reserved.
  *
  * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
  *
@@ -43,6 +43,9 @@
 #include <sys/socket.h>
 #include <sys/socketvar.h>
 #include <sys/coalition.h>
+#include <sys/ubc.h>
+#include <sys/codesign.h>
+#include <kern/cs_blobs.h>
 #include <netinet/ip.h>
 #include <netinet/ip6.h>
 #include <netinet/tcp.h>
 #include <sys/kern_event.h>
 #include <sys/file_internal.h>
 #include <IOKit/IOBSD.h>
+#include <libkern/crypto/rand.h>
+#include <corecrypto/cchmac.h>
+#include <corecrypto/ccsha2.h>
+#include <os/refcnt.h>
 #include <net/network_agent.h>
 #include <net/necp.h>
 
@@ -68,7 +75,7 @@
  * NECP - Network Extension Control Policy database
  * ------------------------------------------------
  * The goal of this module is to allow clients connecting via a
- * kernel control socket to create high-level policy sessions, which
+ * policy file descriptor to create high-level policy sessions, which
  * are ingested into low-level kernel policies that control and tag
  * traffic at the application, socket, and IP layers.
  *
@@ -84,7 +91,7 @@
  * can specify the sub-order for each policy it creates which will be
  * used to further sort the kernel policies.
  *
- *  Kernel Control Socket --> 1 necp_session --> list of necp_session_policy structs
+ *  Policy fd --> 1 necp_session --> list of necp_session_policy structs
  *
  * ------------------------------------------------
  * Kernel Policies
@@ -136,6 +143,14 @@ u_int32_t necp_drop_all_level = 0;
 
 u_int32_t necp_pass_loopback = 1; // 0=Off, 1=On
 u_int32_t necp_pass_keepalives = 1; // 0=Off, 1=On
+u_int32_t necp_pass_interpose = 1; // 0=Off, 1=On
+
+u_int32_t necp_drop_unentitled_order = 0;
+#ifdef XNU_TARGET_OS_WATCH
+u_int32_t necp_drop_unentitled_level = NECP_SESSION_PRIORITY_CONTROL + 1; // Block all unentitled traffic from policies below control level
+#else // XNU_TARGET_OS_WATCH
+u_int32_t necp_drop_unentitled_level = 0;
+#endif // XNU_TARGET_OS_WATCH
 
 u_int32_t necp_debug = 0; // 0=None, 1=Basic, 2=EveryMatch
 
@@ -182,6 +197,9 @@ u_int32_t necp_session_count = 0;
 
 #define IS_NECP_ROUTE_RULE_ALLOW_OR_DENY(x)     ((x) == NECP_ROUTE_RULE_DENY_INTERFACE || (x) == NECP_ROUTE_RULE_ALLOW_INTERFACE)
 
+#define IS_NECP_DEST_IN_LOCAL_NETWORKS(rt) \
+    ((rt) != NULL && !((rt)->rt_flags & RTF_GATEWAY) && ((rt)->rt_ifa && (rt)->rt_ifa->ifa_ifp && !((rt)->rt_ifa->ifa_ifp->if_flags & IFF_POINTOPOINT)))
+
 #define NECP_KERNEL_CONDITION_ALL_INTERFACES            0x000001
 #define NECP_KERNEL_CONDITION_BOUND_INTERFACE           0x000002
 #define NECP_KERNEL_CONDITION_PROTOCOL                          0x000004
@@ -203,6 +221,12 @@ u_int32_t necp_session_count = 0;
 #define NECP_KERNEL_CONDITION_ENTITLEMENT                       0x040000
 #define NECP_KERNEL_CONDITION_CUSTOM_ENTITLEMENT        0x080000
 #define NECP_KERNEL_CONDITION_AGENT_TYPE                        0x100000
+#define NECP_KERNEL_CONDITION_HAS_CLIENT                        0x200000
+#define NECP_KERNEL_CONDITION_LOCAL_NETWORKS                    0x400000
+#define NECP_KERNEL_CONDITION_CLIENT_FLAGS                      0x800000
+#define NECP_KERNEL_CONDITION_LOCAL_EMPTY                       0x1000000
+#define NECP_KERNEL_CONDITION_REMOTE_EMPTY                      0x2000000
+#define NECP_KERNEL_CONDITION_PLATFORM_BINARY                   0x4000000
 
 #define NECP_MAX_POLICY_RESULT_SIZE                                     512
 #define NECP_MAX_ROUTE_RULES_ARRAY_SIZE                         1024
@@ -256,13 +280,15 @@ struct necp_socket_info {
        u_int32_t application_id;
        u_int32_t real_application_id;
        u_int32_t account_id;
+       u_int32_t drop_order;
+       u_int32_t client_flags;
        char *domain;
        errno_t cred_result;
+       unsigned has_client : 1;
+       unsigned is_platform_binary : 1;
+       unsigned __pad_bits : 6;
 };
 
-static kern_ctl_ref     necp_kctlref;
-static u_int32_t        necp_family;
-static OSMallocTag      necp_malloc_tag;
 static  lck_grp_attr_t  *necp_kernel_policy_grp_attr    = NULL;
 static  lck_attr_t              *necp_kernel_policy_mtx_attr    = NULL;
 static  lck_grp_t               *necp_kernel_policy_mtx_grp             = NULL;
@@ -273,6 +299,8 @@ static  lck_attr_t              *necp_route_rule_mtx_attr       = NULL;
 static  lck_grp_t               *necp_route_rule_mtx_grp        = NULL;
 decl_lck_rw_data(static, necp_route_rule_lock);
 
+os_refgrp_decl(static, necp_refgrp, "NECPRefGroup", NULL);
+
 /*
  * On modification, invalidate cached lookups by bumping the generation count.
  * Other calls will need to take the slowpath of taking
@@ -285,6 +313,22 @@ static volatile int32_t necp_kernel_socket_policies_gencount;
        }                                                                                                                                                               \
 } while (0)
 
+/*
+ * Drop-all Bypass:
+ * Allow priviledged processes to bypass the default drop-all
+ * via entitlement check.  For OSX, since entitlement check is
+ * not supported for configd, configd signing identity is checked
+ * instead.
+ */
+#define SIGNING_ID_CONFIGD "com.apple.configd"
+#define SIGNING_ID_CONFIGD_LEN (sizeof(SIGNING_ID_CONFIGD) - 1)
+
+typedef enum {
+       NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE = 0,
+       NECP_DROP_ALL_BYPASS_CHECK_RESULT_TRUE = 1,
+       NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE = 2,
+} necp_drop_all_bypass_check_result_t;
+
 static u_int32_t necp_kernel_application_policies_condition_mask;
 static size_t necp_kernel_application_policies_count;
 static u_int32_t necp_kernel_socket_policies_condition_mask;
@@ -310,23 +354,18 @@ static LIST_HEAD(_necpkernelipoutputpolicies, necp_kernel_ip_output_policy) necp
 #define NECP_KERNEL_IP_OUTPUT_POLICIES_MAP_NUM_ID_BUCKETS 5
 #define NECP_IP_OUTPUT_MAP_ID_TO_BUCKET(id) (id ? (id%(NECP_KERNEL_IP_OUTPUT_POLICIES_MAP_NUM_ID_BUCKETS - 1) + 1) : 0)
 static struct necp_kernel_ip_output_policy **necp_kernel_ip_output_policies_map[NECP_KERNEL_IP_OUTPUT_POLICIES_MAP_NUM_ID_BUCKETS];
+static struct necp_kernel_socket_policy pass_policy =
+{
+       .id = NECP_KERNEL_POLICY_ID_NO_MATCH,
+       .result = NECP_KERNEL_POLICY_RESULT_PASS,
+};
 
 static struct necp_session *necp_create_session(void);
 static void necp_delete_session(struct necp_session *session);
 
-static necp_policy_id necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_t packet,
+static necp_policy_id necp_handle_policy_add(struct necp_session *session,
     u_int8_t *tlv_buffer, size_t tlv_buffer_length, int offset, int *error);
-static void necp_handle_policy_get(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_policy_delete(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_policy_apply_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_policy_list_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_policy_delete_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static int necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet,
-    user_addr_t out_buffer, size_t out_buffer_length, int offset);
-static void necp_handle_set_session_priority(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_lock_session_to_proc(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_register_service(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
-static void necp_handle_unregister_service(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset);
+static int necp_handle_policy_dump_all(user_addr_t out_buffer, size_t out_buffer_length);
 
 #define MAX_RESULT_STRING_LEN 64
 static inline const char * necp_get_result_description(char *result_string, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter);
@@ -338,11 +377,11 @@ static bool necp_policy_mark_all_for_deletion(struct necp_session *session);
 static bool necp_policy_delete(struct necp_session *session, struct necp_session_policy *policy);
 static void necp_policy_apply_all(struct necp_session *session);
 
-static necp_kernel_policy_id necp_kernel_socket_policy_add(necp_policy_order order, u_int32_t session_order, int session_pid, u_int32_t condition_mask, u_int32_t condition_negated_mask, necp_app_id cond_app_id, necp_app_id cond_real_app_id, char *cond_custom_entitlement, u_int32_t cond_account_id, char *domain, pid_t cond_pid, uid_t cond_uid, ifnet_t cond_bound_interface, struct necp_policy_condition_tc_range cond_traffic_class, u_int16_t cond_protocol, union necp_sockaddr_union *cond_local_start, union necp_sockaddr_union *cond_local_end, u_int8_t cond_local_prefix, union necp_sockaddr_union *cond_remote_start, union necp_sockaddr_union *cond_remote_end, u_int8_t cond_remote_prefix, struct necp_policy_condition_agent_type *cond_agent_type, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter);
+static necp_kernel_policy_id necp_kernel_socket_policy_add(necp_policy_order order, u_int32_t session_order, int session_pid, u_int32_t condition_mask, u_int32_t condition_negated_mask, necp_app_id cond_app_id, necp_app_id cond_real_app_id, char *cond_custom_entitlement, u_int32_t cond_account_id, char *domain, pid_t cond_pid, uid_t cond_uid, ifnet_t cond_bound_interface, struct necp_policy_condition_tc_range cond_traffic_class, u_int16_t cond_protocol, union necp_sockaddr_union *cond_local_start, union necp_sockaddr_union *cond_local_end, u_int8_t cond_local_prefix, union necp_sockaddr_union *cond_remote_start, union necp_sockaddr_union *cond_remote_end, u_int8_t cond_remote_prefix, struct necp_policy_condition_agent_type *cond_agent_type, u_int32_t cond_client_flags, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter);
 static bool necp_kernel_socket_policy_delete(necp_kernel_policy_id policy_id);
 static bool necp_kernel_socket_policies_reprocess(void);
 static bool necp_kernel_socket_policies_update_uuid_table(void);
-static inline struct necp_kernel_socket_policy *necp_socket_find_policy_match_with_info_locked(struct necp_kernel_socket_policy **policy_search_array, struct necp_socket_info *info, necp_kernel_policy_filter *return_filter, u_int32_t *return_route_rule_id, necp_kernel_policy_result *return_service_action, necp_kernel_policy_service *return_service, u_int32_t *return_netagent_array, u_int32_t *return_netagent_use_flags_array, size_t netagent_array_count, struct necp_client_parameter_netagent_type *required_agent_types, u_int32_t num_required_agent_types, proc_t proc, necp_kernel_policy_id *skip_policy_id);
+static inline struct necp_kernel_socket_policy *necp_socket_find_policy_match_with_info_locked(struct necp_kernel_socket_policy **policy_search_array, struct necp_socket_info *info, necp_kernel_policy_filter *return_filter, u_int32_t *return_route_rule_id_array, size_t *return_route_rule_id_array_count, size_t route_rule_id_array_count, necp_kernel_policy_result *return_service_action, necp_kernel_policy_service *return_service, u_int32_t *return_netagent_array, u_int32_t *return_netagent_use_flags_array, size_t netagent_array_count, struct necp_client_parameter_netagent_type *required_agent_types, u_int32_t num_required_agent_types, proc_t proc, necp_kernel_policy_id *skip_policy_id, struct rtentry *rt, necp_kernel_policy_result *return_drop_dest_policy_result, necp_drop_all_bypass_check_result_t *return_drop_all_bypass);
 
 static necp_kernel_policy_id necp_kernel_ip_output_policy_add(necp_policy_order order, necp_policy_order suborder, u_int32_t session_order, int session_pid, u_int32_t condition_mask, u_int32_t condition_negated_mask, necp_kernel_policy_id cond_policy_id, ifnet_t cond_bound_interface, u_int32_t cond_last_interface_index, u_int16_t cond_protocol, union necp_sockaddr_union *cond_local_start, union necp_sockaddr_union *cond_local_end, u_int8_t cond_local_prefix, union necp_sockaddr_union *cond_remote_start, union necp_sockaddr_union *cond_remote_end, u_int8_t cond_remote_prefix, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter);
 static bool necp_kernel_ip_output_policy_delete(necp_kernel_policy_id policy_id);
@@ -353,15 +392,16 @@ static bool necp_is_range_in_range(struct sockaddr *inner_range_start, struct so
 static bool necp_is_addr_in_subnet(struct sockaddr *addr, struct sockaddr *subnet_addr, u_int8_t subnet_prefix);
 static int necp_addr_compare(struct sockaddr *sa1, struct sockaddr *sa2, int check_port);
 static bool necp_buffer_compare_with_bit_prefix(u_int8_t *p1, u_int8_t *p2, u_int32_t bits);
-static bool necp_is_loopback(struct sockaddr *local_addr, struct sockaddr *remote_addr, struct inpcb *inp, struct mbuf *packet);
+static bool necp_addr_is_empty(struct sockaddr *addr);
+static bool necp_is_loopback(struct sockaddr *local_addr, struct sockaddr *remote_addr, struct inpcb *inp, struct mbuf *packet, u_int32_t bound_interface_index);
 static bool necp_is_intcoproc(struct inpcb *inp, struct mbuf *packet);
 
 struct necp_uuid_id_mapping {
        LIST_ENTRY(necp_uuid_id_mapping) chain;
        uuid_t          uuid;
        u_int32_t       id;
-       u_int32_t       refcount;
-       u_int32_t       table_refcount; // Add to UUID policy table count
+       os_refcnt_t     refcount;
+       u_int32_t       table_usecount; // Add to UUID policy table count
 };
 static size_t necp_num_uuid_app_id_mappings;
 static bool necp_uuid_app_id_mappings_dirty;
@@ -383,7 +423,7 @@ struct necp_string_id_mapping {
        LIST_ENTRY(necp_string_id_mapping) chain;
        char            *string;
        necp_app_id     id;
-       u_int32_t       refcount;
+       os_refcnt_t     refcount;
 };
 static LIST_HEAD(necp_string_id_mapping_list, necp_string_id_mapping) necp_account_id_list;
 static u_int32_t necp_create_string_to_id_mapping(struct necp_string_id_mapping_list *list, char *domain);
@@ -412,9 +452,10 @@ struct necp_route_rule {
        u_int8_t        wifi_action;
        u_int8_t        wired_action;
        u_int8_t        expensive_action;
+       u_int8_t        constrained_action;
        u_int           exception_if_indices[MAX_ROUTE_RULE_INTERFACES];
        u_int8_t        exception_if_actions[MAX_ROUTE_RULE_INTERFACES];
-       u_int32_t       refcount;
+       os_refcnt_t     refcount;
 };
 static LIST_HEAD(necp_route_rule_list, necp_route_rule) necp_route_rules;
 static u_int32_t necp_create_route_rule(struct necp_route_rule_list *list, u_int8_t *route_rules_array, u_int32_t route_rules_array_size);
@@ -434,17 +475,30 @@ static u_int32_t necp_create_aggregate_route_rule(u_int32_t *rule_ids);
 
 // Sysctl definitions
 static int sysctl_handle_necp_level SYSCTL_HANDLER_ARGS;
+static int sysctl_handle_necp_unentitled_level SYSCTL_HANDLER_ARGS;
 
 SYSCTL_NODE(_net, OID_AUTO, necp, CTLFLAG_RW | CTLFLAG_LOCKED, 0, "NECP");
 SYSCTL_INT(_net_necp, NECPCTL_PASS_LOOPBACK, pass_loopback, CTLFLAG_LOCKED | CTLFLAG_RW, &necp_pass_loopback, 0, "");
 SYSCTL_INT(_net_necp, NECPCTL_PASS_KEEPALIVES, pass_keepalives, CTLFLAG_LOCKED | CTLFLAG_RW, &necp_pass_keepalives, 0, "");
+SYSCTL_INT(_net_necp, NECPCTL_PASS_INTERPOSE, pass_interpose, CTLFLAG_LOCKED | CTLFLAG_RW, &necp_pass_interpose, 0, "");
 SYSCTL_INT(_net_necp, NECPCTL_DEBUG, debug, CTLFLAG_LOCKED | CTLFLAG_RW, &necp_debug, 0, "");
+SYSCTL_PROC(_net_necp, NECPCTL_DROP_UNENTITLED_LEVEL, drop_unentitled_level, CTLTYPE_INT | CTLFLAG_LOCKED | CTLFLAG_RW, &necp_drop_unentitled_level, 0, &sysctl_handle_necp_unentitled_level, "IU", "");
 SYSCTL_PROC(_net_necp, NECPCTL_DROP_ALL_LEVEL, drop_all_level, CTLTYPE_INT | CTLFLAG_LOCKED | CTLFLAG_RW, &necp_drop_all_level, 0, &sysctl_handle_necp_level, "IU", "");
 SYSCTL_LONG(_net_necp, NECPCTL_SOCKET_POLICY_COUNT, socket_policy_count, CTLFLAG_LOCKED | CTLFLAG_RD, &necp_kernel_socket_policies_count, "");
 SYSCTL_LONG(_net_necp, NECPCTL_SOCKET_NON_APP_POLICY_COUNT, socket_non_app_policy_count, CTLFLAG_LOCKED | CTLFLAG_RD, &necp_kernel_socket_policies_non_app_count, "");
 SYSCTL_LONG(_net_necp, NECPCTL_IP_POLICY_COUNT, ip_policy_count, CTLFLAG_LOCKED | CTLFLAG_RD, &necp_kernel_ip_output_policies_count, "");
 SYSCTL_INT(_net_necp, NECPCTL_SESSION_COUNT, session_count, CTLFLAG_LOCKED | CTLFLAG_RD, &necp_session_count, 0, "");
 
+static struct necp_drop_dest_policy necp_drop_dest_policy;
+static int necp_drop_dest_debug = 0;    // 0: off, 1: match, >1: every evaluation
+SYSCTL_INT(_net_necp, OID_AUTO, drop_dest_debug, CTLFLAG_LOCKED | CTLFLAG_RW, &necp_drop_dest_debug, 0, "");
+
+static int sysctl_handle_necp_drop_dest_level SYSCTL_HANDLER_ARGS;
+SYSCTL_PROC(_net_necp, OID_AUTO, drop_dest_level, CTLTYPE_STRUCT | CTLFLAG_LOCKED | CTLFLAG_ANYBODY | CTLFLAG_RW,
+    0, 0, &sysctl_handle_necp_drop_dest_level, "S,necp_drop_dest_level", "");
+
+static bool necp_address_matches_drop_dest_policy(union necp_sockaddr_union *, u_int32_t);
+
 // Session order allocation
 static u_int32_t
 necp_allocate_new_session_order(u_int32_t priority, u_int32_t control_unit)
@@ -465,6 +519,9 @@ necp_allocate_new_session_order(u_int32_t priority, u_int32_t control_unit)
 static inline u_int32_t
 necp_get_first_order_for_priority(u_int32_t priority)
 {
+       if (priority == 0) {
+               return 0;
+       }
        return ((priority - 1) * 1000) + 1;
 }
 
@@ -474,72 +531,76 @@ sysctl_handle_necp_level SYSCTL_HANDLER_ARGS
 {
 #pragma unused(arg1, arg2)
        int error = sysctl_handle_int(oidp, oidp->oid_arg1, oidp->oid_arg2, req);
-       if (necp_drop_all_level == 0) {
-               necp_drop_all_order = 0;
+       necp_drop_all_order = necp_get_first_order_for_priority(necp_drop_all_level);
+       return error;
+}
+
+static int
+sysctl_handle_necp_unentitled_level SYSCTL_HANDLER_ARGS
+{
+#pragma unused(arg1, arg2)
+       int error = sysctl_handle_int(oidp, oidp->oid_arg1, oidp->oid_arg2, req);
+       necp_drop_unentitled_order = necp_get_first_order_for_priority(necp_drop_unentitled_level);
+       return error;
+}
+
+// Use a macro here to avoid computing the kauth_cred_t when necp_drop_unentitled_level is 0
+static inline u_int32_t
+_necp_process_drop_order_inner(kauth_cred_t cred)
+{
+       if (priv_check_cred(cred, PRIV_NET_PRIVILEGED_CLIENT_ACCESS, 0) != 0 &&
+           priv_check_cred(cred, PRIV_NET_PRIVILEGED_SERVER_ACCESS, 0) != 0) {
+               return necp_drop_unentitled_order;
        } else {
-               necp_drop_all_order = necp_get_first_order_for_priority(necp_drop_all_level);
+               return 0;
        }
-       return error;
 }
 
+#define necp_process_drop_order(_cred) (necp_drop_unentitled_order != 0 ? _necp_process_drop_order_inner(_cred) : necp_drop_unentitled_order)
+#pragma GCC poison _necp_process_drop_order_inner
+
 // Session fd
 
-static int noop_read(struct fileproc *, struct uio *, int, vfs_context_t);
-static int noop_write(struct fileproc *, struct uio *, int, vfs_context_t);
-static int noop_ioctl(struct fileproc *, unsigned long, caddr_t,
-    vfs_context_t);
-static int noop_select(struct fileproc *, int, void *, vfs_context_t);
 static int necp_session_op_close(struct fileglob *, vfs_context_t);
-static int noop_kqfilter(struct fileproc *, struct knote *,
-    struct kevent_internal_s *, vfs_context_t);
 
 static const struct fileops necp_session_fd_ops = {
-       .fo_type = DTYPE_NETPOLICY,
-       .fo_read = noop_read,
-       .fo_write = noop_write,
-       .fo_ioctl = noop_ioctl,
-       .fo_select = noop_select,
-       .fo_close = necp_session_op_close,
-       .fo_kqfilter = noop_kqfilter,
-       .fo_drain = NULL,
+       .fo_type     = DTYPE_NETPOLICY,
+       .fo_read     = fo_no_read,
+       .fo_write    = fo_no_write,
+       .fo_ioctl    = fo_no_ioctl,
+       .fo_select   = fo_no_select,
+       .fo_close    = necp_session_op_close,
+       .fo_drain    = fo_no_drain,
+       .fo_kqfilter = fo_no_kqfilter,
 };
 
-static int
-noop_read(struct fileproc *fp, struct uio *uio, int flags, vfs_context_t ctx)
-{
-#pragma unused(fp, uio, flags, ctx)
-       return ENXIO;
-}
-
-static int
-noop_write(struct fileproc *fp, struct uio *uio, int flags,
-    vfs_context_t ctx)
-{
-#pragma unused(fp, uio, flags, ctx)
-       return ENXIO;
-}
-
-static int
-noop_ioctl(struct fileproc *fp, unsigned long com, caddr_t data,
-    vfs_context_t ctx)
+static inline necp_drop_all_bypass_check_result_t
+necp_check_drop_all_bypass_result(proc_t proc)
 {
-#pragma unused(fp, com, data, ctx)
-       return ENOTTY;
-}
+       if (proc == NULL) {
+               proc = current_proc();
+               if (proc == NULL) {
+                       return NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE;
+               }
+       }
 
-static int
-noop_select(struct fileproc *fp, int which, void *wql, vfs_context_t ctx)
-{
-#pragma unused(fp, which, wql, ctx)
-       return ENXIO;
-}
+#if defined(XNU_TARGET_OS_OSX)
+       const char *signing_id = NULL;
+       const bool isConfigd = (csproc_get_platform_binary(proc) &&
+           (signing_id = cs_identity_get(proc)) &&
+           (strlen(signing_id) == SIGNING_ID_CONFIGD_LEN) &&
+           (memcmp(signing_id, SIGNING_ID_CONFIGD, SIGNING_ID_CONFIGD_LEN) == 0));
+       if (isConfigd) {
+               return NECP_DROP_ALL_BYPASS_CHECK_RESULT_TRUE;
+       }
+#endif
 
-static int
-noop_kqfilter(struct fileproc *fp, struct knote *kn,
-    struct kevent_internal_s *kev, vfs_context_t ctx)
-{
-#pragma unused(fp, kn, kev, ctx)
-       return ENXIO;
+       const task_t task = proc_task(proc);
+       if (task == NULL || !IOTaskHasEntitlement(task, "com.apple.private.necp.drop_all_bypass")) {
+               return NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE;
+       } else {
+               return NECP_DROP_ALL_BYPASS_CHECK_RESULT_TRUE;
+       }
 }
 
 int
@@ -667,7 +728,7 @@ necp_session_add_policy(struct necp_session *session, struct necp_session_action
                goto done;
        }
 
-       necp_policy_id new_policy_id = necp_handle_policy_add(session, 0, NULL, tlv_buffer, uap->in_buffer_length, 0, &error);
+       necp_policy_id new_policy_id = necp_handle_policy_add(session, tlv_buffer, uap->in_buffer_length, 0, &error);
        if (error != 0) {
                NECPLOG(LOG_ERR, "necp_session_add_policy failed to add policy (%d)", error);
                goto done;
@@ -1010,6 +1071,7 @@ done:
 static int
 necp_session_dump_all(struct necp_session *session, struct necp_session_action_args *uap, int *retval)
 {
+#pragma unused(session)
        int error = 0;
 
        if (uap->out_buffer_length == 0 || uap->out_buffer == 0) {
@@ -1018,7 +1080,7 @@ necp_session_dump_all(struct necp_session *session, struct necp_session_action_a
                goto done;
        }
 
-       error = necp_handle_policy_dump_all(session, 0, NULL, uap->out_buffer, uap->out_buffer_length, 0);
+       error = necp_handle_policy_dump_all(uap->out_buffer, uap->out_buffer_length);
 done:
        *retval = error;
        return error;
@@ -1113,27 +1175,122 @@ done:
        return return_value;
 }
 
-// Kernel Control functions
-static errno_t necp_register_control(void);
-static errno_t necp_ctl_connect(kern_ctl_ref kctlref, struct sockaddr_ctl *sac, void **unitinfo);
-static errno_t necp_ctl_disconnect(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo);
-static errno_t necp_ctl_send(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, mbuf_t m, int flags);
-static void necp_ctl_rcvd(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, int flags);
-static errno_t necp_ctl_getopt(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, int opt, void *data, size_t *len);
-static errno_t necp_ctl_setopt(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, int opt, void *data, size_t len);
+struct necp_resolver_key_state {
+       const struct ccdigest_info *digest_info;
+       uint8_t key[CCSHA256_OUTPUT_SIZE];
+};
+static struct necp_resolver_key_state s_necp_resolver_key_state;
+
+static void
+necp_generate_resolver_key(void)
+{
+       s_necp_resolver_key_state.digest_info = ccsha256_di();
+       cc_rand_generate(s_necp_resolver_key_state.key, sizeof(s_necp_resolver_key_state.key));
+}
+
+static void
+necp_sign_update_context(const struct ccdigest_info *di,
+    cchmac_ctx_t ctx,
+    uuid_t client_id,
+    u_int8_t *query,
+    u_int32_t query_length,
+    u_int8_t *answer,
+    u_int32_t answer_length)
+{
+       const uint8_t context[32] = {[0 ... 31] = 0x20}; // 0x20 repeated 32 times
+       const char *context_string = "NECP Resolver Binder";
+       uint8_t separator = 0;
+       cchmac_update(di, ctx, sizeof(context), context);
+       cchmac_update(di, ctx, strlen(context_string), context_string);
+       cchmac_update(di, ctx, sizeof(separator), &separator);
+       cchmac_update(di, ctx, sizeof(uuid_t), client_id);
+       cchmac_update(di, ctx, sizeof(query_length), &query_length);
+       cchmac_update(di, ctx, query_length, query);
+       cchmac_update(di, ctx, sizeof(answer_length), &answer_length);
+       cchmac_update(di, ctx, answer_length, answer);
+}
+
+int
+necp_sign_resolver_answer(uuid_t client_id, u_int8_t *query, u_int32_t query_length,
+    u_int8_t *answer, u_int32_t answer_length,
+    u_int8_t *tag, u_int32_t *out_tag_length)
+{
+       if (s_necp_resolver_key_state.digest_info == NULL) {
+               return EINVAL;
+       }
+
+       if (query == NULL ||
+           query_length == 0 ||
+           answer == NULL ||
+           answer_length == 0 ||
+           tag == NULL ||
+           out_tag_length == NULL) {
+               return EINVAL;
+       }
+
+       size_t required_tag_length = s_necp_resolver_key_state.digest_info->output_size;
+       if (*out_tag_length < required_tag_length) {
+               return ERANGE;
+       }
+
+       *out_tag_length = required_tag_length;
+
+       cchmac_ctx_decl(s_necp_resolver_key_state.digest_info->state_size,
+           s_necp_resolver_key_state.digest_info->block_size, ctx);
+       cchmac_init(s_necp_resolver_key_state.digest_info, ctx,
+           sizeof(s_necp_resolver_key_state.key),
+           s_necp_resolver_key_state.key);
+       necp_sign_update_context(s_necp_resolver_key_state.digest_info,
+           ctx, client_id, query, query_length,
+           answer, answer_length);
+       cchmac_final(s_necp_resolver_key_state.digest_info, ctx, tag);
+
+       return 0;
+}
+
+bool
+necp_validate_resolver_answer(uuid_t client_id, u_int8_t *query, u_int32_t query_length,
+    u_int8_t *answer, u_int32_t answer_length,
+    u_int8_t *tag, u_int32_t tag_length)
+{
+       if (s_necp_resolver_key_state.digest_info == NULL) {
+               return false;
+       }
+
+       if (query == NULL ||
+           query_length == 0 ||
+           answer == NULL ||
+           answer_length == 0 ||
+           tag == NULL ||
+           tag_length == 0) {
+               return false;
+       }
+
+       size_t required_tag_length = s_necp_resolver_key_state.digest_info->output_size;
+       if (tag_length != required_tag_length) {
+               return false;
+       }
+
+       uint8_t actual_tag[required_tag_length];
+
+       cchmac_ctx_decl(s_necp_resolver_key_state.digest_info->state_size,
+           s_necp_resolver_key_state.digest_info->block_size, ctx);
+       cchmac_init(s_necp_resolver_key_state.digest_info, ctx,
+           sizeof(s_necp_resolver_key_state.key),
+           s_necp_resolver_key_state.key);
+       necp_sign_update_context(s_necp_resolver_key_state.digest_info,
+           ctx, client_id, query, query_length,
+           answer, answer_length);
+       cchmac_final(s_necp_resolver_key_state.digest_info, ctx, actual_tag);
 
-static bool necp_send_ctl_data(struct necp_session *session, u_int8_t *buffer, size_t buffer_size);
+       return cc_cmp_safe(s_necp_resolver_key_state.digest_info->output_size, tag, actual_tag) == 0;
+}
 
 errno_t
 necp_init(void)
 {
        errno_t result = 0;
 
-       result = necp_register_control();
-       if (result != 0) {
-               goto done;
-       }
-
        necp_kernel_policy_grp_attr = lck_grp_attr_alloc_init();
        if (necp_kernel_policy_grp_attr == NULL) {
                NECPLOG0(LOG_ERR, "lck_grp_attr_alloc_init failed");
@@ -1196,6 +1353,8 @@ necp_init(void)
        LIST_INIT(&necp_route_rules);
        LIST_INIT(&necp_aggregate_route_rules);
 
+       necp_generate_resolver_key();
+
        necp_uuid_app_id_hashtbl = hashinit(NECP_UUID_APP_ID_HASH_SIZE, M_NECP, &necp_uuid_app_id_hash_mask);
        necp_uuid_app_id_hash_num_buckets = necp_uuid_app_id_hash_mask + 1;
        necp_num_uuid_app_id_mappings = 0;
@@ -1217,6 +1376,8 @@ necp_init(void)
        memset(&necp_kernel_ip_output_policies_map, 0, sizeof(necp_kernel_ip_output_policies_map));
        necp_kernel_socket_policies_app_layer_map = NULL;
 
+       necp_drop_unentitled_order = necp_get_first_order_for_priority(necp_drop_unentitled_level);
+
 done:
        if (result != 0) {
                if (necp_kernel_policy_mtx_attr != NULL) {
@@ -1243,52 +1404,10 @@ done:
                        lck_grp_attr_free(necp_route_rule_grp_attr);
                        necp_route_rule_grp_attr = NULL;
                }
-               if (necp_kctlref != NULL) {
-                       ctl_deregister(necp_kctlref);
-                       necp_kctlref = NULL;
-               }
        }
        return result;
 }
 
-static errno_t
-necp_register_control(void)
-{
-       struct kern_ctl_reg     kern_ctl;
-       errno_t                         result = 0;
-
-       // Create a tag to allocate memory
-       necp_malloc_tag = OSMalloc_Tagalloc(NECP_CONTROL_NAME, OSMT_DEFAULT);
-
-       // Find a unique value for our interface family
-       result = mbuf_tag_id_find(NECP_CONTROL_NAME, &necp_family);
-       if (result != 0) {
-               NECPLOG(LOG_ERR, "mbuf_tag_id_find_internal failed: %d", result);
-               return result;
-       }
-
-       bzero(&kern_ctl, sizeof(kern_ctl));
-       strlcpy(kern_ctl.ctl_name, NECP_CONTROL_NAME, sizeof(kern_ctl.ctl_name));
-       kern_ctl.ctl_name[sizeof(kern_ctl.ctl_name) - 1] = 0;
-       kern_ctl.ctl_flags = CTL_FLAG_PRIVILEGED; // Require root
-       kern_ctl.ctl_sendsize = 64 * 1024;
-       kern_ctl.ctl_recvsize = 64 * 1024;
-       kern_ctl.ctl_connect = necp_ctl_connect;
-       kern_ctl.ctl_disconnect = necp_ctl_disconnect;
-       kern_ctl.ctl_send = necp_ctl_send;
-       kern_ctl.ctl_rcvd = necp_ctl_rcvd;
-       kern_ctl.ctl_setopt = necp_ctl_setopt;
-       kern_ctl.ctl_getopt = necp_ctl_getopt;
-
-       result = ctl_register(&kern_ctl, &necp_kctlref);
-       if (result != 0) {
-               NECPLOG(LOG_ERR, "ctl_register failed: %d", result);
-               return result;
-       }
-
-       return 0;
-}
-
 static void
 necp_post_change_event(struct kev_necp_policies_changed_data *necp_event_data)
 {
@@ -1307,117 +1426,6 @@ necp_post_change_event(struct kev_necp_policies_changed_data *necp_event_data)
        kev_post_msg(&ev_msg);
 }
 
-static errno_t
-necp_ctl_connect(kern_ctl_ref kctlref, struct sockaddr_ctl *sac, void **unitinfo)
-{
-#pragma unused(kctlref, sac)
-       *unitinfo = necp_create_session();
-       if (*unitinfo == NULL) {
-               // Could not allocate session
-               return ENOBUFS;
-       }
-
-       return 0;
-}
-
-static errno_t
-necp_ctl_disconnect(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo)
-{
-#pragma unused(kctlref, unit)
-       struct necp_session *session = (struct necp_session *)unitinfo;
-       if (session != NULL) {
-               necp_policy_mark_all_for_deletion(session);
-               necp_policy_apply_all(session);
-               necp_delete_session((struct necp_session *)unitinfo);
-       }
-
-       return 0;
-}
-
-
-// Message handling
-static int
-necp_packet_find_tlv(mbuf_t packet, int offset, u_int8_t type, int *err, int next)
-{
-       size_t  cursor                  = offset;
-       int             error                   = 0;
-       u_int32_t       curr_length;
-       u_int8_t        curr_type;
-
-       *err = 0;
-
-       do {
-               if (!next) {
-                       error = mbuf_copydata(packet, cursor, sizeof(curr_type), &curr_type);
-                       if (error) {
-                               *err = ENOENT;
-                               return -1;
-                       }
-               } else {
-                       next = 0;
-                       curr_type = NECP_TLV_NIL;
-               }
-
-               if (curr_type != type) {
-                       cursor += sizeof(curr_type);
-                       error = mbuf_copydata(packet, cursor, sizeof(curr_length), &curr_length);
-                       if (error) {
-                               *err = error;
-                               return -1;
-                       }
-                       cursor += (sizeof(curr_length) + curr_length);
-               }
-       } while (curr_type != type);
-
-       return cursor;
-}
-
-static int
-necp_packet_get_tlv_at_offset(mbuf_t packet, int tlv_offset, u_int32_t buff_len, void *buff, u_int32_t *value_size)
-{
-       int                     error   = 0;
-       u_int32_t       length;
-
-       if (tlv_offset < 0) {
-               return EINVAL;
-       }
-
-       error = mbuf_copydata(packet, tlv_offset + sizeof(u_int8_t), sizeof(length), &length);
-       if (error) {
-               return error;
-       }
-
-       u_int32_t total_len = m_length2(packet, NULL);
-       if (total_len < (tlv_offset + sizeof(u_int8_t) + sizeof(length) + length)) {
-               NECPLOG(LOG_ERR, "Got a bad TLV, length (%u) + offset (%d) < total length (%u)",
-                   length, (tlv_offset + sizeof(u_int8_t) + sizeof(length)), total_len);
-               return EINVAL;
-       }
-
-       if (value_size != NULL) {
-               *value_size = length;
-       }
-
-       if (buff != NULL && buff_len > 0) {
-               u_int32_t to_copy = (length < buff_len) ? length : buff_len;
-               error = mbuf_copydata(packet, tlv_offset + sizeof(u_int8_t) + sizeof(length), to_copy, buff);
-               if (error) {
-                       return error;
-               }
-       }
-
-       return 0;
-}
-
-static u_int8_t *
-necp_buffer_write_packet_header(u_int8_t *buffer, u_int8_t packet_type, u_int8_t flags, u_int32_t message_id)
-{
-       ((struct necp_packet_header *)(void *)buffer)->packet_type = packet_type;
-       ((struct necp_packet_header *)(void *)buffer)->flags = flags;
-       ((struct necp_packet_header *)(void *)buffer)->message_id = message_id;
-       return buffer + sizeof(struct necp_packet_header);
-}
-
 static inline bool
 necp_buffer_write_tlv_validate(u_int8_t *cursor, u_int8_t type, u_int32_t length,
     u_int8_t *buffer, u_int32_t buffer_length)
@@ -1442,7 +1450,8 @@ necp_buffer_write_tlv_if_different(u_int8_t *cursor, u_int8_t type,
     u_int8_t *buffer, u_int32_t buffer_length)
 {
        if (!necp_buffer_write_tlv_validate(cursor, type, length, buffer, buffer_length)) {
-               return NULL;
+               // If we can't fit this TLV, return the current cursor
+               return cursor;
        }
        u_int8_t *next_tlv = (u_int8_t *)(cursor + sizeof(type) + sizeof(length) + length);
        if (*updated || *(u_int8_t *)(cursor) != type) {
@@ -1524,9 +1533,15 @@ necp_buffer_get_tlv_value(u_int8_t *buffer, int tlv_offset, u_int32_t *value_siz
 }
 
 int
-necp_buffer_find_tlv(u_int8_t *buffer, u_int32_t buffer_length, int offset, u_int8_t type, int next)
+necp_buffer_find_tlv(u_int8_t *buffer, u_int32_t buffer_length, int offset, u_int8_t type, int *err, int next)
 {
+       if (err != NULL) {
+               *err = ENOENT;
+       }
        if (offset < 0) {
+               if (err != NULL) {
+                       *err = EINVAL;
+               }
                return -1;
        }
        int cursor = offset;
@@ -1553,6 +1568,9 @@ necp_buffer_find_tlv(u_int8_t *buffer, u_int32_t buffer_length, int offset, u_in
                if (curr_type == type) {
                        // check if entire TLV fits inside buffer
                        if (((u_int32_t)next_cursor) <= buffer_length) {
+                               if (err != NULL) {
+                                       *err = 0;
+                               }
                                return cursor;
                        } else {
                                return -1;
@@ -1563,26 +1581,19 @@ necp_buffer_find_tlv(u_int8_t *buffer, u_int32_t buffer_length, int offset, u_in
 }
 
 static int
-necp_find_tlv(mbuf_t packet, u_int8_t *buffer, u_int32_t buffer_length, int offset, u_int8_t type, int *err, int next)
+necp_find_tlv(u_int8_t *buffer, u_int32_t buffer_length, int offset, u_int8_t type, int *err, int next)
 {
        int cursor = -1;
-       if (packet != NULL) {
-               cursor = necp_packet_find_tlv(packet, offset, type, err, next);
-       } else if (buffer != NULL) {
-               cursor = necp_buffer_find_tlv(buffer, buffer_length, offset, type, next);
+       if (buffer != NULL) {
+               cursor = necp_buffer_find_tlv(buffer, buffer_length, offset, type, err, next);
        }
        return cursor;
 }
 
 static int
-necp_get_tlv_at_offset(mbuf_t packet, u_int8_t *buffer, u_int32_t buffer_length,
+necp_get_tlv_at_offset(u_int8_t *buffer, u_int32_t buffer_length,
     int tlv_offset, u_int32_t out_buffer_length, void *out_buffer, u_int32_t *value_size)
 {
-       if (packet != NULL) {
-               // Handle mbuf parsing
-               return necp_packet_get_tlv_at_offset(packet, tlv_offset, out_buffer_length, out_buffer, value_size);
-       }
-
        if (buffer == NULL) {
                NECPLOG0(LOG_ERR, "necp_get_tlv_at_offset buffer is NULL");
                return EINVAL;
@@ -1633,300 +1644,98 @@ necp_get_tlv_at_offset(mbuf_t packet, u_int8_t *buffer, u_int32_t buffer_length,
 }
 
 static int
-necp_get_tlv(mbuf_t packet, u_int8_t *buffer, u_int32_t buffer_length,
+necp_get_tlv(u_int8_t *buffer, u_int32_t buffer_length,
     int offset, u_int8_t type, u_int32_t buff_len, void *buff, u_int32_t *value_size)
 {
        int error = 0;
 
-       int tlv_offset = necp_find_tlv(packet, buffer, buffer_length, offset, type, &error, 0);
+       int tlv_offset = necp_find_tlv(buffer, buffer_length, offset, type, &error, 0);
        if (tlv_offset < 0) {
                return error;
        }
 
-       return necp_get_tlv_at_offset(packet, buffer, buffer_length, tlv_offset, buff_len, buff, value_size);
+       return necp_get_tlv_at_offset(buffer, buffer_length, tlv_offset, buff_len, buff, value_size);
 }
 
-static bool
-necp_send_ctl_data(struct necp_session *session, u_int8_t *buffer, size_t buffer_size)
+// Session Management
+
+static struct necp_session *
+necp_create_session(void)
 {
-       int             error;
+       struct necp_session *new_session = NULL;
 
-       if (necp_kctlref == NULL || session == NULL || buffer == NULL || buffer_size == 0) {
-               return FALSE;
+       MALLOC(new_session, struct necp_session *, sizeof(*new_session), M_NECP, M_WAITOK | M_ZERO);
+       if (new_session == NULL) {
+               goto done;
        }
 
-       error = ctl_enqueuedata(necp_kctlref, session->control_unit, buffer, buffer_size, CTL_DATA_EOR);
+       new_session->necp_fd_type = necp_fd_type_session;
+       new_session->session_priority = NECP_SESSION_PRIORITY_UNKNOWN;
+       new_session->dirty = FALSE;
+       LIST_INIT(&new_session->policies);
+       lck_mtx_init(&new_session->lock, necp_kernel_policy_mtx_grp, necp_kernel_policy_mtx_attr);
 
-       return error == 0;
-}
+       // Take the lock
+       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
 
-static bool
-necp_send_success_response(struct necp_session *session, u_int8_t packet_type, u_int32_t message_id)
-{
-       bool success = TRUE;
-       u_int8_t *response = NULL;
-       u_int8_t *cursor = NULL;
-       size_t response_size = sizeof(struct necp_packet_header) + sizeof(u_int8_t) + sizeof(u_int32_t);
-       MALLOC(response, u_int8_t *, response_size, M_NECP, M_WAITOK);
-       if (response == NULL) {
-               return FALSE;
-       }
-       cursor = response;
-       cursor = necp_buffer_write_packet_header(cursor, packet_type, NECP_PACKET_FLAGS_RESPONSE, message_id);
-       cursor = necp_buffer_write_tlv(cursor, NECP_TLV_NIL, 0, NULL, response, response_size);
+       // Find the next available control unit
+       u_int32_t control_unit = 1;
+       struct necp_session *next_session = NULL;
+       TAILQ_FOREACH(next_session, &necp_session_list, chain) {
+               if (next_session->control_unit > control_unit) {
+                       // Found a gap, grab this control unit
+                       break;
+               }
 
-       if (!(success = necp_send_ctl_data(session, (u_int8_t *)response, response_size))) {
-               NECPLOG0(LOG_ERR, "Failed to send response");
+               // Try the next control unit, loop around
+               control_unit = next_session->control_unit + 1;
        }
 
-       FREE(response, M_NECP);
-       return success;
-}
+       new_session->control_unit = control_unit;
+       new_session->session_order = necp_allocate_new_session_order(new_session->session_priority, control_unit);
 
-static bool
-necp_send_error_response(struct necp_session *session, u_int8_t packet_type, u_int32_t message_id, u_int32_t error)
-{
-       bool success = TRUE;
-       u_int8_t *response = NULL;
-       u_int8_t *cursor = NULL;
-       size_t response_size = sizeof(struct necp_packet_header) + sizeof(u_int8_t) + sizeof(u_int32_t) + sizeof(u_int32_t);
-       MALLOC(response, u_int8_t *, response_size, M_NECP, M_WAITOK);
-       if (response == NULL) {
-               return FALSE;
+       if (next_session != NULL) {
+               TAILQ_INSERT_BEFORE(next_session, new_session, chain);
+       } else {
+               TAILQ_INSERT_TAIL(&necp_session_list, new_session, chain);
        }
-       cursor = response;
-       cursor = necp_buffer_write_packet_header(cursor, packet_type, NECP_PACKET_FLAGS_RESPONSE, message_id);
-       cursor = necp_buffer_write_tlv(cursor, NECP_TLV_ERROR, sizeof(error), &error, response, response_size);
 
-       if (!(success = necp_send_ctl_data(session, (u_int8_t *)response, response_size))) {
-               NECPLOG0(LOG_ERR, "Failed to send response");
+       necp_session_count++;
+       lck_rw_done(&necp_kernel_policy_lock);
+
+       if (necp_debug) {
+               NECPLOG(LOG_DEBUG, "Created NECP session, control unit %d", control_unit);
        }
 
-       FREE(response, M_NECP);
-       return success;
+done:
+       return new_session;
 }
 
-static bool
-necp_send_policy_id_response(struct necp_session *session, u_int8_t packet_type, u_int32_t message_id, necp_policy_id policy_id)
+static void
+necp_delete_session(struct necp_session *session)
 {
-       bool success = TRUE;
-       u_int8_t *response = NULL;
-       u_int8_t *cursor = NULL;
-       size_t response_size = sizeof(struct necp_packet_header) + sizeof(u_int8_t) + sizeof(u_int32_t) + sizeof(u_int32_t);
-       MALLOC(response, u_int8_t *, response_size, M_NECP, M_WAITOK);
-       if (response == NULL) {
-               return FALSE;
-       }
-       cursor = response;
-       cursor = necp_buffer_write_packet_header(cursor, packet_type, NECP_PACKET_FLAGS_RESPONSE, message_id);
-       cursor = necp_buffer_write_tlv(cursor, NECP_TLV_POLICY_ID, sizeof(policy_id), &policy_id, response, response_size);
+       if (session != NULL) {
+               struct necp_service_registration *service = NULL;
+               struct necp_service_registration *temp_service = NULL;
+               LIST_FOREACH_SAFE(service, &session->services, session_chain, temp_service) {
+                       LIST_REMOVE(service, session_chain);
+                       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
+                       LIST_REMOVE(service, kernel_chain);
+                       lck_rw_done(&necp_kernel_policy_lock);
+                       FREE(service, M_NECP);
+               }
+               if (necp_debug) {
+                       NECPLOG0(LOG_DEBUG, "Deleted NECP session");
+               }
 
-       if (!(success = necp_send_ctl_data(session, (u_int8_t *)response, response_size))) {
-               NECPLOG0(LOG_ERR, "Failed to send response");
-       }
+               lck_rw_lock_exclusive(&necp_kernel_policy_lock);
+               TAILQ_REMOVE(&necp_session_list, session, chain);
+               necp_session_count--;
+               lck_rw_done(&necp_kernel_policy_lock);
 
-       FREE(response, M_NECP);
-       return success;
-}
-
-static errno_t
-necp_ctl_send(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, mbuf_t packet, int flags)
-{
-#pragma unused(kctlref, unit, flags)
-       struct necp_session *session = (struct necp_session *)unitinfo;
-       struct necp_packet_header header;
-       int error = 0;
-
-       if (session == NULL) {
-               NECPLOG0(LOG_ERR, "Got a NULL session");
-               error = EINVAL;
-               goto done;
-       }
-
-       if (mbuf_pkthdr_len(packet) < sizeof(header)) {
-               NECPLOG(LOG_ERR, "Got a bad packet, length (%lu) < sizeof header (%lu)", mbuf_pkthdr_len(packet), sizeof(header));
-               error = EINVAL;
-               goto done;
-       }
-
-       error = mbuf_copydata(packet, 0, sizeof(header), &header);
-       if (error) {
-               NECPLOG(LOG_ERR, "mbuf_copydata failed for the header: %d", error);
-               error = ENOBUFS;
-               goto done;
-       }
-
-       if (session->proc_locked) {
-               // Verify that the calling process is allowed to send messages
-               uuid_t proc_uuid;
-               proc_getexecutableuuid(current_proc(), proc_uuid, sizeof(proc_uuid));
-               if (uuid_compare(proc_uuid, session->proc_uuid) != 0) {
-                       necp_send_error_response(session, header.packet_type, header.message_id, NECP_ERROR_INVALID_PROCESS);
-                       goto done;
-               }
-       } else {
-               // If not locked, update the proc_uuid and proc_pid of the session
-               proc_getexecutableuuid(current_proc(), session->proc_uuid, sizeof(session->proc_uuid));
-               session->proc_pid = proc_pid(current_proc());
-       }
-
-       switch (header.packet_type) {
-       case NECP_PACKET_TYPE_POLICY_ADD: {
-               necp_handle_policy_add(session, header.message_id, packet, NULL, 0, sizeof(header), NULL);
-               break;
-       }
-       case NECP_PACKET_TYPE_POLICY_GET: {
-               necp_handle_policy_get(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_POLICY_DELETE: {
-               necp_handle_policy_delete(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_POLICY_APPLY_ALL: {
-               necp_handle_policy_apply_all(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_POLICY_LIST_ALL: {
-               necp_handle_policy_list_all(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_POLICY_DELETE_ALL: {
-               necp_handle_policy_delete_all(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_POLICY_DUMP_ALL: {
-               necp_handle_policy_dump_all(session, header.message_id, packet, 0, 0, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_SET_SESSION_PRIORITY: {
-               necp_handle_set_session_priority(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_LOCK_SESSION_TO_PROC: {
-               necp_handle_lock_session_to_proc(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_REGISTER_SERVICE: {
-               necp_handle_register_service(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       case NECP_PACKET_TYPE_UNREGISTER_SERVICE: {
-               necp_handle_unregister_service(session, header.message_id, packet, sizeof(header));
-               break;
-       }
-       default: {
-               NECPLOG(LOG_ERR, "Received unknown message type %d", header.packet_type);
-               necp_send_error_response(session, header.packet_type, header.message_id, NECP_ERROR_UNKNOWN_PACKET_TYPE);
-               break;
-       }
-       }
-
-done:
-       mbuf_freem(packet);
-       return error;
-}
-
-static void
-necp_ctl_rcvd(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, int flags)
-{
-#pragma unused(kctlref, unit, unitinfo, flags)
-       return;
-}
-
-static errno_t
-necp_ctl_getopt(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, int opt, void *data, size_t *len)
-{
-#pragma unused(kctlref, unit, unitinfo, opt, data, len)
-       return 0;
-}
-
-static errno_t
-necp_ctl_setopt(kern_ctl_ref kctlref, u_int32_t unit, void *unitinfo, int opt, void *data, size_t len)
-{
-#pragma unused(kctlref, unit, unitinfo, opt, data, len)
-       return 0;
-}
-
-// Session Management
-
-static struct necp_session *
-necp_create_session(void)
-{
-       struct necp_session *new_session = NULL;
-
-       MALLOC(new_session, struct necp_session *, sizeof(*new_session), M_NECP, M_WAITOK | M_ZERO);
-       if (new_session == NULL) {
-               goto done;
-       }
-
-       new_session->necp_fd_type = necp_fd_type_session;
-       new_session->session_priority = NECP_SESSION_PRIORITY_UNKNOWN;
-       new_session->dirty = FALSE;
-       LIST_INIT(&new_session->policies);
-       lck_mtx_init(&new_session->lock, necp_kernel_policy_mtx_grp, necp_kernel_policy_mtx_attr);
-
-       // Take the lock
-       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
-
-       // Find the next available control unit
-       u_int32_t control_unit = 1;
-       struct necp_session *next_session = NULL;
-       TAILQ_FOREACH(next_session, &necp_session_list, chain) {
-               if (next_session->control_unit > control_unit) {
-                       // Found a gap, grab this control unit
-                       break;
-               }
-
-               // Try the next control unit, loop around
-               control_unit = next_session->control_unit + 1;
-       }
-
-       new_session->control_unit = control_unit;
-       new_session->session_order = necp_allocate_new_session_order(new_session->session_priority, control_unit);
-
-       if (next_session != NULL) {
-               TAILQ_INSERT_BEFORE(next_session, new_session, chain);
-       } else {
-               TAILQ_INSERT_TAIL(&necp_session_list, new_session, chain);
-       }
-
-       necp_session_count++;
-       lck_rw_done(&necp_kernel_policy_lock);
-
-       if (necp_debug) {
-               NECPLOG(LOG_DEBUG, "Created NECP session, control unit %d", control_unit);
-       }
-
-done:
-       return new_session;
-}
-
-static void
-necp_delete_session(struct necp_session *session)
-{
-       if (session != NULL) {
-               struct necp_service_registration *service = NULL;
-               struct necp_service_registration *temp_service = NULL;
-               LIST_FOREACH_SAFE(service, &session->services, session_chain, temp_service) {
-                       LIST_REMOVE(service, session_chain);
-                       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
-                       LIST_REMOVE(service, kernel_chain);
-                       lck_rw_done(&necp_kernel_policy_lock);
-                       FREE(service, M_NECP);
-               }
-               if (necp_debug) {
-                       NECPLOG0(LOG_DEBUG, "Deleted NECP session");
-               }
-
-               lck_rw_lock_exclusive(&necp_kernel_policy_lock);
-               TAILQ_REMOVE(&necp_session_list, session, chain);
-               necp_session_count--;
-               lck_rw_done(&necp_kernel_policy_lock);
-
-               lck_mtx_destroy(&session->lock, necp_kernel_policy_mtx_grp);
-               FREE(session, M_NECP);
-       }
+               lck_mtx_destroy(&session->lock, necp_kernel_policy_mtx_grp);
+               FREE(session, M_NECP);
+       }
 }
 
 // Session Policy Management
@@ -1981,7 +1790,8 @@ necp_policy_result_is_valid(u_int8_t *buffer, u_int32_t length)
        case NECP_POLICY_RESULT_PASS:
        case NECP_POLICY_RESULT_DROP:
        case NECP_POLICY_RESULT_ROUTE_RULES:
-       case NECP_POLICY_RESULT_SCOPED_DIRECT: {
+       case NECP_POLICY_RESULT_SCOPED_DIRECT:
+       case NECP_POLICY_RESULT_ALLOW_UNENTITLED: {
                validated = TRUE;
                break;
        }
@@ -2099,7 +1909,8 @@ necp_policy_condition_is_valid(u_int8_t *buffer, u_int32_t length, u_int8_t poli
            policy_result_type == NECP_POLICY_RESULT_ROUTE_RULES ||
            policy_result_type == NECP_POLICY_RESULT_USE_NETAGENT ||
            policy_result_type == NECP_POLICY_RESULT_NETAGENT_SCOPED ||
-           policy_result_type == NECP_POLICY_RESULT_SCOPED_DIRECT) ? TRUE : FALSE;
+           policy_result_type == NECP_POLICY_RESULT_SCOPED_DIRECT ||
+           policy_result_type == NECP_POLICY_RESULT_ALLOW_UNENTITLED) ? TRUE : FALSE;
        u_int32_t condition_length = necp_policy_condition_get_value_length_from_buffer(buffer, length);
        u_int8_t *condition_value = necp_policy_condition_get_value_pointer_from_buffer(buffer, length);
        u_int8_t type = necp_policy_condition_get_type_from_buffer(buffer, length);
@@ -2131,7 +1942,10 @@ necp_policy_condition_is_valid(u_int8_t *buffer, u_int32_t length, u_int8_t poli
        }
        case NECP_POLICY_CONDITION_DEFAULT:
        case NECP_POLICY_CONDITION_ALL_INTERFACES:
-       case NECP_POLICY_CONDITION_ENTITLEMENT: {
+       case NECP_POLICY_CONDITION_ENTITLEMENT:
+       case NECP_POLICY_CONDITION_PLATFORM_BINARY:
+       case NECP_POLICY_CONDITION_HAS_CLIENT:
+       case NECP_POLICY_CONDITION_LOCAL_NETWORKS: {
                if (!(flags & NECP_POLICY_CONDITION_FLAGS_NEGATIVE)) {
                        validated = TRUE;
                }
@@ -2181,6 +1995,43 @@ necp_policy_condition_is_valid(u_int8_t *buffer, u_int32_t length, u_int8_t poli
                }
                break;
        }
+       case NECP_POLICY_CONDITION_FLOW_IP_PROTOCOL: {
+               if (condition_length >= sizeof(u_int16_t)) {
+                       validated = TRUE;
+               }
+               break;
+       }
+       case NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR:
+       case NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR: {
+               if (condition_length >= sizeof(struct necp_policy_condition_addr) &&
+                   necp_address_is_valid(&((struct necp_policy_condition_addr *)(void *)condition_value)->address.sa)) {
+                       validated = TRUE;
+               }
+               break;
+       }
+       case NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR_RANGE:
+       case NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR_RANGE: {
+               if (condition_length >= sizeof(struct necp_policy_condition_addr_range) &&
+                   necp_address_is_valid(&((struct necp_policy_condition_addr_range *)(void *)condition_value)->start_address.sa) &&
+                   necp_address_is_valid(&((struct necp_policy_condition_addr_range *)(void *)condition_value)->end_address.sa)) {
+                       validated = TRUE;
+               }
+               break;
+       }
+       case NECP_POLICY_CONDITION_CLIENT_FLAGS: {
+               if (condition_length == 0 || condition_length >= sizeof(u_int32_t)) {
+                       validated = TRUE;
+               }
+               break;
+       }
+       case NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR_EMPTY: {
+               validated = TRUE;
+               break;
+       }
+       case NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR_EMPTY: {
+               validated = TRUE;
+               break;
+       }
        default: {
                validated = FALSE;
                break;
@@ -2260,164 +2111,8 @@ necp_get_posix_error_for_necp_error(int response_error)
        }
 }
 
-static void
-necp_handle_set_session_priority(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-       int error;
-       struct necp_session_policy *policy = NULL;
-       struct necp_session_policy *temp_policy = NULL;
-       u_int32_t response_error = NECP_ERROR_INTERNAL;
-       u_int32_t requested_session_priority = NECP_SESSION_PRIORITY_UNKNOWN;
-
-       // Read policy id
-       error = necp_get_tlv(packet, NULL, 0, offset, NECP_TLV_SESSION_PRIORITY, sizeof(requested_session_priority), &requested_session_priority, NULL);
-       if (error) {
-               NECPLOG(LOG_ERR, "Failed to get session priority: %d", error);
-               response_error = NECP_ERROR_INVALID_TLV;
-               goto fail;
-       }
-
-       if (session == NULL) {
-               NECPLOG0(LOG_ERR, "Failed to find session");
-               response_error = NECP_ERROR_INTERNAL;
-               goto fail;
-       }
-
-       // Enforce special session priorities with entitlements
-       if (requested_session_priority == NECP_SESSION_PRIORITY_CONTROL ||
-           requested_session_priority == NECP_SESSION_PRIORITY_PRIVILEGED_TUNNEL) {
-               errno_t cred_result = priv_check_cred(kauth_cred_get(), PRIV_NET_PRIVILEGED_NECP_POLICIES, 0);
-               if (cred_result != 0) {
-                       NECPLOG(LOG_ERR, "Session does not hold necessary entitlement to claim priority level %d", requested_session_priority);
-                       goto fail;
-               }
-       }
-
-       if (session->session_priority != requested_session_priority) {
-               session->session_priority = requested_session_priority;
-               session->session_order = necp_allocate_new_session_order(session->session_priority, session->control_unit);
-               session->dirty = TRUE;
-
-               // Mark all policies as needing updates
-               LIST_FOREACH_SAFE(policy, &session->policies, chain, temp_policy) {
-                       policy->pending_update = TRUE;
-               }
-       }
-
-       necp_send_success_response(session, NECP_PACKET_TYPE_SET_SESSION_PRIORITY, message_id);
-       return;
-
-fail:
-       necp_send_error_response(session, NECP_PACKET_TYPE_SET_SESSION_PRIORITY, message_id, response_error);
-}
-
-static void
-necp_handle_lock_session_to_proc(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-#pragma unused(packet, offset)
-       // proc_uuid already filled out
-       session->proc_locked = TRUE;
-       necp_send_success_response(session, NECP_PACKET_TYPE_LOCK_SESSION_TO_PROC, message_id);
-}
-
-static void
-necp_handle_register_service(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-       int error;
-       struct necp_service_registration *new_service = NULL;
-       u_int32_t response_error = NECP_ERROR_INTERNAL;
-       uuid_t service_uuid;
-       uuid_clear(service_uuid);
-
-       if (session == NULL) {
-               NECPLOG0(LOG_ERR, "Failed to find session");
-               response_error = NECP_ERROR_INTERNAL;
-               goto fail;
-       }
-
-       // Enforce entitlements
-       errno_t cred_result = priv_check_cred(kauth_cred_get(), PRIV_NET_PRIVILEGED_NECP_POLICIES, 0);
-       if (cred_result != 0) {
-               NECPLOG0(LOG_ERR, "Session does not hold necessary entitlement to register service");
-               goto fail;
-       }
-
-       // Read service uuid
-       error = necp_get_tlv(packet, NULL, 0, offset, NECP_TLV_SERVICE_UUID, sizeof(uuid_t), service_uuid, NULL);
-       if (error) {
-               NECPLOG(LOG_ERR, "Failed to get service UUID: %d", error);
-               response_error = NECP_ERROR_INVALID_TLV;
-               goto fail;
-       }
-
-       MALLOC(new_service, struct necp_service_registration *, sizeof(*new_service), M_NECP, M_WAITOK);
-       if (new_service == NULL) {
-               NECPLOG0(LOG_ERR, "Failed to allocate service registration");
-               response_error = NECP_ERROR_INTERNAL;
-               goto fail;
-       }
-
-       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
-       memset(new_service, 0, sizeof(*new_service));
-       new_service->service_id = necp_create_uuid_service_id_mapping(service_uuid);
-       LIST_INSERT_HEAD(&session->services, new_service, session_chain);
-       LIST_INSERT_HEAD(&necp_registered_service_list, new_service, kernel_chain);
-       lck_rw_done(&necp_kernel_policy_lock);
-
-       necp_send_success_response(session, NECP_PACKET_TYPE_REGISTER_SERVICE, message_id);
-       return;
-fail:
-       necp_send_error_response(session, NECP_PACKET_TYPE_REGISTER_SERVICE, message_id, response_error);
-}
-
-static void
-necp_handle_unregister_service(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-       int error;
-       struct necp_service_registration *service = NULL;
-       struct necp_service_registration *temp_service = NULL;
-       u_int32_t response_error = NECP_ERROR_INTERNAL;
-       struct necp_uuid_id_mapping *mapping = NULL;
-       uuid_t service_uuid;
-       uuid_clear(service_uuid);
-
-       if (session == NULL) {
-               NECPLOG0(LOG_ERR, "Failed to find session");
-               response_error = NECP_ERROR_INTERNAL;
-               goto fail;
-       }
-
-       // Read service uuid
-       error = necp_get_tlv(packet, NULL, 0, offset, NECP_TLV_SERVICE_UUID, sizeof(uuid_t), service_uuid, NULL);
-       if (error) {
-               NECPLOG(LOG_ERR, "Failed to get service UUID: %d", error);
-               response_error = NECP_ERROR_INVALID_TLV;
-               goto fail;
-       }
-
-       // Mark remove all matching services for this session
-       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
-       mapping = necp_uuid_lookup_service_id_locked(service_uuid);
-       if (mapping != NULL) {
-               LIST_FOREACH_SAFE(service, &session->services, session_chain, temp_service) {
-                       if (service->service_id == mapping->id) {
-                               LIST_REMOVE(service, session_chain);
-                               LIST_REMOVE(service, kernel_chain);
-                               FREE(service, M_NECP);
-                       }
-               }
-               necp_remove_uuid_service_id_mapping(service_uuid);
-       }
-       lck_rw_done(&necp_kernel_policy_lock);
-
-       necp_send_success_response(session, NECP_PACKET_TYPE_UNREGISTER_SERVICE, message_id);
-       return;
-fail:
-       necp_send_error_response(session, NECP_PACKET_TYPE_UNREGISTER_SERVICE, message_id, response_error);
-}
-
 static necp_policy_id
-necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_t packet,
+necp_handle_policy_add(struct necp_session *session,
     u_int8_t *tlv_buffer, size_t tlv_buffer_length, int offset, int *return_error)
 {
        bool has_default_condition = FALSE;
@@ -2445,7 +2140,7 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
        u_int32_t policy_result_size = 0;
 
        // Read policy order
-       error = necp_get_tlv(packet, tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_ORDER, sizeof(order), &order, NULL);
+       error = necp_get_tlv(tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_ORDER, sizeof(order), &order, NULL);
        if (error) {
                NECPLOG(LOG_ERR, "Failed to get policy order: %d", error);
                response_error = NECP_ERROR_INVALID_TLV;
@@ -2453,8 +2148,13 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
        }
 
        // Read policy result
-       cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_RESULT, &error, 0);
-       error = necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &policy_result_size);
+       cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_RESULT, &error, 0);
+       if (error || cursor < 0) {
+               NECPLOG(LOG_ERR, "Failed to find policy result TLV: %d", error);
+               response_error = NECP_ERROR_INVALID_TLV;
+               goto fail;
+       }
+       error = necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &policy_result_size);
        if (error || policy_result_size == 0) {
                NECPLOG(LOG_ERR, "Failed to get policy result length: %d", error);
                response_error = NECP_ERROR_INVALID_TLV;
@@ -2471,7 +2171,7 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
                response_error = NECP_ERROR_INTERNAL;
                goto fail;
        }
-       error = necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, policy_result_size, policy_result, NULL);
+       error = necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, policy_result_size, policy_result, NULL);
        if (error) {
                NECPLOG(LOG_ERR, "Failed to get policy result: %d", error);
                response_error = NECP_ERROR_POLICY_RESULT_INVALID;
@@ -2485,13 +2185,17 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
 
        if (necp_policy_result_requires_route_rules(policy_result, policy_result_size)) {
                // Read route rules conditions
-               for (cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, offset, NECP_TLV_ROUTE_RULE, &error, 0);
+               for (cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, offset, NECP_TLV_ROUTE_RULE, &error, 0);
                    cursor >= 0;
-                   cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_ROUTE_RULE, &error, 1)) {
+                   cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_ROUTE_RULE, &error, 1)) {
                        u_int32_t route_rule_size = 0;
-                       necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &route_rule_size);
-                       if (route_rule_size > 0) {
-                               route_rules_array_size += (sizeof(u_int8_t) + sizeof(u_int32_t) + route_rule_size);
+                       necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &route_rule_size);
+                       if (os_add_overflow(route_rules_array_size,
+                           (sizeof(u_int8_t) + sizeof(u_int32_t) + route_rule_size),
+                           &route_rules_array_size)) {
+                               NECPLOG0(LOG_ERR, "Route rules size overflowed, too large");
+                               response_error = NECP_ERROR_INVALID_TLV;
+                               goto fail;
                        }
                }
 
@@ -2513,13 +2217,14 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
                }
 
                route_rules_array_cursor = 0;
-               for (cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, offset, NECP_TLV_ROUTE_RULE, &error, 0);
+               for (cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, offset, NECP_TLV_ROUTE_RULE, &error, 0);
                    cursor >= 0;
-                   cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_ROUTE_RULE, &error, 1)) {
+                   cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_ROUTE_RULE, &error, 1)) {
                        u_int8_t route_rule_type = NECP_TLV_ROUTE_RULE;
                        u_int32_t route_rule_size = 0;
-                       necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &route_rule_size);
-                       if (route_rule_size > 0 && route_rule_size <= (route_rules_array_size - route_rules_array_cursor)) {
+                       necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &route_rule_size);
+                       if (route_rule_size > 0 &&
+                           (sizeof(route_rule_type) + sizeof(route_rule_size) + route_rule_size) <= (route_rules_array_size - route_rules_array_cursor)) {
                                // Add type
                                memcpy((route_rules_array + route_rules_array_cursor), &route_rule_type, sizeof(route_rule_type));
                                route_rules_array_cursor += sizeof(route_rule_type);
@@ -2529,7 +2234,7 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
                                route_rules_array_cursor += sizeof(route_rule_size);
 
                                // Add value
-                               necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, route_rule_size, (route_rules_array + route_rules_array_cursor), NULL);
+                               necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, route_rule_size, (route_rules_array + route_rules_array_cursor), NULL);
 
                                if (!necp_policy_route_rule_is_valid((route_rules_array + route_rules_array_cursor), route_rule_size)) {
                                        NECPLOG0(LOG_ERR, "Failed to validate policy route rule");
@@ -2552,14 +2257,20 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
        }
 
        // Read policy conditions
-       for (cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_CONDITION, &error, 0);
+       for (cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_CONDITION, &error, 0);
            cursor >= 0;
-           cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_POLICY_CONDITION, &error, 1)) {
+           cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_POLICY_CONDITION, &error, 1)) {
                u_int32_t condition_size = 0;
-               necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &condition_size);
+               necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &condition_size);
 
                if (condition_size > 0) {
-                       conditions_array_size += (sizeof(u_int8_t) + sizeof(u_int32_t) + condition_size);
+                       if (os_add_overflow(conditions_array_size,
+                           (sizeof(u_int8_t) + sizeof(u_int32_t) + condition_size),
+                           &conditions_array_size)) {
+                               NECPLOG0(LOG_ERR, "Conditions size overflowed, too large");
+                               response_error = NECP_ERROR_INVALID_TLV;
+                               goto fail;
+                       }
                }
        }
 
@@ -2581,13 +2292,14 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
        }
 
        conditions_array_cursor = 0;
-       for (cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_CONDITION, &error, 0);
+       for (cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, offset, NECP_TLV_POLICY_CONDITION, &error, 0);
            cursor >= 0;
-           cursor = necp_find_tlv(packet, tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_POLICY_CONDITION, &error, 1)) {
+           cursor = necp_find_tlv(tlv_buffer, tlv_buffer_length, cursor, NECP_TLV_POLICY_CONDITION, &error, 1)) {
                u_int8_t condition_type = NECP_TLV_POLICY_CONDITION;
                u_int32_t condition_size = 0;
-               necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &condition_size);
-               if (condition_size > 0 && condition_size <= (conditions_array_size - conditions_array_cursor)) {
+               necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, 0, NULL, &condition_size);
+               if (condition_size > 0 &&
+                   (sizeof(condition_type) + sizeof(condition_size) + condition_size) <= (conditions_array_size - conditions_array_cursor)) {
                        // Add type
                        memcpy((conditions_array + conditions_array_cursor), &condition_type, sizeof(condition_type));
                        conditions_array_cursor += sizeof(condition_type);
@@ -2597,7 +2309,7 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
                        conditions_array_cursor += sizeof(condition_size);
 
                        // Add value
-                       necp_get_tlv_at_offset(packet, tlv_buffer, tlv_buffer_length, cursor, condition_size, (conditions_array + conditions_array_cursor), NULL);
+                       necp_get_tlv_at_offset(tlv_buffer, tlv_buffer_length, cursor, condition_size, (conditions_array + conditions_array_cursor), NULL);
                        if (!necp_policy_condition_is_valid((conditions_array + conditions_array_cursor), condition_size, necp_policy_result_get_type_from_buffer(policy_result, policy_result_size))) {
                                NECPLOG0(LOG_ERR, "Failed to validate policy condition");
                                response_error = NECP_ERROR_POLICY_CONDITIONS_INVALID;
@@ -2652,9 +2364,6 @@ necp_handle_policy_add(struct necp_session *session, u_int32_t message_id, mbuf_
                goto fail;
        }
 
-       if (packet != NULL) {
-               necp_send_policy_id_response(session, NECP_PACKET_TYPE_POLICY_ADD, message_id, policy->local_id);
-       }
        return policy->local_id;
 
 fail:
@@ -2668,168 +2377,12 @@ fail:
                FREE(route_rules_array, M_NECP);
        }
 
-       if (packet != NULL) {
-               necp_send_error_response(session, NECP_PACKET_TYPE_POLICY_ADD, message_id, response_error);
-       }
        if (return_error != NULL) {
                *return_error = necp_get_posix_error_for_necp_error(response_error);
        }
        return 0;
 }
 
-static void
-necp_handle_policy_get(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-#pragma unused(offset)
-       int error;
-       u_int8_t *response = NULL;
-       u_int8_t *cursor = NULL;
-       u_int32_t response_error = NECP_ERROR_INTERNAL;
-       necp_policy_id policy_id = 0;
-       u_int32_t order_tlv_size = 0;
-       u_int32_t result_tlv_size = 0;
-       u_int32_t response_size = 0;
-
-       struct necp_session_policy *policy = NULL;
-
-       // Read policy id
-       error = necp_get_tlv(packet, NULL, 0, offset, NECP_TLV_POLICY_ID, sizeof(policy_id), &policy_id, NULL);
-       if (error) {
-               NECPLOG(LOG_ERR, "Failed to get policy id: %d", error);
-               response_error = NECP_ERROR_INVALID_TLV;
-               goto fail;
-       }
-
-       policy = necp_policy_find(session, policy_id);
-       if (policy == NULL || policy->pending_deletion) {
-               NECPLOG(LOG_ERR, "Failed to find policy with id %d", policy_id);
-               response_error = NECP_ERROR_POLICY_ID_NOT_FOUND;
-               goto fail;
-       }
-
-       order_tlv_size = sizeof(u_int8_t) + sizeof(u_int32_t) + sizeof(necp_policy_order);
-       result_tlv_size = (policy->result_size ? (sizeof(u_int8_t) + sizeof(u_int32_t) + policy->result_size) : 0);
-       response_size = sizeof(struct necp_packet_header) + order_tlv_size + result_tlv_size + policy->conditions_size;
-       MALLOC(response, u_int8_t *, response_size, M_NECP, M_WAITOK);
-       if (response == NULL) {
-               necp_send_error_response(session, NECP_PACKET_TYPE_POLICY_GET, message_id, NECP_ERROR_INTERNAL);
-               return;
-       }
-
-       cursor = response;
-       cursor = necp_buffer_write_packet_header(cursor, NECP_PACKET_TYPE_POLICY_GET, NECP_PACKET_FLAGS_RESPONSE, message_id);
-       cursor = necp_buffer_write_tlv(cursor, NECP_TLV_POLICY_ORDER, sizeof(necp_policy_order), &policy->order, response, response_size);
-
-       if (result_tlv_size) {
-               cursor = necp_buffer_write_tlv(cursor, NECP_TLV_POLICY_RESULT, policy->result_size, &policy->result, response, response_size);
-       }
-       if (policy->conditions_size) {
-               memcpy(((u_int8_t *)(void *)(cursor)), policy->conditions, policy->conditions_size);
-       }
-
-       if (!necp_send_ctl_data(session, (u_int8_t *)response, response_size)) {
-               NECPLOG0(LOG_ERR, "Failed to send response");
-       }
-
-       FREE(response, M_NECP);
-       return;
-
-fail:
-       necp_send_error_response(session, NECP_PACKET_TYPE_POLICY_GET, message_id, response_error);
-}
-
-static void
-necp_handle_policy_delete(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-       int error;
-       u_int32_t response_error = NECP_ERROR_INTERNAL;
-       necp_policy_id policy_id = 0;
-
-       struct necp_session_policy *policy = NULL;
-
-       // Read policy id
-       error = necp_get_tlv(packet, NULL, 0, offset, NECP_TLV_POLICY_ID, sizeof(policy_id), &policy_id, NULL);
-       if (error) {
-               NECPLOG(LOG_ERR, "Failed to get policy id: %d", error);
-               response_error = NECP_ERROR_INVALID_TLV;
-               goto fail;
-       }
-
-       policy = necp_policy_find(session, policy_id);
-       if (policy == NULL || policy->pending_deletion) {
-               NECPLOG(LOG_ERR, "Failed to find policy with id %d", policy_id);
-               response_error = NECP_ERROR_POLICY_ID_NOT_FOUND;
-               goto fail;
-       }
-
-       necp_policy_mark_for_deletion(session, policy);
-
-       necp_send_success_response(session, NECP_PACKET_TYPE_POLICY_DELETE, message_id);
-       return;
-
-fail:
-       necp_send_error_response(session, NECP_PACKET_TYPE_POLICY_DELETE, message_id, response_error);
-}
-
-static void
-necp_handle_policy_apply_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-#pragma unused(packet, offset)
-       necp_policy_apply_all(session);
-       necp_send_success_response(session, NECP_PACKET_TYPE_POLICY_APPLY_ALL, message_id);
-}
-
-static void
-necp_handle_policy_list_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-#pragma unused(packet, offset)
-       u_int32_t tlv_size = (sizeof(u_int8_t) + sizeof(u_int32_t) + sizeof(u_int32_t));
-       u_int32_t response_size = 0;
-       u_int8_t *response = NULL;
-       u_int8_t *cursor = NULL;
-       int num_policies = 0;
-       int cur_policy_index = 0;
-       struct necp_session_policy *policy;
-
-       LIST_FOREACH(policy, &session->policies, chain) {
-               if (!policy->pending_deletion) {
-                       num_policies++;
-               }
-       }
-
-       // Create a response with one Policy ID TLV for each policy
-       response_size = sizeof(struct necp_packet_header) + num_policies * tlv_size;
-       MALLOC(response, u_int8_t *, response_size, M_NECP, M_WAITOK);
-       if (response == NULL) {
-               necp_send_error_response(session, NECP_PACKET_TYPE_POLICY_LIST_ALL, message_id, NECP_ERROR_INTERNAL);
-               return;
-       }
-
-       cursor = response;
-       cursor = necp_buffer_write_packet_header(cursor, NECP_PACKET_TYPE_POLICY_LIST_ALL, NECP_PACKET_FLAGS_RESPONSE, message_id);
-
-       LIST_FOREACH(policy, &session->policies, chain) {
-               if (!policy->pending_deletion && cur_policy_index < num_policies) {
-                       cursor = necp_buffer_write_tlv(cursor, NECP_TLV_POLICY_ID, sizeof(u_int32_t), &policy->local_id, response, response_size);
-                       cur_policy_index++;
-               }
-       }
-
-       if (!necp_send_ctl_data(session, (u_int8_t *)response, response_size)) {
-               NECPLOG0(LOG_ERR, "Failed to send response");
-       }
-
-       FREE(response, M_NECP);
-}
-
-static void
-necp_handle_policy_delete_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet, int offset)
-{
-#pragma unused(packet, offset)
-       necp_policy_mark_all_for_deletion(session);
-       necp_send_success_response(session, NECP_PACKET_TYPE_POLICY_DELETE_ALL, message_id);
-}
-
 static necp_policy_id
 necp_policy_get_new_id(struct necp_session *session)
 {
@@ -2945,10 +2498,8 @@ necp_policy_get_new_id(struct necp_session *session)
  *     ...
  */
 static int
-necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id, mbuf_t packet,
-    user_addr_t out_buffer, size_t out_buffer_length, int offset)
+necp_handle_policy_dump_all(user_addr_t out_buffer, size_t out_buffer_length)
 {
-#pragma unused(offset)
        struct necp_kernel_socket_policy *policy = NULL;
        int policy_i;
        int policy_count = 0;
@@ -3035,6 +2586,9 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
                        if (condition_mask & NECP_KERNEL_CONDITION_ALL_INTERFACES) {
                                num_conditions++;
                        }
+                       if (condition_mask & NECP_KERNEL_CONDITION_HAS_CLIENT) {
+                               num_conditions++;
+                       }
                        if (condition_mask & NECP_KERNEL_CONDITION_BOUND_INTERFACE) {
                                snprintf(if_name, IFXNAMSIZ, "%s%d", ifnet_name(policy->cond_bound_interface), ifnet_unit(policy->cond_bound_interface));
                                condition_tlv_length += strlen(if_name) + 1;
@@ -3086,6 +2640,12 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
                                condition_tlv_length += entitlement_len;
                                num_conditions++;
                        }
+                       if (condition_mask & NECP_KERNEL_CONDITION_PLATFORM_BINARY) {
+                               num_conditions++;
+                       }
+                       if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) {
+                               num_conditions++;
+                       }
                        if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_START) {
                                if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_END) {
                                        condition_tlv_length += sizeof(struct necp_policy_condition_addr_range);
@@ -3106,6 +2666,16 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
                                condition_tlv_length += sizeof(struct necp_policy_condition_agent_type);
                                num_conditions++;
                        }
+                       if (condition_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS) {
+                               condition_tlv_length += sizeof(u_int32_t);
+                               num_conditions++;
+                       }
+                       if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_EMPTY) {
+                               num_conditions++;
+                       }
+                       if (condition_mask & NECP_KERNEL_CONDITION_REMOTE_EMPTY) {
+                               num_conditions++;
+                       }
                }
 
                condition_tlv_length += num_conditions * (sizeof(u_int8_t) + sizeof(u_int32_t)); // These are for the condition TLVs. The space for "value" is already accounted for above.
@@ -3148,6 +2718,12 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
                        if (condition_mask & NECP_KERNEL_CONDITION_ALL_INTERFACES) {
                                cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_ALL_INTERFACES, 0, "", cond_buf, condition_tlv_length);
                        }
+                       if (condition_mask & NECP_KERNEL_CONDITION_HAS_CLIENT) {
+                               cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_HAS_CLIENT, 0, "", cond_buf, condition_tlv_length);
+                       }
+                       if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) {
+                               cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_LOCAL_NETWORKS, 0, "", cond_buf, condition_tlv_length);
+                       }
                        if (condition_mask & NECP_KERNEL_CONDITION_BOUND_INTERFACE) {
                                cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_BOUND_INTERFACE, strlen(if_name) + 1,
                                    if_name, cond_buf, condition_tlv_length);
@@ -3200,6 +2776,9 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
                                cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_ENTITLEMENT, strlen(policy->cond_custom_entitlement) + 1, policy->cond_custom_entitlement,
                                    cond_buf, condition_tlv_length);
                        }
+                       if (condition_mask & NECP_KERNEL_CONDITION_PLATFORM_BINARY) {
+                               cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_PLATFORM_BINARY, 0, "", cond_buf, condition_tlv_length);
+                       }
                        if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_START) {
                                if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_END) {
                                        struct necp_policy_condition_addr_range range;
@@ -3235,6 +2814,15 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
                                    sizeof(policy->cond_agent_type), &policy->cond_agent_type,
                                    cond_buf, condition_tlv_length);
                        }
+                       if (condition_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS) {
+                               cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_CLIENT_FLAGS, sizeof(policy->cond_client_flags), &policy->cond_client_flags, cond_buf, condition_tlv_length);
+                       }
+                       if (condition_mask & NECP_KERNEL_CONDITION_LOCAL_EMPTY) {
+                               cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR_EMPTY, 0, "", cond_buf, condition_tlv_length);
+                       }
+                       if (condition_mask & NECP_KERNEL_CONDITION_REMOTE_EMPTY) {
+                               cond_buf_cursor = necp_buffer_write_tlv(cond_buf_cursor, NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR_EMPTY, 0, "", cond_buf, condition_tlv_length);
+                       }
                }
 
                cursor = necp_buffer_write_tlv(cursor, NECP_TLV_POLICY_CONDITION, cond_buf_cursor - cond_buf, cond_buf, tlv_buffer, total_allocated_bytes);
@@ -3252,33 +2840,6 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
        // UNLOCK
        lck_rw_done(&necp_kernel_policy_lock);
 
-       // Send packet
-       if (packet != NULL) {
-               u_int32_t total_result_length = sizeof(struct necp_packet_header) + total_tlv_len;
-
-               // Allow malloc to wait, since the total buffer may be large and we are not holding any locks
-               MALLOC(result_buf, u_int8_t *, total_result_length, M_NECP, M_WAITOK | M_ZERO);
-               if (result_buf == NULL) {
-                       NECPLOG(LOG_DEBUG, "Failed to allocate result_buffer (%u bytes)", total_result_length);
-                       REPORT_ERROR(NECP_ERROR_INTERNAL);
-               }
-
-               result_buf_cursor = result_buf;
-               result_buf_cursor = necp_buffer_write_packet_header(result_buf_cursor, NECP_PACKET_TYPE_POLICY_DUMP_ALL, NECP_PACKET_FLAGS_RESPONSE, message_id);
-
-               for (int i = 0; i < policy_count; i++) {
-                       if (tlv_buffer_pointers[i] != NULL) {
-                               result_buf_cursor = necp_buffer_write_tlv(result_buf_cursor, NECP_TLV_POLICY_DUMP, tlv_buffer_lengths[i], tlv_buffer_pointers[i], result_buf, total_result_length);
-                       }
-               }
-
-               if (!necp_send_ctl_data(session, result_buf, result_buf_cursor - result_buf)) {
-                       NECPLOG(LOG_ERR, "Failed to send response (%u bytes)", result_buf_cursor - result_buf);
-               } else {
-                       NECPLOG(LOG_ERR, "Sent data worth %u bytes. Total result buffer length was %u bytes", result_buf_cursor - result_buf, total_result_length);
-               }
-       }
-
        // Copy out
        if (out_buffer != 0) {
                if (out_buffer_length < total_tlv_len + sizeof(u_int32_t)) {
@@ -3315,13 +2876,6 @@ necp_handle_policy_dump_all(struct necp_session *session, u_int32_t message_id,
 done:
 
        if (error_occured) {
-               if (packet != NULL) {
-                       if (!necp_send_error_response(session, NECP_PACKET_TYPE_POLICY_DUMP_ALL, message_id, response_error)) {
-                               NECPLOG0(LOG_ERR, "Failed to send error response");
-                       } else {
-                               NECPLOG0(LOG_ERR, "Sent error response");
-                       }
-               }
                error_code = necp_get_posix_error_for_necp_error(response_error);
        }
 
@@ -3606,6 +3160,7 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
        union necp_sockaddr_union cond_remote_start;
        union necp_sockaddr_union cond_remote_end;
        u_int8_t cond_remote_prefix = 0;
+       u_int32_t cond_client_flags = 0;
        u_int32_t offset = 0;
        u_int8_t ultimate_result = 0;
        u_int32_t secondary_result = 0;
@@ -3642,6 +3197,11 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                        socket_ip_conditions = TRUE;
                        break;
                }
+               case NECP_POLICY_CONDITION_HAS_CLIENT: {
+                       master_condition_mask |= NECP_KERNEL_CONDITION_HAS_CLIENT;
+                       socket_only_conditions = TRUE;
+                       break;
+               }
                case NECP_POLICY_CONDITION_ENTITLEMENT: {
                        if (condition_length > 0) {
                                if (cond_custom_entitlement == NULL) {
@@ -3657,6 +3217,11 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                        }
                        break;
                }
+               case NECP_POLICY_CONDITION_PLATFORM_BINARY: {
+                       master_condition_mask |= NECP_KERNEL_CONDITION_PLATFORM_BINARY;
+                       socket_only_conditions = TRUE;
+                       break;
+               }
                case NECP_POLICY_CONDITION_DOMAIN: {
                        // Make sure there is only one such rule
                        if (condition_length > 0 && cond_domain == NULL) {
@@ -3781,18 +3346,29 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                        }
                        break;
                }
-               case NECP_POLICY_CONDITION_IP_PROTOCOL: {
+               case NECP_POLICY_CONDITION_IP_PROTOCOL:
+               case NECP_POLICY_CONDITION_FLOW_IP_PROTOCOL: {
                        if (condition_length >= sizeof(u_int16_t)) {
                                master_condition_mask |= NECP_KERNEL_CONDITION_PROTOCOL;
                                if (condition_is_negative) {
                                        master_condition_negated_mask |= NECP_KERNEL_CONDITION_PROTOCOL;
                                }
-                               memcpy(&cond_protocol, condition_value, sizeof(cond_protocol));
-                               socket_ip_conditions = TRUE;
+                               memcpy(&cond_protocol, condition_value, sizeof(cond_protocol));
+                               if (condition_type == NECP_POLICY_CONDITION_FLOW_IP_PROTOCOL) {
+                                       socket_only_conditions = TRUE;
+                               } else {
+                                       socket_ip_conditions = TRUE;
+                               }
                        }
                        break;
                }
-               case NECP_POLICY_CONDITION_LOCAL_ADDR: {
+               case NECP_POLICY_CONDITION_LOCAL_NETWORKS: {
+                       master_condition_mask |= NECP_KERNEL_CONDITION_LOCAL_NETWORKS;
+                       socket_ip_conditions = TRUE;
+                       break;
+               }
+               case NECP_POLICY_CONDITION_LOCAL_ADDR:
+               case NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR: {
                        struct necp_policy_condition_addr *address_struct = (struct necp_policy_condition_addr *)(void *)condition_value;
                        if (!necp_address_is_valid(&address_struct->address.sa)) {
                                break;
@@ -3806,10 +3382,15 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_LOCAL_START;
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_LOCAL_PREFIX;
                        }
-                       socket_ip_conditions = TRUE;
+                       if (condition_type == NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR) {
+                               socket_only_conditions = TRUE;
+                       } else {
+                               socket_ip_conditions = TRUE;
+                       }
                        break;
                }
-               case NECP_POLICY_CONDITION_REMOTE_ADDR: {
+               case NECP_POLICY_CONDITION_REMOTE_ADDR:
+               case NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR: {
                        struct necp_policy_condition_addr *address_struct = (struct necp_policy_condition_addr *)(void *)condition_value;
                        if (!necp_address_is_valid(&address_struct->address.sa)) {
                                break;
@@ -3823,10 +3404,15 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_REMOTE_START;
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_REMOTE_PREFIX;
                        }
-                       socket_ip_conditions = TRUE;
+                       if (condition_type == NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR) {
+                               socket_only_conditions = TRUE;
+                       } else {
+                               socket_ip_conditions = TRUE;
+                       }
                        break;
                }
-               case NECP_POLICY_CONDITION_LOCAL_ADDR_RANGE: {
+               case NECP_POLICY_CONDITION_LOCAL_ADDR_RANGE:
+               case NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR_RANGE: {
                        struct necp_policy_condition_addr_range *address_struct = (struct necp_policy_condition_addr_range *)(void *)condition_value;
                        if (!necp_address_is_valid(&address_struct->start_address.sa) ||
                            !necp_address_is_valid(&address_struct->end_address.sa)) {
@@ -3841,10 +3427,15 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_LOCAL_START;
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_LOCAL_END;
                        }
-                       socket_ip_conditions = TRUE;
+                       if (condition_type == NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR_RANGE) {
+                               socket_only_conditions = TRUE;
+                       } else {
+                               socket_ip_conditions = TRUE;
+                       }
                        break;
                }
-               case NECP_POLICY_CONDITION_REMOTE_ADDR_RANGE: {
+               case NECP_POLICY_CONDITION_REMOTE_ADDR_RANGE:
+               case NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR_RANGE: {
                        struct necp_policy_condition_addr_range *address_struct = (struct necp_policy_condition_addr_range *)(void *)condition_value;
                        if (!necp_address_is_valid(&address_struct->start_address.sa) ||
                            !necp_address_is_valid(&address_struct->end_address.sa)) {
@@ -3859,7 +3450,11 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_REMOTE_START;
                                master_condition_negated_mask |= NECP_KERNEL_CONDITION_REMOTE_END;
                        }
-                       socket_ip_conditions = TRUE;
+                       if (condition_type == NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR_RANGE) {
+                               socket_only_conditions = TRUE;
+                       } else {
+                               socket_ip_conditions = TRUE;
+                       }
                        break;
                }
                case NECP_POLICY_CONDITION_AGENT_TYPE: {
@@ -3870,6 +3465,36 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                        }
                        break;
                }
+               case NECP_POLICY_CONDITION_CLIENT_FLAGS: {
+                       if (condition_is_negative) {
+                               master_condition_negated_mask |= NECP_KERNEL_CONDITION_CLIENT_FLAGS;
+                       }
+                       master_condition_mask |= NECP_KERNEL_CONDITION_CLIENT_FLAGS;
+                       socket_only_conditions = TRUE;
+                       if (condition_length >= sizeof(u_int32_t)) {
+                               memcpy(&cond_client_flags, condition_value, sizeof(cond_client_flags));
+                       } else {
+                               // Empty means match on fallback traffic
+                               cond_client_flags = NECP_CLIENT_PARAMETER_FLAG_FALLBACK_TRAFFIC;
+                       }
+                       break;
+               }
+               case NECP_POLICY_CONDITION_FLOW_LOCAL_ADDR_EMPTY: {
+                       master_condition_mask |= NECP_KERNEL_CONDITION_LOCAL_EMPTY;
+                       if (condition_is_negative) {
+                               master_condition_negated_mask |= NECP_KERNEL_CONDITION_LOCAL_EMPTY;
+                       }
+                       socket_only_conditions = TRUE;
+                       break;
+               }
+               case NECP_POLICY_CONDITION_FLOW_REMOTE_ADDR_EMPTY: {
+                       master_condition_mask |= NECP_KERNEL_CONDITION_REMOTE_EMPTY;
+                       if (condition_is_negative) {
+                               master_condition_negated_mask |= NECP_KERNEL_CONDITION_REMOTE_EMPTY;
+                       }
+                       socket_only_conditions = TRUE;
+                       break;
+               }
                default: {
                        break;
                }
@@ -4017,13 +3642,23 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                socket_layer_non_id_conditions = TRUE;
                break;
        }
+       case NECP_POLICY_RESULT_ALLOW_UNENTITLED: {
+               socket_layer_non_id_conditions = TRUE;
+               break;
+       }
        case NECP_POLICY_RESULT_ROUTE_RULES: {
                if (policy->route_rules != NULL && policy->route_rules_size > 0) {
                        u_int32_t route_rule_id = necp_create_route_rule(&necp_route_rules, policy->route_rules, policy->route_rules_size);
                        if (route_rule_id > 0) {
                                policy->applied_route_rules_id = route_rule_id;
                                ultimate_result_parameter.route_rule_id = route_rule_id;
-                               socket_layer_non_id_conditions = TRUE;
+                               if (socket_only_conditions) { // socket_ip_conditions can be TRUE or FALSE
+                                       socket_layer_non_id_conditions = TRUE;
+                               } else if (socket_ip_conditions) {
+                                       socket_layer_non_id_conditions = TRUE;
+                                       ip_output_layer_non_id_conditions = TRUE;
+                                       ip_output_layer_non_id_only = TRUE; // Only apply route rules to packets that didn't go through socket layer
+                               }
                        }
                }
                break;
@@ -4034,7 +3669,7 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
        }
 
        if (socket_layer_non_id_conditions) {
-               necp_kernel_policy_id policy_id = necp_kernel_socket_policy_add(policy->order, session->session_order, session->proc_pid, master_condition_mask, master_condition_negated_mask, cond_app_id, cond_real_app_id, cond_custom_entitlement, cond_account_id, cond_domain, cond_pid, cond_uid, cond_bound_interface, cond_traffic_class, cond_protocol, &cond_local_start, &cond_local_end, cond_local_prefix, &cond_remote_start, &cond_remote_end, cond_remote_prefix, &cond_agent_type, ultimate_result, ultimate_result_parameter);
+               necp_kernel_policy_id policy_id = necp_kernel_socket_policy_add(policy->order, session->session_order, session->proc_pid, master_condition_mask, master_condition_negated_mask, cond_app_id, cond_real_app_id, cond_custom_entitlement, cond_account_id, cond_domain, cond_pid, cond_uid, cond_bound_interface, cond_traffic_class, cond_protocol, &cond_local_start, &cond_local_end, cond_local_prefix, &cond_remote_start, &cond_remote_end, cond_remote_prefix, &cond_agent_type, cond_client_flags, ultimate_result, ultimate_result_parameter);
 
                if (policy_id == 0) {
                        NECPLOG0(LOG_DEBUG, "Error applying socket kernel policy");
@@ -4050,6 +3685,7 @@ necp_policy_apply(struct necp_session *session, struct necp_session_policy *poli
                if (ip_output_layer_non_id_only) {
                        condition_mask |= NECP_KERNEL_CONDITION_POLICY_ID;
                }
+
                necp_kernel_policy_id policy_id = necp_kernel_ip_output_policy_add(policy->order, NECP_KERNEL_POLICY_SUBORDER_NON_ID_CONDITIONS, session->session_order, session->proc_pid, condition_mask, master_condition_negated_mask, NECP_KERNEL_POLICY_ID_NONE, cond_bound_interface, 0, cond_protocol, &cond_local_start, &cond_local_end, cond_local_prefix, &cond_remote_start, &cond_remote_end, cond_remote_prefix, ultimate_result, ultimate_result_parameter);
 
                if (policy_id == 0) {
@@ -4202,9 +3838,10 @@ necp_kernel_policy_get_new_id(bool socket_level)
        return newid;
 }
 
-#define NECP_KERNEL_VALID_SOCKET_CONDITIONS (NECP_KERNEL_CONDITION_APP_ID | NECP_KERNEL_CONDITION_REAL_APP_ID | NECP_KERNEL_CONDITION_DOMAIN | NECP_KERNEL_CONDITION_ACCOUNT_ID | NECP_KERNEL_CONDITION_PID | NECP_KERNEL_CONDITION_UID | NECP_KERNEL_CONDITION_ALL_INTERFACES | NECP_KERNEL_CONDITION_BOUND_INTERFACE | NECP_KERNEL_CONDITION_TRAFFIC_CLASS | NECP_KERNEL_CONDITION_PROTOCOL | NECP_KERNEL_CONDITION_LOCAL_START | NECP_KERNEL_CONDITION_LOCAL_END | NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_REMOTE_START | NECP_KERNEL_CONDITION_REMOTE_END | NECP_KERNEL_CONDITION_REMOTE_PREFIX | NECP_KERNEL_CONDITION_ENTITLEMENT | NECP_KERNEL_CONDITION_CUSTOM_ENTITLEMENT | NECP_KERNEL_CONDITION_AGENT_TYPE)
+#define NECP_KERNEL_VALID_SOCKET_CONDITIONS (NECP_KERNEL_CONDITION_APP_ID | NECP_KERNEL_CONDITION_REAL_APP_ID | NECP_KERNEL_CONDITION_DOMAIN | NECP_KERNEL_CONDITION_ACCOUNT_ID | NECP_KERNEL_CONDITION_PID | NECP_KERNEL_CONDITION_UID | NECP_KERNEL_CONDITION_ALL_INTERFACES | NECP_KERNEL_CONDITION_BOUND_INTERFACE | NECP_KERNEL_CONDITION_TRAFFIC_CLASS | NECP_KERNEL_CONDITION_PROTOCOL | NECP_KERNEL_CONDITION_LOCAL_START | NECP_KERNEL_CONDITION_LOCAL_END | NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_REMOTE_START | NECP_KERNEL_CONDITION_REMOTE_END | NECP_KERNEL_CONDITION_REMOTE_PREFIX | NECP_KERNEL_CONDITION_ENTITLEMENT | NECP_KERNEL_CONDITION_CUSTOM_ENTITLEMENT | NECP_KERNEL_CONDITION_AGENT_TYPE | NECP_KERNEL_CONDITION_HAS_CLIENT | NECP_KERNEL_CONDITION_LOCAL_NETWORKS | NECP_KERNEL_CONDITION_CLIENT_FLAGS | NECP_KERNEL_CONDITION_LOCAL_EMPTY | NECP_KERNEL_CONDITION_REMOTE_EMPTY | NECP_KERNEL_CONDITION_PLATFORM_BINARY)
+
 static necp_kernel_policy_id
-necp_kernel_socket_policy_add(necp_policy_order order, u_int32_t session_order, int session_pid, u_int32_t condition_mask, u_int32_t condition_negated_mask, necp_app_id cond_app_id, necp_app_id cond_real_app_id, char *cond_custom_entitlement, u_int32_t cond_account_id, char *cond_domain, pid_t cond_pid, uid_t cond_uid, ifnet_t cond_bound_interface, struct necp_policy_condition_tc_range cond_traffic_class, u_int16_t cond_protocol, union necp_sockaddr_union *cond_local_start, union necp_sockaddr_union *cond_local_end, u_int8_t cond_local_prefix, union necp_sockaddr_union *cond_remote_start, union necp_sockaddr_union *cond_remote_end, u_int8_t cond_remote_prefix, struct necp_policy_condition_agent_type *cond_agent_type, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter)
+necp_kernel_socket_policy_add(necp_policy_order order, u_int32_t session_order, int session_pid, u_int32_t condition_mask, u_int32_t condition_negated_mask, necp_app_id cond_app_id, necp_app_id cond_real_app_id, char *cond_custom_entitlement, u_int32_t cond_account_id, char *cond_domain, pid_t cond_pid, uid_t cond_uid, ifnet_t cond_bound_interface, struct necp_policy_condition_tc_range cond_traffic_class, u_int16_t cond_protocol, union necp_sockaddr_union *cond_local_start, union necp_sockaddr_union *cond_local_end, u_int8_t cond_local_prefix, union necp_sockaddr_union *cond_remote_start, union necp_sockaddr_union *cond_remote_end, u_int8_t cond_remote_prefix, struct necp_policy_condition_agent_type *cond_agent_type, u_int32_t cond_client_flags, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter)
 {
        struct necp_kernel_socket_policy *new_kernel_policy = NULL;
        struct necp_kernel_socket_policy *tmp_kernel_policy = NULL;
@@ -4237,6 +3874,12 @@ necp_kernel_socket_policy_add(necp_policy_order order, u_int32_t session_order,
        if ((new_kernel_policy->condition_mask & NECP_KERNEL_CONDITION_REMOTE_END) && (new_kernel_policy->condition_mask & NECP_KERNEL_CONDITION_REMOTE_PREFIX)) {
                new_kernel_policy->condition_mask &= ~NECP_KERNEL_CONDITION_REMOTE_PREFIX;
        }
+       if (new_kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_EMPTY) {
+               new_kernel_policy->condition_mask &= ~(NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_LOCAL_END);
+       }
+       if ((new_kernel_policy->condition_mask & NECP_KERNEL_CONDITION_REMOTE_EMPTY)) {
+               new_kernel_policy->condition_mask &= ~(NECP_KERNEL_CONDITION_REMOTE_PREFIX | NECP_KERNEL_CONDITION_REMOTE_END);
+       }
        new_kernel_policy->condition_negated_mask = condition_negated_mask & new_kernel_policy->condition_mask;
 
        // Set condition values
@@ -4296,6 +3939,9 @@ necp_kernel_socket_policy_add(necp_policy_order order, u_int32_t session_order,
        if (new_kernel_policy->condition_mask & NECP_KERNEL_CONDITION_AGENT_TYPE) {
                memcpy(&new_kernel_policy->cond_agent_type, cond_agent_type, sizeof(*cond_agent_type));
        }
+       if (new_kernel_policy->condition_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS) {
+               new_kernel_policy->cond_client_flags = cond_client_flags;
+       }
 
        new_kernel_policy->result = result;
        memcpy(&new_kernel_policy->result_parameter, &result_parameter, sizeof(result_parameter));
@@ -4407,9 +4053,13 @@ necp_get_result_description(char *result_string, necp_kernel_policy_result resul
                snprintf(result_string, MAX_RESULT_STRING_LEN, "ScopedDirect");
                break;
        }
+       case NECP_KERNEL_POLICY_RESULT_ALLOW_UNENTITLED: {
+               snprintf(result_string, MAX_RESULT_STRING_LEN, "AllowUnentitled");
+               break;
+       }
        case NECP_KERNEL_POLICY_RESULT_ROUTE_RULES: {
                int index = 0;
-               char interface_names[IFXNAMSIZ][MAX_ROUTE_RULE_INTERFACES];
+               char interface_names[MAX_ROUTE_RULE_INTERFACES][IFXNAMSIZ];
                struct necp_route_rule *route_rule = necp_lookup_route_rule_locked(&necp_route_rules, result_parameter.route_rule_id);
                if (route_rule != NULL) {
                        for (index = 0; index < MAX_ROUTE_RULE_INTERFACES; index++) {
@@ -4422,11 +4072,12 @@ necp_get_result_description(char *result_string, necp_kernel_policy_result resul
                        }
                        switch (route_rule->default_action) {
                        case NECP_ROUTE_RULE_DENY_INTERFACE:
-                               snprintf(result_string, MAX_RESULT_STRING_LEN, "RouteRules (Only %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s)",
+                               snprintf(result_string, MAX_RESULT_STRING_LEN, "RouteRules (Only %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s)",
                                    (route_rule->cellular_action == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? "Cell " : "",
                                    (route_rule->wifi_action == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? "WiFi " : "",
                                    (route_rule->wired_action == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? "Wired " : "",
                                    (route_rule->expensive_action == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? "Exp " : "",
+                                   (route_rule->constrained_action == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? "Constrained " : "",
                                    (route_rule->exception_if_actions[0] == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? interface_names[0] : "",
                                    (route_rule->exception_if_actions[0] == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? " " : "",
                                    (route_rule->exception_if_actions[1] == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? interface_names[1] : "",
@@ -4448,11 +4099,12 @@ necp_get_result_description(char *result_string, necp_kernel_policy_result resul
                                    (route_rule->exception_if_actions[9] == NECP_ROUTE_RULE_ALLOW_INTERFACE) ? interface_names[9] : "");
                                break;
                        case NECP_ROUTE_RULE_ALLOW_INTERFACE:
-                               snprintf(result_string, MAX_RESULT_STRING_LEN, "RouteRules (%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s)",
+                               snprintf(result_string, MAX_RESULT_STRING_LEN, "RouteRules (%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s)",
                                    (route_rule->cellular_action == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!Cell " : "",
                                    (route_rule->wifi_action == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!WiFi " : "",
                                    (route_rule->wired_action == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!Wired " : "",
                                    (route_rule->expensive_action == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!Exp " : "",
+                                   (route_rule->constrained_action == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!Constrained " : "",
                                    (route_rule->exception_if_actions[0] == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!" : "",
                                    (route_rule->exception_if_actions[0] == NECP_ROUTE_RULE_DENY_INTERFACE) ? interface_names[0] : "",
                                    (route_rule->exception_if_actions[1] == NECP_ROUTE_RULE_DENY_INTERFACE) ? "!" : "",
@@ -4475,11 +4127,12 @@ necp_get_result_description(char *result_string, necp_kernel_policy_result resul
                                    (route_rule->exception_if_actions[9] == NECP_ROUTE_RULE_DENY_INTERFACE) ? interface_names[9] : "");
                                break;
                        case NECP_ROUTE_RULE_QOS_MARKING:
-                               snprintf(result_string, MAX_RESULT_STRING_LEN, "RouteRules (QoSMarking %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s)",
+                               snprintf(result_string, MAX_RESULT_STRING_LEN, "RouteRules (QoSMarking %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s)",
                                    (route_rule->cellular_action == NECP_ROUTE_RULE_QOS_MARKING) ? "Cell " : "",
                                    (route_rule->wifi_action == NECP_ROUTE_RULE_QOS_MARKING) ? "WiFi " : "",
                                    (route_rule->wired_action == NECP_ROUTE_RULE_QOS_MARKING) ? "Wired " : "",
                                    (route_rule->expensive_action == NECP_ROUTE_RULE_QOS_MARKING) ? "Exp " : "",
+                                   (route_rule->constrained_action == NECP_ROUTE_RULE_QOS_MARKING) ? "Constrained " : "",
                                    (route_rule->exception_if_actions[0] == NECP_ROUTE_RULE_QOS_MARKING) ? interface_names[0] : "",
                                    (route_rule->exception_if_actions[0] == NECP_ROUTE_RULE_QOS_MARKING) ? " " : "",
                                    (route_rule->exception_if_actions[1] == NECP_ROUTE_RULE_QOS_MARKING) ? interface_names[1] : "",
@@ -4627,7 +4280,8 @@ necp_kernel_socket_policy_results_overlap(struct necp_kernel_socket_policy *uppe
        } else if (upper_policy->result == NECP_KERNEL_POLICY_RESULT_SOCKET_FILTER ||
            upper_policy->result == NECP_KERNEL_POLICY_RESULT_ROUTE_RULES ||
            upper_policy->result == NECP_KERNEL_POLICY_RESULT_USE_NETAGENT ||
-           upper_policy->result == NECP_KERNEL_POLICY_RESULT_NETAGENT_SCOPED) {
+           upper_policy->result == NECP_KERNEL_POLICY_RESULT_NETAGENT_SCOPED ||
+           upper_policy->result == NECP_KERNEL_POLICY_RESULT_ALLOW_UNENTITLED) {
                // Filters and route rules never cancel out lower policies
                return FALSE;
        } else if (necp_kernel_socket_result_is_trigger_service_type(upper_policy)) {
@@ -4699,6 +4353,11 @@ necp_kernel_socket_policy_is_unnecessary(struct necp_kernel_socket_policy *polic
                        continue;
                }
 
+               // If new policy matches Local Networks, compared policy must also
+               if ((policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) && !(compared_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS)) {
+                       continue;
+               }
+
                // Default makes lower policies unecessary always
                if (compared_policy->condition_mask == 0) {
                        return TRUE;
@@ -4764,6 +4423,11 @@ necp_kernel_socket_policy_is_unnecessary(struct necp_kernel_socket_policy *polic
                        continue;
                }
 
+               if (compared_policy->condition_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS &&
+                   compared_policy->cond_client_flags != policy->cond_client_flags) {
+                       continue;
+               }
+
                if (compared_policy->condition_mask & NECP_KERNEL_CONDITION_TRAFFIC_CLASS &&
                    !(compared_policy->cond_traffic_class.start_tc <= policy->cond_traffic_class.start_tc &&
                    compared_policy->cond_traffic_class.end_tc >= policy->cond_traffic_class.end_tc)) {
@@ -5020,7 +4684,7 @@ necp_create_string_to_id_mapping(struct necp_string_id_mapping_list *list, char
        existing_mapping = necp_lookup_string_to_id_locked(list, string);
        if (existing_mapping != NULL) {
                string_id = existing_mapping->id;
-               existing_mapping->refcount++;
+               os_ref_retain_locked(&existing_mapping->refcount);
        } else {
                struct necp_string_id_mapping *new_mapping = NULL;
                MALLOC(new_mapping, struct necp_string_id_mapping *, sizeof(struct necp_string_id_mapping), M_NECP, M_WAITOK);
@@ -5032,7 +4696,7 @@ necp_create_string_to_id_mapping(struct necp_string_id_mapping_list *list, char
                        if (new_mapping->string != NULL) {
                                memcpy(new_mapping->string, string, length);
                                new_mapping->id = necp_get_new_string_id();
-                               new_mapping->refcount = 1;
+                               os_ref_init(&new_mapping->refcount, &necp_refgrp);
                                LIST_INSERT_HEAD(list, new_mapping, chain);
                                string_id = new_mapping->id;
                        } else {
@@ -5053,7 +4717,7 @@ necp_remove_string_to_id_mapping(struct necp_string_id_mapping_list *list, char
 
        existing_mapping = necp_lookup_string_to_id_locked(list, string);
        if (existing_mapping != NULL) {
-               if (--existing_mapping->refcount == 0) {
+               if (os_ref_release_locked(&existing_mapping->refcount) == 0) {
                        LIST_REMOVE(existing_mapping, chain);
                        FREE(existing_mapping->string, M_NECP);
                        FREE(existing_mapping, M_NECP);
@@ -5138,7 +4802,7 @@ necp_lookup_route_rule_locked(struct necp_route_rule_list *list, u_int32_t route
 }
 
 static struct necp_route_rule *
-necp_lookup_route_rule_by_contents_locked(struct necp_route_rule_list *list, u_int32_t default_action, u_int8_t cellular_action, u_int8_t wifi_action, u_int8_t wired_action, u_int8_t expensive_action, u_int32_t *if_indices, u_int8_t *if_actions)
+necp_lookup_route_rule_by_contents_locked(struct necp_route_rule_list *list, u_int32_t default_action, u_int8_t cellular_action, u_int8_t wifi_action, u_int8_t wired_action, u_int8_t expensive_action, u_int8_t constrained_action, u_int32_t *if_indices, u_int8_t *if_actions)
 {
        struct necp_route_rule *searchentry = NULL;
        struct necp_route_rule *foundentry = NULL;
@@ -5148,7 +4812,8 @@ necp_lookup_route_rule_by_contents_locked(struct necp_route_rule_list *list, u_i
                    searchentry->cellular_action == cellular_action &&
                    searchentry->wifi_action == wifi_action &&
                    searchentry->wired_action == wired_action &&
-                   searchentry->expensive_action == expensive_action) {
+                   searchentry->expensive_action == expensive_action &&
+                   searchentry->constrained_action == constrained_action) {
                        bool match_failed = FALSE;
                        size_t index_a = 0;
                        size_t index_b = 0;
@@ -5199,6 +4864,7 @@ necp_create_route_rule(struct necp_route_rule_list *list, u_int8_t *route_rules_
        u_int8_t wifi_action = NECP_ROUTE_RULE_NONE;
        u_int8_t wired_action = NECP_ROUTE_RULE_NONE;
        u_int8_t expensive_action = NECP_ROUTE_RULE_NONE;
+       u_int8_t constrained_action = NECP_ROUTE_RULE_NONE;
        u_int32_t if_indices[MAX_ROUTE_RULE_INTERFACES];
        size_t num_valid_indices = 0;
        memset(&if_indices, 0, sizeof(if_indices));
@@ -5241,6 +4907,9 @@ necp_create_route_rule(struct necp_route_rule_list *list, u_int8_t *route_rules_
                        if (rule_flags & NECP_ROUTE_RULE_FLAG_EXPENSIVE) {
                                expensive_action = rule_type;
                        }
+                       if (rule_flags & NECP_ROUTE_RULE_FLAG_CONSTRAINED) {
+                               constrained_action = rule_type;
+                       }
                        if (rule_flags == 0) {
                                default_action = rule_type;
                        }
@@ -5265,10 +4934,10 @@ necp_create_route_rule(struct necp_route_rule_list *list, u_int8_t *route_rules_
                offset += sizeof(u_int8_t) + sizeof(u_int32_t) + length;
        }
 
-       existing_rule = necp_lookup_route_rule_by_contents_locked(list, default_action, cellular_action, wifi_action, wired_action, expensive_action, if_indices, if_actions);
+       existing_rule = necp_lookup_route_rule_by_contents_locked(list, default_action, cellular_action, wifi_action, wired_action, expensive_action, constrained_action, if_indices, if_actions);
        if (existing_rule != NULL) {
                route_rule_id = existing_rule->id;
-               existing_rule->refcount++;
+               os_ref_retain_locked(&existing_rule->refcount);
        } else {
                struct necp_route_rule *new_rule = NULL;
                MALLOC(new_rule, struct necp_route_rule *, sizeof(struct necp_route_rule), M_NECP, M_WAITOK);
@@ -5280,9 +4949,10 @@ necp_create_route_rule(struct necp_route_rule_list *list, u_int8_t *route_rules_
                        new_rule->wifi_action = wifi_action;
                        new_rule->wired_action = wired_action;
                        new_rule->expensive_action = expensive_action;
+                       new_rule->constrained_action =  constrained_action;
                        memcpy(&new_rule->exception_if_indices, &if_indices, sizeof(if_indices));
                        memcpy(&new_rule->exception_if_actions, &if_actions, sizeof(if_actions));
-                       new_rule->refcount = 1;
+                       os_ref_init(&new_rule->refcount, &necp_refgrp);
                        LIST_INSERT_HEAD(list, new_rule, chain);
                }
        }
@@ -5323,7 +4993,7 @@ necp_remove_route_rule(struct necp_route_rule_list *list, u_int32_t route_rule_i
 
        existing_rule = necp_lookup_route_rule_locked(list, route_rule_id);
        if (existing_rule != NULL) {
-               if (--existing_rule->refcount == 0) {
+               if (os_ref_release_locked(&existing_rule->refcount) == 0) {
                        necp_remove_aggregate_route_rule_for_id(existing_rule->id);
                        LIST_REMOVE(existing_rule, chain);
                        FREE(existing_rule, M_NECP);
@@ -5361,16 +5031,10 @@ necp_create_aggregate_route_rule(u_int32_t *rule_ids)
        struct necp_aggregate_route_rule *new_rule = NULL;
        struct necp_aggregate_route_rule *existing_rule = NULL;
 
-       LIST_FOREACH(existing_rule, &necp_aggregate_route_rules, chain) {
-               if (memcmp(existing_rule->rule_ids, rule_ids, (sizeof(u_int32_t) * MAX_AGGREGATE_ROUTE_RULES)) == 0) {
-                       return existing_rule->id;
-               }
-       }
-
        lck_rw_lock_exclusive(&necp_route_rule_lock);
 
+       // Check if the rule already exists
        LIST_FOREACH(existing_rule, &necp_aggregate_route_rules, chain) {
-               // Re-check, in case something else created the rule while we are waiting to lock
                if (memcmp(existing_rule->rule_ids, rule_ids, (sizeof(u_int32_t) * MAX_AGGREGATE_ROUTE_RULES)) == 0) {
                        lck_rw_done(&necp_route_rule_lock);
                        return existing_rule->id;
@@ -5494,9 +5158,9 @@ necp_create_uuid_app_id_mapping(uuid_t uuid, bool *allocated_mapping, bool uuid_
        existing_mapping = necp_uuid_lookup_app_id_locked(uuid);
        if (existing_mapping != NULL) {
                local_id = existing_mapping->id;
-               existing_mapping->refcount++;
+               os_ref_retain_locked(&existing_mapping->refcount);
                if (uuid_policy_table) {
-                       existing_mapping->table_refcount++;
+                       existing_mapping->table_usecount++;
                }
        } else {
                struct necp_uuid_id_mapping *new_mapping = NULL;
@@ -5504,11 +5168,11 @@ necp_create_uuid_app_id_mapping(uuid_t uuid, bool *allocated_mapping, bool uuid_
                if (new_mapping != NULL) {
                        uuid_copy(new_mapping->uuid, uuid);
                        new_mapping->id = necp_get_new_uuid_id(false);
-                       new_mapping->refcount = 1;
+                       os_ref_init(&new_mapping->refcount, &necp_refgrp);
                        if (uuid_policy_table) {
-                               new_mapping->table_refcount = 1;
+                               new_mapping->table_usecount = 1;
                        } else {
-                               new_mapping->table_refcount = 0;
+                               new_mapping->table_usecount = 0;
                        }
 
                        LIST_INSERT_HEAD(APPUUIDHASH(uuid), new_mapping, chain);
@@ -5538,9 +5202,9 @@ necp_remove_uuid_app_id_mapping(uuid_t uuid, bool *removed_mapping, bool uuid_po
        existing_mapping = necp_uuid_lookup_app_id_locked(uuid);
        if (existing_mapping != NULL) {
                if (uuid_policy_table) {
-                       existing_mapping->table_refcount--;
+                       existing_mapping->table_usecount--;
                }
-               if (--existing_mapping->refcount == 0) {
+               if (os_ref_release_locked(&existing_mapping->refcount) == 0) {
                        LIST_REMOVE(existing_mapping, chain);
                        FREE(existing_mapping, M_NECP);
                        if (removed_mapping) {
@@ -5618,14 +5282,14 @@ necp_create_uuid_service_id_mapping(uuid_t uuid)
        existing_mapping = necp_uuid_lookup_service_id_locked(uuid);
        if (existing_mapping != NULL) {
                local_id = existing_mapping->id;
-               existing_mapping->refcount++;
+               os_ref_retain_locked(&existing_mapping->refcount);
        } else {
                struct necp_uuid_id_mapping *new_mapping = NULL;
                MALLOC(new_mapping, struct necp_uuid_id_mapping *, sizeof(*new_mapping), M_NECP, M_WAITOK);
                if (new_mapping != NULL) {
                        uuid_copy(new_mapping->uuid, uuid);
                        new_mapping->id = necp_get_new_uuid_id(true);
-                       new_mapping->refcount = 1;
+                       os_ref_init(&new_mapping->refcount, &necp_refgrp);
 
                        LIST_INSERT_HEAD(&necp_uuid_service_id_list, new_mapping, chain);
 
@@ -5647,9 +5311,9 @@ necp_remove_uuid_service_id_mapping(uuid_t uuid)
 
        LCK_RW_ASSERT(&necp_kernel_policy_lock, LCK_RW_ASSERT_EXCLUSIVE);
 
-       existing_mapping = necp_uuid_lookup_app_id_locked(uuid);
+       existing_mapping = necp_uuid_lookup_service_id_locked(uuid);
        if (existing_mapping != NULL) {
-               if (--existing_mapping->refcount == 0) {
+               if (os_ref_release_locked(&existing_mapping->refcount) == 0) {
                        LIST_REMOVE(existing_mapping, chain);
                        FREE(existing_mapping, M_NECP);
                }
@@ -5676,7 +5340,7 @@ necp_kernel_socket_policies_update_uuid_table(void)
                        for (uuid_list_head = &necp_uuid_app_id_hashtbl[necp_uuid_app_id_hash_num_buckets - 1]; uuid_list_head >= necp_uuid_app_id_hashtbl; uuid_list_head--) {
                                struct necp_uuid_id_mapping *mapping = NULL;
                                LIST_FOREACH(mapping, uuid_list_head, chain) {
-                                       if (mapping->table_refcount > 0 &&
+                                       if (mapping->table_usecount > 0 &&
                                            proc_uuid_policy_kernel(PROC_UUID_POLICY_OPERATION_ADD, mapping->uuid, PROC_UUID_NECP_APP_POLICY) < 0) {
                                                NECPLOG0(LOG_DEBUG, "Error adding uuid to policy table\n");
                                        }
@@ -5690,7 +5354,7 @@ necp_kernel_socket_policies_update_uuid_table(void)
        return TRUE;
 }
 
-#define NECP_KERNEL_VALID_IP_OUTPUT_CONDITIONS (NECP_KERNEL_CONDITION_ALL_INTERFACES | NECP_KERNEL_CONDITION_BOUND_INTERFACE | NECP_KERNEL_CONDITION_PROTOCOL | NECP_KERNEL_CONDITION_LOCAL_START | NECP_KERNEL_CONDITION_LOCAL_END | NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_REMOTE_START | NECP_KERNEL_CONDITION_REMOTE_END | NECP_KERNEL_CONDITION_REMOTE_PREFIX | NECP_KERNEL_CONDITION_POLICY_ID | NECP_KERNEL_CONDITION_LAST_INTERFACE)
+#define NECP_KERNEL_VALID_IP_OUTPUT_CONDITIONS (NECP_KERNEL_CONDITION_ALL_INTERFACES | NECP_KERNEL_CONDITION_BOUND_INTERFACE | NECP_KERNEL_CONDITION_PROTOCOL | NECP_KERNEL_CONDITION_LOCAL_START | NECP_KERNEL_CONDITION_LOCAL_END | NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_REMOTE_START | NECP_KERNEL_CONDITION_REMOTE_END | NECP_KERNEL_CONDITION_REMOTE_PREFIX | NECP_KERNEL_CONDITION_POLICY_ID | NECP_KERNEL_CONDITION_LAST_INTERFACE | NECP_KERNEL_CONDITION_LOCAL_NETWORKS)
 static necp_kernel_policy_id
 necp_kernel_ip_output_policy_add(necp_policy_order order, necp_policy_order suborder, u_int32_t session_order, int session_pid, u_int32_t condition_mask, u_int32_t condition_negated_mask, necp_kernel_policy_id cond_policy_id, ifnet_t cond_bound_interface, u_int32_t cond_last_interface_index, u_int16_t cond_protocol, union necp_sockaddr_union *cond_local_start, union necp_sockaddr_union *cond_local_end, u_int8_t cond_local_prefix, union necp_sockaddr_union *cond_remote_start, union necp_sockaddr_union *cond_remote_end, u_int8_t cond_remote_prefix, necp_kernel_policy_result result, necp_kernel_policy_result_parameter result_parameter)
 {
@@ -5905,6 +5569,11 @@ necp_kernel_ip_output_policy_is_unnecessary(struct necp_kernel_ip_output_policy
                        continue;
                }
 
+               // If new policy matches Local Networks, compared policy must also
+               if ((policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) && !(compared_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS)) {
+                       continue;
+               }
+
                // Default makes lower policies unecessary always
                if (compared_policy->condition_mask == 0) {
                        return TRUE;
@@ -5999,8 +5668,10 @@ necp_kernel_ip_output_policies_reprocess(void)
 
                /* Update bucket counts:
                 * Non-id and SKIP policies will be added to all buckets
+                * Add local networks policy to all buckets for incoming IP
                 */
                if (!(kernel_policy->condition_mask & NECP_KERNEL_CONDITION_POLICY_ID) ||
+                   (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) ||
                    kernel_policy->result == NECP_KERNEL_POLICY_RESULT_SKIP) {
                        for (i = 0; i < NECP_KERNEL_IP_OUTPUT_POLICIES_MAP_NUM_ID_BUCKETS; i++) {
                                bucket_allocation_counts[i]++;
@@ -6030,6 +5701,7 @@ necp_kernel_ip_output_policies_reprocess(void)
        LIST_FOREACH(kernel_policy, &necp_kernel_ip_output_policies, chain) {
                // Insert pointers into map
                if (!(kernel_policy->condition_mask & NECP_KERNEL_CONDITION_POLICY_ID) ||
+                   (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) ||
                    kernel_policy->result == NECP_KERNEL_POLICY_RESULT_SKIP) {
                        for (i = 0; i < NECP_KERNEL_IP_OUTPUT_POLICIES_MAP_NUM_ID_BUCKETS; i++) {
                                if (!necp_kernel_ip_output_policy_is_unnecessary(kernel_policy, necp_kernel_ip_output_policies_map[i], bucket_current_free_index[i])) {
@@ -6167,11 +5839,35 @@ necp_hostname_matches_domain(struct substring hostname_substring, u_int8_t hostn
        return FALSE;
 }
 
+bool
+net_domain_contains_hostname(char *hostname_string, char *domain_string)
+{
+       if (hostname_string == NULL ||
+           domain_string == NULL) {
+               return false;
+       }
+
+       struct substring hostname_substring;
+       hostname_substring.string = hostname_string;
+       hostname_substring.length = strlen(hostname_string);
+
+       return necp_hostname_matches_domain(hostname_substring,
+                  necp_count_dots(hostname_string, hostname_substring.length),
+                  domain_string,
+                  necp_count_dots(domain_string, strlen(domain_string)));
+}
+
+#define NECP_MAX_STRING_LEN 1024
+
 static char *
 necp_copy_string(char *string, size_t length)
 {
        char *copied_string = NULL;
 
+       if (length > NECP_MAX_STRING_LEN) {
+               return NULL;
+       }
+
        MALLOC(copied_string, char *, length + 1, M_NECP, M_WAITOK);
        if (copied_string == NULL) {
                return NULL;
@@ -6208,34 +5904,31 @@ static inline void
 necp_get_parent_cred_result(proc_t proc, struct necp_socket_info *info)
 {
        task_t task = proc_task(proc ? proc : current_proc());
-       coalition_t coal = COALITION_NULL;
-       Boolean is_leader = coalition_is_leader(task, COALITION_TYPE_JETSAM, &coal);
+       coalition_t coal = task_get_coalition(task, COALITION_TYPE_JETSAM);
 
-       if (is_leader == TRUE) {
+       if (coal == COALITION_NULL || coalition_is_leader(task, coal)) {
                // No parent, nothing to do
                return;
        }
 
-       if (coal != NULL) {
-               task_t lead_task = coalition_get_leader(coal);
-               if (lead_task != NULL) {
-                       proc_t lead_proc = get_bsdtask_info(lead_task);
-                       if (lead_proc != NULL) {
-                               kauth_cred_t lead_cred = kauth_cred_proc_ref(lead_proc);
-                               if (lead_cred != NULL) {
-                                       errno_t cred_result = priv_check_cred(lead_cred, PRIV_NET_PRIVILEGED_NECP_MATCH, 0);
-                                       kauth_cred_unref(&lead_cred);
-                                       info->cred_result = cred_result;
-                               }
+       task_t lead_task = coalition_get_leader(coal);
+       if (lead_task != NULL) {
+               proc_t lead_proc = get_bsdtask_info(lead_task);
+               if (lead_proc != NULL) {
+                       kauth_cred_t lead_cred = kauth_cred_proc_ref(lead_proc);
+                       if (lead_cred != NULL) {
+                               errno_t cred_result = priv_check_cred(lead_cred, PRIV_NET_PRIVILEGED_NECP_MATCH, 0);
+                               kauth_cred_unref(&lead_cred);
+                               info->cred_result = cred_result;
                        }
-                       task_deallocate(lead_task);
                }
+               task_deallocate(lead_task);
        }
 }
 
-#define NECP_KERNEL_ADDRESS_TYPE_CONDITIONS (NECP_KERNEL_CONDITION_LOCAL_START | NECP_KERNEL_CONDITION_LOCAL_END | NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_REMOTE_START | NECP_KERNEL_CONDITION_REMOTE_END | NECP_KERNEL_CONDITION_REMOTE_PREFIX)
+#define NECP_KERNEL_ADDRESS_TYPE_CONDITIONS (NECP_KERNEL_CONDITION_LOCAL_START | NECP_KERNEL_CONDITION_LOCAL_END | NECP_KERNEL_CONDITION_LOCAL_PREFIX | NECP_KERNEL_CONDITION_REMOTE_START | NECP_KERNEL_CONDITION_REMOTE_END | NECP_KERNEL_CONDITION_REMOTE_PREFIX | NECP_KERNEL_CONDITION_LOCAL_EMPTY | NECP_KERNEL_CONDITION_REMOTE_EMPTY | NECP_KERNEL_CONDITION_LOCAL_NETWORKS)
 static void
-necp_application_fillout_info_locked(uuid_t application_uuid, uuid_t real_application_uuid, char *account, char *domain, pid_t pid, uid_t uid, u_int16_t protocol, u_int32_t bound_interface_index, u_int32_t traffic_class, union necp_sockaddr_union *local_addr, union necp_sockaddr_union *remote_addr, proc_t proc, struct necp_socket_info *info)
+necp_application_fillout_info_locked(uuid_t application_uuid, uuid_t real_application_uuid, char *account, char *domain, pid_t pid, uid_t uid, u_int16_t protocol, u_int32_t bound_interface_index, u_int32_t traffic_class, union necp_sockaddr_union *local_addr, union necp_sockaddr_union *remote_addr, u_int16_t local_port, u_int16_t remote_port, bool has_client, proc_t proc, u_int32_t drop_order, u_int32_t client_flags, struct necp_socket_info *info)
 {
        memset(info, 0, sizeof(struct necp_socket_info));
 
@@ -6244,6 +5937,9 @@ necp_application_fillout_info_locked(uuid_t application_uuid, uuid_t real_applic
        info->protocol = protocol;
        info->bound_interface_index = bound_interface_index;
        info->traffic_class = traffic_class;
+       info->has_client = has_client;
+       info->drop_order = drop_order;
+       info->client_flags = client_flags;
 
        if (necp_kernel_application_policies_condition_mask & NECP_KERNEL_CONDITION_ENTITLEMENT && proc != NULL) {
                info->cred_result = priv_check_cred(proc_ucred(proc), PRIV_NET_PRIVILEGED_NECP_MATCH, 0);
@@ -6253,6 +5949,10 @@ necp_application_fillout_info_locked(uuid_t application_uuid, uuid_t real_applic
                }
        }
 
+       if (necp_kernel_application_policies_condition_mask & NECP_KERNEL_CONDITION_PLATFORM_BINARY && proc != NULL) {
+               info->is_platform_binary = csproc_get_platform_binary(proc) ? true : false;
+       }
+
        if (necp_kernel_application_policies_condition_mask & NECP_KERNEL_CONDITION_APP_ID && !uuid_is_null(application_uuid)) {
                struct necp_uuid_id_mapping *existing_mapping = necp_uuid_lookup_app_id_locked(application_uuid);
                if (existing_mapping) {
@@ -6285,9 +5985,23 @@ necp_application_fillout_info_locked(uuid_t application_uuid, uuid_t real_applic
        if (necp_kernel_application_policies_condition_mask & NECP_KERNEL_ADDRESS_TYPE_CONDITIONS) {
                if (local_addr && local_addr->sa.sa_len > 0) {
                        memcpy(&info->local_addr, local_addr, local_addr->sa.sa_len);
+                       if (local_port != 0) {
+                               info->local_addr.sin6.sin6_port = local_port;
+                       }
+               } else if (local_port != 0) {
+                       info->local_addr.sin6.sin6_len = sizeof(struct sockaddr_in6);
+                       info->local_addr.sin6.sin6_family = AF_INET6;
+                       info->local_addr.sin6.sin6_port = local_port;
                }
                if (remote_addr && remote_addr->sa.sa_len > 0) {
                        memcpy(&info->remote_addr, remote_addr, remote_addr->sa.sa_len);
+                       if (remote_port != 0) {
+                               info->remote_addr.sin6.sin6_port = remote_port;
+                       }
+               } else if (remote_port != 0) {
+                       info->remote_addr.sin6.sin6_len = sizeof(struct sockaddr_in6);
+                       info->remote_addr.sin6.sin6_family = AF_INET6;
+                       info->remote_addr.sin6.sin6_port = remote_port;
                }
        }
 }
@@ -6325,10 +6039,14 @@ necp_application_find_policy_match_internal(proc_t proc,
     u_int32_t parameters_size,
     struct necp_aggregate_result *returned_result,
     u_int32_t *flags,
+    u_int32_t *reason,
     u_int required_interface_index,
     const union necp_sockaddr_union *override_local_addr,
     const union necp_sockaddr_union *override_remote_addr,
-    struct rtentry **returned_route, bool ignore_address)
+    struct necp_client_endpoint *returned_v4_gateway,
+    struct necp_client_endpoint *returned_v6_gateway,
+    struct rtentry **returned_route, bool ignore_address,
+    bool has_client)
 {
        int error = 0;
        size_t offset = 0;
@@ -6336,7 +6054,6 @@ necp_application_find_policy_match_internal(proc_t proc,
        struct necp_kernel_socket_policy *matched_policy = NULL;
        struct necp_socket_info info;
        necp_kernel_policy_filter filter_control_unit = 0;
-       u_int32_t route_rule_id = 0;
        necp_kernel_policy_result service_action = 0;
        necp_kernel_policy_service service = { 0, 0 };
 
@@ -6349,6 +6066,9 @@ necp_application_find_policy_match_internal(proc_t proc,
        bool no_remote_addr = FALSE;
        u_int8_t remote_family = 0;
        bool no_local_addr = FALSE;
+       u_int16_t local_port = 0;
+       u_int16_t remote_port = 0;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
 
        if (override_local_addr) {
                memcpy(&local_addr, override_local_addr, sizeof(local_addr));
@@ -6392,11 +6112,23 @@ necp_application_find_policy_match_internal(proc_t proc,
                return EINVAL;
        }
 
+       if (returned_v4_gateway != NULL) {
+               memset(returned_v4_gateway, 0, sizeof(struct necp_client_endpoint));
+       }
+
+       if (returned_v6_gateway != NULL) {
+               memset(returned_v6_gateway, 0, sizeof(struct necp_client_endpoint));
+       }
+
        memset(returned_result, 0, sizeof(struct necp_aggregate_result));
 
+       u_int32_t drop_order = necp_process_drop_order(proc_ucred(proc));
+
+       necp_kernel_policy_result drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+
        lck_rw_lock_shared(&necp_kernel_policy_lock);
        if (necp_kernel_application_policies_count == 0) {
-               if (necp_drop_all_order > 0) {
+               if (necp_drop_all_order > 0 || drop_order > 0) {
                        returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
                        lck_rw_done(&necp_kernel_policy_lock);
                        return 0;
@@ -6489,6 +6221,8 @@ necp_application_find_policy_match_internal(proc_t proc,
                                case NECP_CLIENT_PARAMETER_IP_PROTOCOL: {
                                        if (length >= sizeof(u_int16_t)) {
                                                memcpy(&protocol, value, sizeof(u_int16_t));
+                                       } else if (length >= sizeof(u_int8_t)) {
+                                               memcpy(&protocol, value, sizeof(u_int8_t));
                                        }
                                        break;
                                }
@@ -6506,7 +6240,7 @@ necp_application_find_policy_match_internal(proc_t proc,
                                        break;
                                }
                                case NECP_CLIENT_PARAMETER_LOCAL_ADDRESS: {
-                                       if (ignore_address) {
+                                       if (ignore_address || override_local_addr) {
                                                break;
                                        }
 
@@ -6519,7 +6253,7 @@ necp_application_find_policy_match_internal(proc_t proc,
                                        break;
                                }
                                case NECP_CLIENT_PARAMETER_REMOTE_ADDRESS: {
-                                       if (ignore_address) {
+                                       if (ignore_address || override_remote_addr) {
                                                break;
                                        }
 
@@ -6531,13 +6265,44 @@ necp_application_find_policy_match_internal(proc_t proc,
                                        }
                                        break;
                                }
+                               case NECP_CLIENT_PARAMETER_LOCAL_ENDPOINT: {
+                                       if (ignore_address || override_local_addr) {
+                                               break;
+                                       }
+
+                                       if (length >= sizeof(struct necp_client_endpoint)) {
+                                               struct necp_client_endpoint *endpoint = (struct necp_client_endpoint *)(void *)value;
+                                               if (endpoint->u.endpoint.endpoint_family == AF_UNSPEC &&
+                                                   endpoint->u.endpoint.endpoint_port != 0) {
+                                                       // Save port
+                                                       local_port = endpoint->u.endpoint.endpoint_port;
+                                               }
+                                       }
+                                       break;
+                               }
+                               case NECP_CLIENT_PARAMETER_REMOTE_ENDPOINT: {
+                                       if (ignore_address || override_remote_addr) {
+                                               break;
+                                       }
+
+                                       if (length >= sizeof(struct necp_client_endpoint)) {
+                                               struct necp_client_endpoint *endpoint = (struct necp_client_endpoint *)(void *)value;
+                                               if (endpoint->u.endpoint.endpoint_family == AF_UNSPEC &&
+                                                   endpoint->u.endpoint.endpoint_port != 0) {
+                                                       // Save port
+                                                       remote_port = endpoint->u.endpoint.endpoint_port;
+                                               }
+                                       }
+                                       break;
+                               }
                                case NECP_CLIENT_PARAMETER_FLAGS: {
                                        if (length >= sizeof(client_flags)) {
                                                memcpy(&client_flags, value, sizeof(client_flags));
                                        }
                                        break;
                                }
-                               case NECP_CLIENT_PARAMETER_REQUIRE_AGENT_TYPE: {
+                               case NECP_CLIENT_PARAMETER_REQUIRE_AGENT_TYPE:
+                               case NECP_CLIENT_PARAMETER_PREFER_AGENT_TYPE: {
                                        if (num_required_agent_types >= NECP_MAX_REQUIRED_AGENTS) {
                                                break;
                                        }
@@ -6557,24 +6322,48 @@ necp_application_find_policy_match_internal(proc_t proc,
                offset += sizeof(u_int8_t) + sizeof(u_int32_t) + length;
        }
 
+       // Check for loopback exception
+       if (necp_pass_loopback > 0 && necp_is_loopback(&local_addr.sa, &remote_addr.sa, NULL, NULL, bound_interface_index)) {
+               returned_result->policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+               returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_PASS;
+               returned_result->routed_interface_index = lo_ifp->if_index;
+               *flags |= (NECP_CLIENT_RESULT_FLAG_IS_LOCAL | NECP_CLIENT_RESULT_FLAG_IS_DIRECT);
+               return 0;
+       }
+
        // Lock
        lck_rw_lock_shared(&necp_kernel_policy_lock);
 
-       necp_application_fillout_info_locked(application_uuid, real_application_uuid, account, domain, pid, uid, protocol, bound_interface_index, traffic_class, &local_addr, &remote_addr, proc, &info);
-       matched_policy = necp_socket_find_policy_match_with_info_locked(necp_kernel_socket_policies_app_layer_map, &info, &filter_control_unit, &route_rule_id, &service_action, &service, netagent_ids, netagent_use_flags, NECP_MAX_NETAGENTS, required_agent_types, num_required_agent_types, proc, NULL);
+       u_int32_t route_rule_id_array[MAX_AGGREGATE_ROUTE_RULES];
+       size_t route_rule_id_array_count = 0;
+       necp_application_fillout_info_locked(application_uuid, real_application_uuid, account, domain, pid, uid, protocol, bound_interface_index, traffic_class, &local_addr, &remote_addr, local_port, remote_port, has_client, proc, drop_order, client_flags, &info);
+       matched_policy = necp_socket_find_policy_match_with_info_locked(necp_kernel_socket_policies_app_layer_map, &info, &filter_control_unit, route_rule_id_array, &route_rule_id_array_count, MAX_AGGREGATE_ROUTE_RULES, &service_action, &service, netagent_ids, netagent_use_flags, NECP_MAX_NETAGENTS, required_agent_types, num_required_agent_types, proc, NULL, NULL, &drop_dest_policy_result, &drop_all_bypass);
        if (matched_policy) {
                returned_result->policy_id = matched_policy->id;
                returned_result->routing_result = matched_policy->result;
                memcpy(&returned_result->routing_result_parameter, &matched_policy->result_parameter, sizeof(returned_result->routing_result_parameter));
-       } else if (necp_drop_all_order > 0) {
-               // Mark socket as a drop if drop_all is set
-               returned_result->policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
        } else {
-               returned_result->policy_id = 0;
-               returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_NONE;
+               bool drop_all = false;
+               if (necp_drop_all_order > 0 || info.drop_order > 0 || drop_dest_policy_result == NECP_KERNEL_POLICY_RESULT_DROP) {
+                       // Mark socket as a drop if drop_all is set
+                       drop_all = true;
+                       if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                               drop_all_bypass = necp_check_drop_all_bypass_result(proc);
+                       }
+               }
+               if (drop_all && drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                       returned_result->policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
+               } else {
+                       returned_result->policy_id = 0;
+                       returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_NONE;
+               }
+       }
+       if (filter_control_unit == NECP_FILTER_UNIT_NO_FILTER) {
+               returned_result->filter_control_unit = 0;
+       } else {
+               returned_result->filter_control_unit = filter_control_unit;
        }
-       returned_result->filter_control_unit = filter_control_unit;
        returned_result->service_action = service_action;
 
        // Handle trigger service
@@ -6671,7 +6460,7 @@ necp_application_find_policy_match_internal(proc_t proc,
                        returned_result->routed_interface_index = 0;
                }
 
-               if (no_remote_addr && remote_family == 0 &&
+               if (no_remote_addr && remote_family == AF_UNSPEC &&
                    (rt == NULL || rt->rt_ifp == NULL)) {
                        // Route lookup for default IPv4 failed, try IPv6
 
@@ -6834,13 +6623,28 @@ necp_application_find_policy_match_internal(proc_t proc,
                                }
 
                                // Check QoS marking (fastlane)
-                               if (necp_update_qos_marking(rt->rt_ifp, route_rule_id)) {
-                                       *flags |= NECP_CLIENT_RESULT_FLAG_ALLOW_QOS_MARKING;
+                               for (size_t route_rule_index = 0; route_rule_index < route_rule_id_array_count; route_rule_index++) {
+                                       if (necp_update_qos_marking(rt->rt_ifp, route_rule_id_array[route_rule_index])) {
+                                               *flags |= NECP_CLIENT_RESULT_FLAG_ALLOW_QOS_MARKING;
+                                               // If the route can use QoS markings, stop iterating route rules
+                                               break;
+                                       }
                                }
 
                                if (IFNET_IS_LOW_POWER(rt->rt_ifp)) {
                                        *flags |= NECP_CLIENT_RESULT_FLAG_INTERFACE_LOW_POWER;
                                }
+
+                               if (traffic_class == SO_TC_BK_SYS) {
+                                       // Block BK_SYS traffic if interface is throttled
+                                       u_int32_t throttle_level = 0;
+                                       if (ifnet_get_throttle(rt->rt_ifp, &throttle_level) == 0) {
+                                               if (throttle_level == IFNET_THROTTLE_OPPORTUNISTIC) {
+                                                       returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
+                                                       memset(&returned_result->routing_result_parameter, 0, sizeof(returned_result->routing_result_parameter));
+                                               }
+                                       }
+                               }
                        }
                }
 
@@ -6867,6 +6671,12 @@ necp_application_find_policy_match_internal(proc_t proc,
                                if (v4Route->rt_ifp != NULL && !IS_INTF_CLAT46(v4Route->rt_ifp)) {
                                        *flags |= NECP_CLIENT_RESULT_FLAG_HAS_IPV4;
                                }
+                               if (returned_v4_gateway != NULL &&
+                                   v4Route->rt_gateway != NULL &&
+                                   v4Route->rt_gateway->sa_len == sizeof(returned_v4_gateway->u.sin)) {
+                                       memcpy(&returned_v4_gateway->u.sin, v4Route->rt_gateway, sizeof(returned_v4_gateway->u.sin));
+                                       memset(&returned_v4_gateway->u.sin.sin_zero, 0, sizeof(returned_v4_gateway->u.sin.sin_zero));
+                               }
                                rtfree(v4Route);
                                v4Route = NULL;
                        }
@@ -6879,21 +6689,56 @@ necp_application_find_policy_match_internal(proc_t proc,
                                                *flags |= NECP_CLIENT_RESULT_FLAG_HAS_NAT64;
                                        }
                                }
+                               if (returned_v6_gateway != NULL &&
+                                   v6Route->rt_gateway != NULL &&
+                                   v6Route->rt_gateway->sa_len == sizeof(returned_v6_gateway->u.sin6)) {
+                                       memcpy(&returned_v6_gateway->u.sin6, v6Route->rt_gateway, sizeof(returned_v6_gateway->u.sin6));
+                               }
                                rtfree(v6Route);
                                v6Route = NULL;
                        }
                }
        }
 
-       u_int32_t interface_type_denied = IFRTYPE_FUNCTIONAL_UNKNOWN;
-       bool route_is_allowed = necp_route_is_allowed(rt, NULL, route_rule_id, &interface_type_denied);
-       if (!route_is_allowed) {
-               // If the route is blocked, treat the lookup as a drop
-               returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
-               memset(&returned_result->routing_result_parameter, 0, sizeof(returned_result->routing_result_parameter));
+       for (size_t route_rule_index = 0; route_rule_index < route_rule_id_array_count; route_rule_index++) {
+               u_int32_t interface_type_denied = IFRTYPE_FUNCTIONAL_UNKNOWN;
+               bool route_is_allowed = necp_route_is_allowed(rt, NULL, route_rule_id_array[route_rule_index], &interface_type_denied);
+               if (!route_is_allowed) {
+                       // If the route is blocked, treat the lookup as a drop
+                       returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
+                       memset(&returned_result->routing_result_parameter, 0, sizeof(returned_result->routing_result_parameter));
+
+                       if (interface_type_denied != IFRTYPE_FUNCTIONAL_UNKNOWN) {
+                               if (reason != NULL) {
+                                       if (interface_type_denied == IFRTYPE_FUNCTIONAL_CELLULAR) {
+                                               *reason = NECP_CLIENT_RESULT_REASON_CELLULAR_DENIED;
+                                       } else if (interface_type_denied == IFRTYPE_FUNCTIONAL_WIFI_INFRA) {
+                                               *reason = NECP_CLIENT_RESULT_REASON_WIFI_DENIED;
+                                       }
+                               }
+                               necp_send_application_interface_denied_event(pid, application_uuid, interface_type_denied);
+                       }
+                       // If the route gets denied, stop matching rules
+                       break;
+               }
+       }
 
-               if (interface_type_denied != IFRTYPE_FUNCTIONAL_UNKNOWN) {
-                       necp_send_application_interface_denied_event(pid, application_uuid, interface_type_denied);
+       if (rt != NULL && rt->rt_ifp != NULL) {
+               const bool expensive_prohibited = ((client_flags & NECP_CLIENT_PARAMETER_FLAG_PROHIBIT_EXPENSIVE) &&
+                   IFNET_IS_EXPENSIVE(rt->rt_ifp));
+               const bool constrained_prohibited = ((client_flags & NECP_CLIENT_PARAMETER_FLAG_PROHIBIT_CONSTRAINED) &&
+                   IFNET_IS_CONSTRAINED(rt->rt_ifp));
+               if (reason != NULL) {
+                       if (expensive_prohibited) {
+                               *reason = NECP_CLIENT_RESULT_REASON_EXPENSIVE_PROHIBITED;
+                       } else if (constrained_prohibited) {
+                               *reason = NECP_CLIENT_RESULT_REASON_CONSTRAINED_PROHIBITED;
+                       }
+               }
+               if (expensive_prohibited || constrained_prohibited) {
+                       // If the client flags prohibited a property of the interface, treat it as a drop
+                       returned_result->routing_result = NECP_KERNEL_POLICY_RESULT_DROP;
+                       memset(&returned_result->routing_result_parameter, 0, sizeof(returned_result->routing_result_parameter));
                }
        }
 
@@ -6912,7 +6757,58 @@ necp_application_find_policy_match_internal(proc_t proc,
 }
 
 static bool
-necp_socket_check_policy(struct necp_kernel_socket_policy *kernel_policy, necp_app_id app_id, necp_app_id real_app_id, errno_t cred_result, u_int32_t account_id, struct substring domain, u_int8_t domain_dot_count, pid_t pid, uid_t uid, u_int32_t bound_interface_index, u_int32_t traffic_class, u_int16_t protocol, union necp_sockaddr_union *local, union necp_sockaddr_union *remote, struct necp_client_parameter_netagent_type *required_agent_types, u_int32_t num_required_agent_types, proc_t proc)
+necp_is_route_local(union necp_sockaddr_union *remote_addr)
+{
+       bool no_remote_addr = FALSE;
+       u_int8_t remote_family = 0;
+       struct rtentry *rt = NULL;
+       bool is_local = FALSE;
+
+       if (remote_addr == NULL) {
+               return NULL;
+       }
+
+       if (remote_addr->sa.sa_len == 0 ||
+           (remote_addr->sa.sa_family == AF_INET && remote_addr->sin.sin_addr.s_addr == 0) ||
+           (remote_addr->sa.sa_family == AF_INET6 && IN6_IS_ADDR_UNSPECIFIED(&remote_addr->sin6.sin6_addr))) {
+               no_remote_addr = TRUE;
+               remote_family = remote_addr->sa.sa_family;
+       }
+
+       if (no_remote_addr) {
+               memset(remote_addr, 0, sizeof(union necp_sockaddr_union));
+               if (remote_family == AF_INET6) {
+                       // Reset address to ::
+                       remote_addr->sa.sa_family = AF_INET6;
+                       remote_addr->sa.sa_len = sizeof(struct sockaddr_in6);
+               } else {
+                       // Reset address to 0.0.0.0
+                       remote_addr->sa.sa_family = AF_INET;
+                       remote_addr->sa.sa_len = sizeof(struct sockaddr_in);
+               }
+       }
+
+       // Lookup route regardless of the scoped interface to check if
+       // remote address is in a local network.
+       rt = rtalloc1_scoped((struct sockaddr *)remote_addr, 0, 0, 0);
+
+       if (rt == NULL) {
+               goto done;
+       }
+       if (remote_addr->sa.sa_family == AF_INET && IS_INTF_CLAT46(rt->rt_ifp)) {
+               goto free_rt;
+       }
+       is_local = IS_NECP_DEST_IN_LOCAL_NETWORKS(rt);
+
+free_rt:
+       rtfree(rt);
+
+done:
+       return is_local;
+}
+
+static bool
+necp_socket_check_policy(struct necp_kernel_socket_policy *kernel_policy, necp_app_id app_id, necp_app_id real_app_id, errno_t cred_result, u_int32_t account_id, struct substring domain, u_int8_t domain_dot_count, pid_t pid, uid_t uid, u_int32_t bound_interface_index, u_int32_t traffic_class, u_int16_t protocol, union necp_sockaddr_union *local, union necp_sockaddr_union *remote, struct necp_client_parameter_netagent_type *required_agent_types, u_int32_t num_required_agent_types, bool has_client, uint32_t client_flags, int is_platform_binary, proc_t proc, struct rtentry *rt)
 {
        if (!(kernel_policy->condition_mask & NECP_KERNEL_CONDITION_ALL_INTERFACES)) {
                if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_BOUND_INTERFACE) {
@@ -6968,6 +6864,12 @@ necp_socket_check_policy(struct necp_kernel_socket_policy *kernel_policy, necp_a
                }
        }
 
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_HAS_CLIENT) {
+               if (!has_client) {
+                       return FALSE;
+               }
+       }
+
        if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_ENTITLEMENT) {
                if (cred_result != 0) {
                        // Process is missing entitlement
@@ -6975,6 +6877,13 @@ necp_socket_check_policy(struct necp_kernel_socket_policy *kernel_policy, necp_a
                }
        }
 
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_PLATFORM_BINARY) {
+               if (is_platform_binary == 0) {
+                       // Process is not platform binary
+                       return FALSE;
+               }
+       }
+
        if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_CUSTOM_ENTITLEMENT) {
                if (kernel_policy->cond_custom_entitlement_matched == necp_boolean_state_false) {
                        // Process is missing entitlement based on previous check
@@ -7103,6 +7012,21 @@ necp_socket_check_policy(struct necp_kernel_socket_policy *kernel_policy, necp_a
                }
        }
 
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) {
+               bool is_local = FALSE;
+
+               if (rt != NULL) {
+                       is_local = IS_NECP_DEST_IN_LOCAL_NETWORKS(rt);
+               } else {
+                       is_local = necp_is_route_local(remote);
+               }
+
+               if (!is_local) {
+                       // Either no route to validate or no match for local networks
+                       return FALSE;
+               }
+       }
+
        if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_START) {
                if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_END) {
                        bool inRange = necp_is_addr_in_range((struct sockaddr *)local, (struct sockaddr *)&kernel_policy->cond_local_start, (struct sockaddr *)&kernel_policy->cond_local_end);
@@ -7155,6 +7079,46 @@ necp_socket_check_policy(struct necp_kernel_socket_policy *kernel_policy, necp_a
                }
        }
 
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS) {
+               if (kernel_policy->condition_negated_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS) {
+                       if ((client_flags & kernel_policy->cond_client_flags) == kernel_policy->cond_client_flags) {
+                               // Flags do match, and condition is negative, fail.
+                               return FALSE;
+                       }
+               } else {
+                       if ((client_flags & kernel_policy->cond_client_flags) != kernel_policy->cond_client_flags) {
+                               // Flags do not match, fail.
+                               return FALSE;
+                       }
+               }
+       }
+
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_EMPTY) {
+               bool isEmpty = necp_addr_is_empty((struct sockaddr *)local);
+               if (kernel_policy->condition_negated_mask & NECP_KERNEL_CONDITION_LOCAL_EMPTY) {
+                       if (isEmpty) {
+                               return FALSE;
+                       }
+               } else {
+                       if (!isEmpty) {
+                               return FALSE;
+                       }
+               }
+       }
+
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_REMOTE_EMPTY) {
+               bool isEmpty = necp_addr_is_empty((struct sockaddr *)remote);
+               if (kernel_policy->condition_negated_mask & NECP_KERNEL_CONDITION_REMOTE_EMPTY) {
+                       if (isEmpty) {
+                               return FALSE;
+                       }
+               } else {
+                       if (!isEmpty) {
+                               return FALSE;
+                       }
+               }
+       }
+
        return TRUE;
 }
 
@@ -7165,7 +7129,7 @@ necp_socket_calc_flowhash_locked(struct necp_socket_info *info)
 }
 
 static void
-necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_local_addr, struct sockaddr *override_remote_addr, u_int32_t override_bound_interface, struct necp_socket_info *info)
+necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_local_addr, struct sockaddr *override_remote_addr, u_int32_t override_bound_interface, u_int32_t drop_order, struct necp_socket_info *info)
 {
        struct socket *so = NULL;
 
@@ -7173,6 +7137,8 @@ necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_loc
 
        so = inp->inp_socket;
 
+       info->drop_order = drop_order;
+
        if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_CONDITION_PID) {
                info->pid = ((so->so_flags & SOF_DELEGATED) ? so->e_pid : so->last_pid);
        }
@@ -7185,6 +7151,30 @@ necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_loc
                info->traffic_class = so->so_traffic_class;
        }
 
+       if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_CONDITION_HAS_CLIENT) {
+               info->has_client = !uuid_is_null(inp->necp_client_uuid);
+       }
+
+       if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_CONDITION_CLIENT_FLAGS) {
+               info->client_flags = 0;
+               if (INP_NO_CONSTRAINED(inp)) {
+                       info->client_flags |= NECP_CLIENT_PARAMETER_FLAG_PROHIBIT_CONSTRAINED;
+               }
+               if (INP_NO_EXPENSIVE(inp)) {
+                       info->client_flags |= NECP_CLIENT_PARAMETER_FLAG_PROHIBIT_EXPENSIVE;
+               }
+               if (inp->inp_socket->so_flags1 & SOF1_CELLFALLBACK) {
+                       info->client_flags |= NECP_CLIENT_PARAMETER_FLAG_FALLBACK_TRAFFIC;
+               }
+               if (inp->inp_socket->so_flags1 & SOF1_INBOUND) {
+                       info->client_flags |= NECP_CLIENT_PARAMETER_FLAG_INBOUND;
+               }
+               if (inp->inp_socket->so_options & SO_ACCEPTCONN ||
+                   inp->inp_flags2 & INP2_EXTERNAL_PORT) {
+                       info->client_flags |= NECP_CLIENT_PARAMETER_FLAG_LISTENER;
+               }
+       }
+
        if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_CONDITION_PROTOCOL) {
                if (inp->inp_ip_p) {
                        info->protocol = inp->inp_ip_p;
@@ -7215,6 +7205,10 @@ necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_loc
                                necp_get_parent_cred_result(NULL, info);
                        }
                }
+
+               if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_CONDITION_PLATFORM_BINARY) {
+                       info->is_platform_binary = csproc_get_platform_binary(current_proc()) ? true : false;
+               }
        }
 
        if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_CONDITION_ACCOUNT_ID && inp->inp_necp_attributes.inp_account != NULL) {
@@ -7237,45 +7231,51 @@ necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_loc
        }
 
        if (necp_kernel_socket_policies_condition_mask & NECP_KERNEL_ADDRESS_TYPE_CONDITIONS) {
-               if (inp->inp_vflag & INP_IPV4) {
-                       if (override_local_addr) {
-                               if (override_local_addr->sa_len <= sizeof(struct sockaddr_in)) {
-                                       memcpy(&info->local_addr, override_local_addr, override_local_addr->sa_len);
-                               }
-                       } else {
+               if (override_local_addr != NULL) {
+                       if (override_local_addr->sa_family == AF_INET6 && override_local_addr->sa_len <= sizeof(struct sockaddr_in6)) {
+                               memcpy(&info->local_addr, override_local_addr, override_local_addr->sa_len);
+                               if (IN6_IS_ADDR_V4MAPPED(&(info->local_addr.sin6.sin6_addr))) {
+                                       struct sockaddr_in sin;
+                                       in6_sin6_2_sin(&sin, &(info->local_addr.sin6));
+                                       memset(&info->local_addr, 0, sizeof(union necp_sockaddr_union));
+                                       memcpy(&info->local_addr, &sin, sin.sin_len);
+                               }
+                       } else if (override_local_addr->sa_family == AF_INET && override_local_addr->sa_len <= sizeof(struct sockaddr_in)) {
+                               memcpy(&info->local_addr, override_local_addr, override_local_addr->sa_len);
+                       }
+               } else {
+                       if (inp->inp_vflag & INP_IPV4) {
                                ((struct sockaddr_in *)&info->local_addr)->sin_family = AF_INET;
                                ((struct sockaddr_in *)&info->local_addr)->sin_len = sizeof(struct sockaddr_in);
                                ((struct sockaddr_in *)&info->local_addr)->sin_port = inp->inp_lport;
                                memcpy(&((struct sockaddr_in *)&info->local_addr)->sin_addr, &inp->inp_laddr, sizeof(struct in_addr));
-                       }
-
-                       if (override_remote_addr) {
-                               if (override_remote_addr->sa_len <= sizeof(struct sockaddr_in)) {
-                                       memcpy(&info->remote_addr, override_remote_addr, override_remote_addr->sa_len);
-                               }
-                       } else {
-                               ((struct sockaddr_in *)&info->remote_addr)->sin_family = AF_INET;
-                               ((struct sockaddr_in *)&info->remote_addr)->sin_len = sizeof(struct sockaddr_in);
-                               ((struct sockaddr_in *)&info->remote_addr)->sin_port = inp->inp_fport;
-                               memcpy(&((struct sockaddr_in *)&info->remote_addr)->sin_addr, &inp->inp_faddr, sizeof(struct in_addr));
-                       }
-               } else if (inp->inp_vflag & INP_IPV6) {
-                       if (override_local_addr) {
-                               if (override_local_addr->sa_len <= sizeof(struct sockaddr_in6)) {
-                                       memcpy(&info->local_addr, override_local_addr, override_local_addr->sa_len);
-                               }
-                       } else {
+                       } else if (inp->inp_vflag & INP_IPV6) {
                                ((struct sockaddr_in6 *)&info->local_addr)->sin6_family = AF_INET6;
                                ((struct sockaddr_in6 *)&info->local_addr)->sin6_len = sizeof(struct sockaddr_in6);
                                ((struct sockaddr_in6 *)&info->local_addr)->sin6_port = inp->inp_lport;
                                memcpy(&((struct sockaddr_in6 *)&info->local_addr)->sin6_addr, &inp->in6p_laddr, sizeof(struct in6_addr));
                        }
+               }
 
-                       if (override_remote_addr) {
-                               if (override_remote_addr->sa_len <= sizeof(struct sockaddr_in6)) {
-                                       memcpy(&info->remote_addr, override_remote_addr, override_remote_addr->sa_len);
+               if (override_remote_addr != NULL) {
+                       if (override_remote_addr->sa_family == AF_INET6 && override_remote_addr->sa_len <= sizeof(struct sockaddr_in6)) {
+                               memcpy(&info->remote_addr, override_remote_addr, override_remote_addr->sa_len);
+                               if (IN6_IS_ADDR_V4MAPPED(&(info->remote_addr.sin6.sin6_addr))) {
+                                       struct sockaddr_in sin;
+                                       in6_sin6_2_sin(&sin, &(info->remote_addr.sin6));
+                                       memset(&info->remote_addr, 0, sizeof(union necp_sockaddr_union));
+                                       memcpy(&info->remote_addr, &sin, sin.sin_len);
                                }
-                       } else {
+                       } else if (override_remote_addr->sa_family == AF_INET && override_remote_addr->sa_len <= sizeof(struct sockaddr_in)) {
+                               memcpy(&info->remote_addr, override_remote_addr, override_remote_addr->sa_len);
+                       }
+               } else {
+                       if (inp->inp_vflag & INP_IPV4) {
+                               ((struct sockaddr_in *)&info->remote_addr)->sin_family = AF_INET;
+                               ((struct sockaddr_in *)&info->remote_addr)->sin_len = sizeof(struct sockaddr_in);
+                               ((struct sockaddr_in *)&info->remote_addr)->sin_port = inp->inp_fport;
+                               memcpy(&((struct sockaddr_in *)&info->remote_addr)->sin_addr, &inp->inp_faddr, sizeof(struct in_addr));
+                       } else if (inp->inp_vflag & INP_IPV6) {
                                ((struct sockaddr_in6 *)&info->remote_addr)->sin6_family = AF_INET6;
                                ((struct sockaddr_in6 *)&info->remote_addr)->sin6_len = sizeof(struct sockaddr_in6);
                                ((struct sockaddr_in6 *)&info->remote_addr)->sin6_port = inp->inp_fport;
@@ -7287,45 +7287,75 @@ necp_socket_fillout_info_locked(struct inpcb *inp, struct sockaddr *override_loc
 
 static inline struct necp_kernel_socket_policy *
 necp_socket_find_policy_match_with_info_locked(struct necp_kernel_socket_policy **policy_search_array, struct necp_socket_info *info,
-    necp_kernel_policy_filter *return_filter, u_int32_t *return_route_rule_id,
+    necp_kernel_policy_filter *return_filter,
+    u_int32_t *return_route_rule_id_array, size_t *return_route_rule_id_array_count, size_t route_rule_id_array_count,
     necp_kernel_policy_result *return_service_action, necp_kernel_policy_service *return_service,
     u_int32_t *return_netagent_array, u_int32_t *return_netagent_use_flags_array, size_t netagent_array_count,
     struct necp_client_parameter_netagent_type *required_agent_types,
-    u_int32_t num_required_agent_types, proc_t proc, necp_kernel_policy_id *skip_policy_id)
+    u_int32_t num_required_agent_types, proc_t proc, necp_kernel_policy_id *skip_policy_id, struct rtentry *rt,
+    necp_kernel_policy_result *return_drop_dest_policy_result, necp_drop_all_bypass_check_result_t *return_drop_all_bypass)
 {
        struct necp_kernel_socket_policy *matched_policy = NULL;
        u_int32_t skip_order = 0;
        u_int32_t skip_session_order = 0;
-       u_int32_t route_rule_id_array[MAX_AGGREGATE_ROUTE_RULES];
        size_t route_rule_id_count = 0;
        int i;
        size_t netagent_cursor = 0;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
+       if (return_drop_all_bypass != NULL) {
+               *return_drop_all_bypass = drop_all_bypass;
+       }
 
        // Pre-process domain for quick matching
        struct substring domain_substring = necp_trim_dots_and_stars(info->domain, info->domain ? strlen(info->domain) : 0);
        u_int8_t domain_dot_count = necp_count_dots(domain_substring.string, domain_substring.length);
 
-       if (return_filter) {
+       if (return_filter != NULL) {
                *return_filter = 0;
        }
 
-       if (return_route_rule_id) {
-               *return_route_rule_id = 0;
+       if (return_route_rule_id_array_count != NULL) {
+               *return_route_rule_id_array_count = 0;
        }
 
-       if (return_service_action) {
+       if (return_service_action != NULL) {
                *return_service_action = 0;
        }
 
-       if (return_service) {
+       if (return_service != NULL) {
                return_service->identifier = 0;
                return_service->data = 0;
        }
 
+       // Do not subject layer-2 filter to NECP policies, return a PASS policy
+       if (necp_pass_interpose > 0 && info->client_flags & NECP_CLIENT_PARAMETER_FLAG_INTERPOSE) {
+               return &pass_policy;
+       }
+
+       *return_drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+
        if (policy_search_array != NULL) {
                for (i = 0; policy_search_array[i] != NULL; i++) {
                        if (necp_drop_all_order != 0 && policy_search_array[i]->session_order >= necp_drop_all_order) {
                                // We've hit a drop all rule
+                               if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                                       drop_all_bypass = necp_check_drop_all_bypass_result(proc);
+                                       if (return_drop_all_bypass != NULL) {
+                                               *return_drop_all_bypass = drop_all_bypass;
+                                       }
+                               }
+                               if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                                       break;
+                               }
+                       }
+                       if (necp_drop_dest_policy.entry_count != 0 &&
+                           necp_address_matches_drop_dest_policy(&info->remote_addr, policy_search_array[i]->session_order)) {
+                               // We've hit a drop by destination address rule
+                               *return_drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_DROP;
+                               break;
+                       }
+                       if (info->drop_order != 0 && policy_search_array[i]->session_order >= info->drop_order) {
+                               // We've hit a drop order for this socket
                                break;
                        }
                        if (skip_session_order && policy_search_array[i]->session_order >= skip_session_order) {
@@ -7346,18 +7376,24 @@ necp_socket_find_policy_match_with_info_locked(struct necp_kernel_socket_policy
                                // Skip this policy
                                continue;
                        }
-                       if (necp_socket_check_policy(policy_search_array[i], info->application_id, info->real_application_id, info->cred_result, info->account_id, domain_substring, domain_dot_count, info->pid, info->uid, info->bound_interface_index, info->traffic_class, info->protocol, &info->local_addr, &info->remote_addr, required_agent_types, num_required_agent_types, proc)) {
+
+                       if (necp_socket_check_policy(policy_search_array[i], info->application_id, info->real_application_id, info->cred_result, info->account_id, domain_substring, domain_dot_count, info->pid, info->uid, info->bound_interface_index, info->traffic_class, info->protocol, &info->local_addr, &info->remote_addr, required_agent_types, num_required_agent_types, info->has_client, info->client_flags, info->is_platform_binary, proc, rt)) {
                                if (policy_search_array[i]->result == NECP_KERNEL_POLICY_RESULT_SOCKET_FILTER) {
-                                       if (return_filter && *return_filter == 0) {
-                                               *return_filter = policy_search_array[i]->result_parameter.filter_control_unit;
+                                       if (return_filter && *return_filter != NECP_FILTER_UNIT_NO_FILTER) {
+                                               necp_kernel_policy_filter control_unit = policy_search_array[i]->result_parameter.filter_control_unit;
+                                               if (control_unit == NECP_FILTER_UNIT_NO_FILTER) {
+                                                       *return_filter = control_unit;
+                                               } else {
+                                                       *return_filter |= control_unit;
+                                               }
                                                if (necp_debug > 1) {
                                                        NECPLOG(LOG_DEBUG, "Socket Policy: (Application %d Real Application %d BoundInterface %d Proto %d) Filter %d", info->application_id, info->real_application_id, info->bound_interface_index, info->protocol, policy_search_array[i]->result_parameter.filter_control_unit);
                                                }
                                        }
                                        continue;
                                } else if (policy_search_array[i]->result == NECP_KERNEL_POLICY_RESULT_ROUTE_RULES) {
-                                       if (return_route_rule_id && route_rule_id_count < MAX_AGGREGATE_ROUTE_RULES) {
-                                               route_rule_id_array[route_rule_id_count++] = policy_search_array[i]->result_parameter.route_rule_id;
+                                       if (return_route_rule_id_array && route_rule_id_count < route_rule_id_array_count) {
+                                               return_route_rule_id_array[route_rule_id_count++] = policy_search_array[i]->result_parameter.route_rule_id;
                                                if (necp_debug > 1) {
                                                        NECPLOG(LOG_DEBUG, "Socket Policy: (Application %d Real Application %d BoundInterface %d Proto %d) Route Rule %d", info->application_id, info->real_application_id, info->bound_interface_index, info->protocol, policy_search_array[i]->result_parameter.route_rule_id);
                                                }
@@ -7408,6 +7444,12 @@ necp_socket_find_policy_match_with_info_locked(struct necp_kernel_socket_policy
                                        continue;
                                }
 
+                               // Matched an allow unentitled, which clears any drop order
+                               if (policy_search_array[i]->result == NECP_KERNEL_POLICY_RESULT_ALLOW_UNENTITLED) {
+                                       info->drop_order = 0;
+                                       continue;
+                               }
+
                                // Passed all tests, found a match
                                matched_policy = policy_search_array[i];
                                break;
@@ -7415,10 +7457,8 @@ necp_socket_find_policy_match_with_info_locked(struct necp_kernel_socket_policy
                }
        }
 
-       if (route_rule_id_count == 1) {
-               *return_route_rule_id = route_rule_id_array[0];
-       } else if (route_rule_id_count > 1) {
-               *return_route_rule_id = necp_create_aggregate_route_rule(route_rule_id_array);
+       if (return_route_rule_id_array_count != NULL) {
+               *return_route_rule_id_array_count = route_rule_id_count;
        }
        return matched_policy;
 }
@@ -7481,7 +7521,7 @@ necp_socket_is_connected(struct inpcb *inp)
 static inline bool
 necp_socket_bypass(struct sockaddr *override_local_addr, struct sockaddr *override_remote_addr, struct inpcb *inp)
 {
-       if (necp_pass_loopback > 0 && necp_is_loopback(override_local_addr, override_remote_addr, inp, NULL)) {
+       if (necp_pass_loopback > 0 && necp_is_loopback(override_local_addr, override_remote_addr, inp, NULL, IFSCOPE_NONE)) {
                return true;
        } else if (necp_is_intcoproc(inp, NULL)) {
                return true;
@@ -7495,11 +7535,12 @@ necp_socket_find_policy_match(struct inpcb *inp, struct sockaddr *override_local
 {
        struct socket *so = NULL;
        necp_kernel_policy_filter filter_control_unit = 0;
-       u_int32_t route_rule_id = 0;
        struct necp_kernel_socket_policy *matched_policy = NULL;
        necp_kernel_policy_id matched_policy_id = NECP_KERNEL_POLICY_ID_NONE;
        necp_kernel_policy_result service_action = 0;
        necp_kernel_policy_service service = { 0, 0 };
+       u_int32_t drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
 
        u_int32_t netagent_ids[NECP_MAX_NETAGENTS];
        memset(&netagent_ids, 0, sizeof(netagent_ids));
@@ -7523,10 +7564,12 @@ necp_socket_find_policy_match(struct inpcb *inp, struct sockaddr *override_local
 
        so = inp->inp_socket;
 
+       u_int32_t drop_order = necp_process_drop_order(so->so_cred);
+
        // Don't lock. Possible race condition, but we don't want the performance hit.
        if (necp_kernel_socket_policies_count == 0 ||
            (!(inp->inp_flags2 & INP2_WANT_APP_POLICY) && necp_kernel_socket_policies_non_app_count == 0)) {
-               if (necp_drop_all_order > 0) {
+               if (necp_drop_all_order > 0 || drop_order > 0) {
                        inp->inp_policyresult.policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
                        inp->inp_policyresult.skip_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
                        inp->inp_policyresult.policy_gencount = 0;
@@ -7560,7 +7603,7 @@ necp_socket_find_policy_match(struct inpcb *inp, struct sockaddr *override_local
        // Lock
        lck_rw_lock_shared(&necp_kernel_policy_lock);
 
-       necp_socket_fillout_info_locked(inp, override_local_addr, override_remote_addr, override_bound_interface, &info);
+       necp_socket_fillout_info_locked(inp, override_local_addr, override_remote_addr, override_bound_interface, drop_order, &info);
        inp->inp_policyresult.app_id = info.application_id;
 
        // Check info
@@ -7578,7 +7621,10 @@ necp_socket_find_policy_match(struct inpcb *inp, struct sockaddr *override_local
 
        // Match socket to policy
        necp_kernel_policy_id skip_policy_id;
-       matched_policy = necp_socket_find_policy_match_with_info_locked(necp_kernel_socket_policies_map[NECP_SOCKET_MAP_APP_ID_TO_BUCKET(info.application_id)], &info, &filter_control_unit, &route_rule_id, &service_action, &service, netagent_ids, NULL, NECP_MAX_NETAGENTS, NULL, 0, current_proc(), &skip_policy_id);
+       u_int32_t route_rule_id_array[MAX_AGGREGATE_ROUTE_RULES];
+       size_t route_rule_id_array_count = 0;
+       matched_policy = necp_socket_find_policy_match_with_info_locked(necp_kernel_socket_policies_map[NECP_SOCKET_MAP_APP_ID_TO_BUCKET(info.application_id)], &info, &filter_control_unit, route_rule_id_array, &route_rule_id_array_count, MAX_AGGREGATE_ROUTE_RULES, &service_action, &service, netagent_ids, NULL, NECP_MAX_NETAGENTS, NULL, 0, current_proc(), &skip_policy_id, inp->inp_route.ro_rt, &drop_dest_policy_result, &drop_all_bypass);
+
        // If the socket matched a scoped service policy, mark as Drop if not registered.
        // This covers the cases in which a service is required (on demand) but hasn't started yet.
        if ((service_action == NECP_KERNEL_POLICY_RESULT_TRIGGER_SCOPED ||
@@ -7655,6 +7701,15 @@ necp_socket_find_policy_match(struct inpcb *inp, struct sockaddr *override_local
                        }
                }
        }
+
+       u_int32_t route_rule_id = 0;
+       if (route_rule_id_array_count == 1) {
+               route_rule_id = route_rule_id_array[0];
+       } else if (route_rule_id_array_count > 1) {
+               route_rule_id = necp_create_aggregate_route_rule(route_rule_id_array);
+       }
+
+       bool reset_tcp_mss = false;
        if (matched_policy) {
                matched_policy_id = matched_policy->id;
                inp->inp_policyresult.policy_id = matched_policy->id;
@@ -7677,40 +7732,54 @@ necp_socket_find_policy_match(struct inpcb *inp, struct sockaddr *override_local
                    matched_policy->result == NECP_KERNEL_POLICY_RESULT_IP_TUNNEL &&
                    info.protocol == IPPROTO_TCP) {
                        // Reset MSS on TCP socket if tunnel policy changes
-                       tcp_mtudisc(inp, 0);
+                       reset_tcp_mss = true;
                }
 
                if (necp_debug > 1) {
                        NECPLOG(LOG_DEBUG, "Socket Policy: %p (BoundInterface %d Proto %d) Policy %d Result %d Parameter %d", inp->inp_socket, info.bound_interface_index, info.protocol, matched_policy->id, matched_policy->result, matched_policy->result_parameter.tunnel_interface_index);
                }
-       } else if (necp_drop_all_order > 0) {
-               // Mark socket as a drop if set
-               inp->inp_policyresult.policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               inp->inp_policyresult.skip_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               inp->inp_policyresult.policy_gencount = necp_kernel_socket_policies_gencount;
-               inp->inp_policyresult.flowhash = flowhash;
-               inp->inp_policyresult.results.filter_control_unit = 0;
-               inp->inp_policyresult.results.route_rule_id = 0;
-               inp->inp_policyresult.results.result = NECP_KERNEL_POLICY_RESULT_DROP;
        } else {
-               // Mark non-matching socket so we don't re-check it
-               inp->inp_policyresult.policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               inp->inp_policyresult.skip_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               inp->inp_policyresult.policy_gencount = necp_kernel_socket_policies_gencount;
-               inp->inp_policyresult.flowhash = flowhash;
-               inp->inp_policyresult.results.filter_control_unit = filter_control_unit; // We may have matched a filter, so mark it!
-               inp->inp_policyresult.results.route_rule_id = route_rule_id; // We may have matched a route rule, so mark it!
-               inp->inp_policyresult.results.result = NECP_KERNEL_POLICY_RESULT_NONE;
+               bool drop_all = false;
+               if (necp_drop_all_order > 0 || info.drop_order > 0 || drop_dest_policy_result == NECP_KERNEL_POLICY_RESULT_DROP) {
+                       // Mark socket as a drop if set
+                       drop_all = true;
+                       if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                               drop_all_bypass = necp_check_drop_all_bypass_result(NULL);
+                       }
+               }
+               if (drop_all && drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                       inp->inp_policyresult.policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       inp->inp_policyresult.skip_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       inp->inp_policyresult.policy_gencount = necp_kernel_socket_policies_gencount;
+                       inp->inp_policyresult.flowhash = flowhash;
+                       inp->inp_policyresult.results.filter_control_unit = 0;
+                       inp->inp_policyresult.results.route_rule_id = 0;
+                       inp->inp_policyresult.results.result = NECP_KERNEL_POLICY_RESULT_DROP;
+               } else {
+                       // Mark non-matching socket so we don't re-check it
+                       inp->inp_policyresult.policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       inp->inp_policyresult.skip_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       inp->inp_policyresult.policy_gencount = necp_kernel_socket_policies_gencount;
+                       inp->inp_policyresult.flowhash = flowhash;
+                       inp->inp_policyresult.results.filter_control_unit = filter_control_unit; // We may have matched a filter, so mark it!
+                       inp->inp_policyresult.results.route_rule_id = route_rule_id; // We may have matched a route rule, so mark it!
+                       inp->inp_policyresult.results.result = NECP_KERNEL_POLICY_RESULT_NONE;
+               }
        }
 
        // Unlock
        lck_rw_done(&necp_kernel_policy_lock);
 
+       if (reset_tcp_mss) {
+               // Update MSS when not holding the policy lock to avoid recursive locking
+               tcp_mtudisc(inp, 0);
+       }
+
        return matched_policy_id;
 }
 
 static bool
-necp_ip_output_check_policy(struct necp_kernel_ip_output_policy *kernel_policy, necp_kernel_policy_id socket_policy_id, necp_kernel_policy_id socket_skip_policy_id, u_int32_t bound_interface_index, u_int32_t last_interface_index, u_int16_t protocol, union necp_sockaddr_union *local, union necp_sockaddr_union *remote)
+necp_ip_output_check_policy(struct necp_kernel_ip_output_policy *kernel_policy, necp_kernel_policy_id socket_policy_id, necp_kernel_policy_id socket_skip_policy_id, u_int32_t bound_interface_index, u_int32_t last_interface_index, u_int16_t protocol, union necp_sockaddr_union *local, union necp_sockaddr_union *remote, struct rtentry *rt)
 {
        if (!(kernel_policy->condition_mask & NECP_KERNEL_CONDITION_ALL_INTERFACES)) {
                if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_BOUND_INTERFACE) {
@@ -7767,6 +7836,21 @@ necp_ip_output_check_policy(struct necp_kernel_ip_output_policy *kernel_policy,
                }
        }
 
+       if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_NETWORKS) {
+               bool is_local = FALSE;
+
+               if (rt != NULL) {
+                       is_local = IS_NECP_DEST_IN_LOCAL_NETWORKS(rt);
+               } else {
+                       is_local = necp_is_route_local(remote);
+               }
+
+               if (!is_local) {
+                       // Either no route to validate or no match for local networks
+                       return FALSE;
+               }
+       }
+
        if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_START) {
                if (kernel_policy->condition_mask & NECP_KERNEL_CONDITION_LOCAL_END) {
                        bool inRange = necp_is_addr_in_range((struct sockaddr *)local, (struct sockaddr *)&kernel_policy->cond_local_start, (struct sockaddr *)&kernel_policy->cond_local_end);
@@ -7823,17 +7907,43 @@ necp_ip_output_check_policy(struct necp_kernel_ip_output_policy *kernel_policy,
 }
 
 static inline struct necp_kernel_ip_output_policy *
-necp_ip_output_find_policy_match_locked(necp_kernel_policy_id socket_policy_id, necp_kernel_policy_id socket_skip_policy_id, u_int32_t bound_interface_index, u_int32_t last_interface_index, u_int16_t protocol, union necp_sockaddr_union *local_addr, union necp_sockaddr_union *remote_addr)
+necp_ip_output_find_policy_match_locked(necp_kernel_policy_id socket_policy_id, necp_kernel_policy_id socket_skip_policy_id, u_int32_t bound_interface_index, u_int32_t last_interface_index, u_int16_t protocol, union necp_sockaddr_union *local_addr, union necp_sockaddr_union *remote_addr, struct rtentry *rt, u_int32_t *return_route_rule_id, necp_kernel_policy_result *return_drop_dest_policy_result, necp_drop_all_bypass_check_result_t *return_drop_all_bypass)
 {
        u_int32_t skip_order = 0;
        u_int32_t skip_session_order = 0;
-       int i;
        struct necp_kernel_ip_output_policy *matched_policy = NULL;
        struct necp_kernel_ip_output_policy **policy_search_array = necp_kernel_ip_output_policies_map[NECP_IP_OUTPUT_MAP_ID_TO_BUCKET(socket_policy_id)];
+       u_int32_t route_rule_id_array[MAX_AGGREGATE_ROUTE_RULES];
+       size_t route_rule_id_count = 0;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
+       if (return_drop_all_bypass != NULL) {
+               *return_drop_all_bypass = drop_all_bypass;
+       }
+
+       if (return_route_rule_id != NULL) {
+               *return_route_rule_id = 0;
+       }
+
+       *return_drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+
        if (policy_search_array != NULL) {
-               for (i = 0; policy_search_array[i] != NULL; i++) {
+               for (int i = 0; policy_search_array[i] != NULL; i++) {
                        if (necp_drop_all_order != 0 && policy_search_array[i]->session_order >= necp_drop_all_order) {
                                // We've hit a drop all rule
+                               if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                                       drop_all_bypass = necp_check_drop_all_bypass_result(NULL);
+                                       if (return_drop_all_bypass != NULL) {
+                                               *return_drop_all_bypass = drop_all_bypass;
+                                       }
+                               }
+                               if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                                       break;
+                               }
+                       }
+                       if (necp_drop_dest_policy.entry_count > 0 &&
+                           necp_address_matches_drop_dest_policy(remote_addr, policy_search_array[i]->session_order)) {
+                               // We've hit a drop by destination address rule
+                               *return_drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_DROP;
                                break;
                        }
                        if (skip_session_order && policy_search_array[i]->session_order >= skip_session_order) {
@@ -7854,28 +7964,39 @@ necp_ip_output_find_policy_match_locked(necp_kernel_policy_id socket_policy_id,
                                // Skip this policy
                                continue;
                        }
-                       if (necp_ip_output_check_policy(policy_search_array[i], socket_policy_id, socket_skip_policy_id, bound_interface_index, last_interface_index, protocol, local_addr, remote_addr)) {
-                               // Passed all tests, found a match
-                               matched_policy = policy_search_array[i];
 
-                               if (policy_search_array[i]->result == NECP_KERNEL_POLICY_RESULT_SKIP) {
+                       if (necp_ip_output_check_policy(policy_search_array[i], socket_policy_id, socket_skip_policy_id, bound_interface_index, last_interface_index, protocol, local_addr, remote_addr, rt)) {
+                               if (policy_search_array[i]->result == NECP_KERNEL_POLICY_RESULT_ROUTE_RULES) {
+                                       if (return_route_rule_id != NULL && route_rule_id_count < MAX_AGGREGATE_ROUTE_RULES) {
+                                               route_rule_id_array[route_rule_id_count++] = policy_search_array[i]->result_parameter.route_rule_id;
+                                       }
+                                       continue;
+                               } else if (policy_search_array[i]->result == NECP_KERNEL_POLICY_RESULT_SKIP) {
                                        skip_order = policy_search_array[i]->result_parameter.skip_policy_order;
                                        skip_session_order = policy_search_array[i]->session_order + 1;
                                        continue;
                                }
 
+                               // Passed all tests, found a match
+                               matched_policy = policy_search_array[i];
                                break;
                        }
                }
        }
 
+       if (route_rule_id_count == 1) {
+               *return_route_rule_id = route_rule_id_array[0];
+       } else if (route_rule_id_count > 1) {
+               *return_route_rule_id = necp_create_aggregate_route_rule(route_rule_id_array);
+       }
+
        return matched_policy;
 }
 
 static inline bool
 necp_output_bypass(struct mbuf *packet)
 {
-       if (necp_pass_loopback > 0 && necp_is_loopback(NULL, NULL, NULL, packet)) {
+       if (necp_pass_loopback > 0 && necp_is_loopback(NULL, NULL, NULL, packet, IFSCOPE_NONE)) {
                return true;
        }
        if (necp_pass_keepalives > 0 && necp_get_is_keepalive_from_packet(packet)) {
@@ -7888,7 +8009,8 @@ necp_output_bypass(struct mbuf *packet)
 }
 
 necp_kernel_policy_id
-necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_args *ipoa, necp_kernel_policy_result *result, necp_kernel_policy_result_parameter *result_parameter)
+necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_args *ipoa, struct rtentry *rt,
+    necp_kernel_policy_result *result, necp_kernel_policy_result_parameter *result_parameter)
 {
        struct ip *ip = NULL;
        int hlen = sizeof(struct ip);
@@ -7901,6 +8023,8 @@ necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_a
        u_int32_t last_interface_index = 0;
        union necp_sockaddr_union local_addr;
        union necp_sockaddr_union remote_addr;
+       u_int32_t drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
 
        if (result) {
                *result = 0;
@@ -7920,7 +8044,7 @@ necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_a
        // Exit early for an empty list
        // Don't lock. Possible race condition, but we don't want the performance hit.
        if (necp_kernel_ip_output_policies_count == 0 ||
-           ((socket_policy_id == NECP_KERNEL_POLICY_ID_NONE) && necp_kernel_ip_output_policies_non_id_count == 0)) {
+           (socket_policy_id == NECP_KERNEL_POLICY_ID_NONE && necp_kernel_ip_output_policies_non_id_count == 0 && necp_drop_dest_policy.entry_count == 0)) {
                if (necp_drop_all_order > 0) {
                        matched_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
                        if (result) {
@@ -7998,7 +8122,8 @@ necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_a
 
        // Match packet to policy
        lck_rw_lock_shared(&necp_kernel_policy_lock);
-       matched_policy = necp_ip_output_find_policy_match_locked(socket_policy_id, socket_skip_policy_id, bound_interface_index, last_interface_index, protocol, &local_addr, &remote_addr);
+       u_int32_t route_rule_id = 0;
+       matched_policy = necp_ip_output_find_policy_match_locked(socket_policy_id, socket_skip_policy_id, bound_interface_index, last_interface_index, protocol, &local_addr, &remote_addr, rt, &route_rule_id, &drop_dest_policy_result, &drop_all_bypass);
        if (matched_policy) {
                matched_policy_id = matched_policy->id;
                if (result) {
@@ -8009,13 +8134,36 @@ necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_a
                        memcpy(result_parameter, &matched_policy->result_parameter, sizeof(matched_policy->result_parameter));
                }
 
+               if (route_rule_id != 0 &&
+                   packet->m_pkthdr.necp_mtag.necp_route_rule_id == 0) {
+                       packet->m_pkthdr.necp_mtag.necp_route_rule_id = route_rule_id;
+               }
+
                if (necp_debug > 1) {
-                       NECPLOG(LOG_DEBUG, "IP Output: (ID %d BoundInterface %d LastInterface %d Proto %d) Policy %d Result %d Parameter %d", socket_policy_id, bound_interface_index, last_interface_index, protocol, matched_policy->id, matched_policy->result, matched_policy->result_parameter.tunnel_interface_index);
+                       NECPLOG(LOG_DEBUG, "IP Output: (ID %d BoundInterface %d LastInterface %d Proto %d) Policy %d Result %d Parameter %d Route Rule %u", socket_policy_id, bound_interface_index, last_interface_index, protocol, matched_policy->id, matched_policy->result, matched_policy->result_parameter.tunnel_interface_index, route_rule_id);
                }
-       } else if (necp_drop_all_order > 0) {
-               matched_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               if (result) {
-                       *result = NECP_KERNEL_POLICY_RESULT_DROP;
+       } else {
+               bool drop_all = false;
+               /*
+                * Apply drop-all only to packets which have never matched a primary policy (check
+                * if the packet saved policy id is none or falls within the socket policy id range).
+                */
+               if (socket_policy_id < NECP_KERNEL_POLICY_ID_FIRST_VALID_IP &&
+                   (necp_drop_all_order > 0 || drop_dest_policy_result == NECP_KERNEL_POLICY_RESULT_DROP)) {
+                       drop_all = true;
+                       if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                               drop_all_bypass = necp_check_drop_all_bypass_result(NULL);
+                       }
+               }
+               if (drop_all && drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                       matched_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       if (result) {
+                               *result = NECP_KERNEL_POLICY_RESULT_DROP;
+                       }
+               } else if (route_rule_id != 0 &&
+                   packet->m_pkthdr.necp_mtag.necp_route_rule_id == 0) {
+                       // If we matched a route rule, mark it
+                       packet->m_pkthdr.necp_mtag.necp_route_rule_id = route_rule_id;
                }
        }
 
@@ -8025,7 +8173,8 @@ necp_ip_output_find_policy_match(struct mbuf *packet, int flags, struct ip_out_a
 }
 
 necp_kernel_policy_id
-necp_ip6_output_find_policy_match(struct mbuf *packet, int flags, struct ip6_out_args *ip6oa, necp_kernel_policy_result *result, necp_kernel_policy_result_parameter *result_parameter)
+necp_ip6_output_find_policy_match(struct mbuf *packet, int flags, struct ip6_out_args *ip6oa, struct rtentry *rt,
+    necp_kernel_policy_result *result, necp_kernel_policy_result_parameter *result_parameter)
 {
        struct ip6_hdr *ip6 = NULL;
        int next = -1;
@@ -8039,6 +8188,8 @@ necp_ip6_output_find_policy_match(struct mbuf *packet, int flags, struct ip6_out
        u_int32_t last_interface_index = 0;
        union necp_sockaddr_union local_addr;
        union necp_sockaddr_union remote_addr;
+       u_int32_t drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
 
        if (result) {
                *result = 0;
@@ -8058,7 +8209,7 @@ necp_ip6_output_find_policy_match(struct mbuf *packet, int flags, struct ip6_out
        // Exit early for an empty list
        // Don't lock. Possible race condition, but we don't want the performance hit.
        if (necp_kernel_ip_output_policies_count == 0 ||
-           ((socket_policy_id == NECP_KERNEL_POLICY_ID_NONE) && necp_kernel_ip_output_policies_non_id_count == 0)) {
+           (socket_policy_id == NECP_KERNEL_POLICY_ID_NONE && necp_kernel_ip_output_policies_non_id_count == 0 && necp_drop_dest_policy.entry_count == 0)) {
                if (necp_drop_all_order > 0) {
                        matched_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
                        if (result) {
@@ -8133,7 +8284,8 @@ necp_ip6_output_find_policy_match(struct mbuf *packet, int flags, struct ip6_out
 
        // Match packet to policy
        lck_rw_lock_shared(&necp_kernel_policy_lock);
-       matched_policy = necp_ip_output_find_policy_match_locked(socket_policy_id, socket_skip_policy_id, bound_interface_index, last_interface_index, protocol, &local_addr, &remote_addr);
+       u_int32_t route_rule_id = 0;
+       matched_policy = necp_ip_output_find_policy_match_locked(socket_policy_id, socket_skip_policy_id, bound_interface_index, last_interface_index, protocol, &local_addr, &remote_addr, rt, &route_rule_id, &drop_dest_policy_result, &drop_all_bypass);
        if (matched_policy) {
                matched_policy_id = matched_policy->id;
                if (result) {
@@ -8144,13 +8296,36 @@ necp_ip6_output_find_policy_match(struct mbuf *packet, int flags, struct ip6_out
                        memcpy(result_parameter, &matched_policy->result_parameter, sizeof(matched_policy->result_parameter));
                }
 
+               if (route_rule_id != 0 &&
+                   packet->m_pkthdr.necp_mtag.necp_route_rule_id == 0) {
+                       packet->m_pkthdr.necp_mtag.necp_route_rule_id = route_rule_id;
+               }
+
                if (necp_debug > 1) {
-                       NECPLOG(LOG_DEBUG, "IP6 Output: (ID %d BoundInterface %d LastInterface %d Proto %d) Policy %d Result %d Parameter %d", socket_policy_id, bound_interface_index, last_interface_index, protocol, matched_policy->id, matched_policy->result, matched_policy->result_parameter.tunnel_interface_index);
+                       NECPLOG(LOG_DEBUG, "IP6 Output: (ID %d BoundInterface %d LastInterface %d Proto %d) Policy %d Result %d Parameter %d Route Rule %u", socket_policy_id, bound_interface_index, last_interface_index, protocol, matched_policy->id, matched_policy->result, matched_policy->result_parameter.tunnel_interface_index, route_rule_id);
                }
-       } else if (necp_drop_all_order > 0) {
-               matched_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
-               if (result) {
-                       *result = NECP_KERNEL_POLICY_RESULT_DROP;
+       } else {
+               bool drop_all = false;
+               /*
+                * Apply drop-all only to packets which have never matched a primary policy (check
+                * if the packet saved policy id is none or falls within the socket policy id range).
+                */
+               if (socket_policy_id < NECP_KERNEL_POLICY_ID_FIRST_VALID_IP &&
+                   (necp_drop_all_order > 0 || drop_dest_policy_result == NECP_KERNEL_POLICY_RESULT_DROP)) {
+                       drop_all = true;
+                       if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                               drop_all_bypass = necp_check_drop_all_bypass_result(NULL);
+                       }
+               }
+               if (drop_all && drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                       matched_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       if (result) {
+                               *result = NECP_KERNEL_POLICY_RESULT_DROP;
+                       }
+               } else if (route_rule_id != 0 &&
+                   packet->m_pkthdr.necp_mtag.necp_route_rule_id == 0) {
+                       // If we matched a route rule, mark it
+                       packet->m_pkthdr.necp_mtag.necp_route_rule_id = route_rule_id;
                }
        }
 
@@ -8360,6 +8535,54 @@ necp_buffer_compare_with_bit_prefix(u_int8_t *p1, u_int8_t *p2, u_int32_t bits)
        return TRUE;
 }
 
+static bool
+necp_addr_is_empty(struct sockaddr *addr)
+{
+       if (addr == NULL) {
+               return TRUE;
+       }
+
+       if (addr->sa_len == 0) {
+               return TRUE;
+       }
+
+       switch (addr->sa_family) {
+       case AF_INET: {
+               static struct sockaddr_in ipv4_empty_address = {
+                       .sin_len = sizeof(struct sockaddr_in),
+                       .sin_family = AF_INET,
+                       .sin_port = 0,
+                       .sin_addr = { .s_addr = 0 }, // 0.0.0.0
+                       .sin_zero = {0},
+               };
+               if (necp_addr_compare(addr, (struct sockaddr *)&ipv4_empty_address, 0) == 0) {
+                       return TRUE;
+               } else {
+                       return FALSE;
+               }
+       }
+       case AF_INET6: {
+               static struct sockaddr_in6 ipv6_empty_address = {
+                       .sin6_len = sizeof(struct sockaddr_in6),
+                       .sin6_family = AF_INET6,
+                       .sin6_port = 0,
+                       .sin6_flowinfo = 0,
+                       .sin6_addr = IN6ADDR_ANY_INIT, // ::
+                       .sin6_scope_id = 0,
+               };
+               if (necp_addr_compare(addr, (struct sockaddr *)&ipv6_empty_address, 0) == 0) {
+                       return TRUE;
+               } else {
+                       return FALSE;
+               }
+       }
+       default:
+               return FALSE;
+       }
+
+       return FALSE;
+}
+
 static bool
 necp_update_qos_marking(struct ifnet *ifp, u_int32_t route_rule_id)
 {
@@ -8399,12 +8622,13 @@ necp_update_qos_marking(struct ifnet *ifp, u_int32_t route_rule_id)
        if ((route_rule->cellular_action == NECP_ROUTE_RULE_QOS_MARKING && IFNET_IS_CELLULAR(ifp)) ||
            (route_rule->wifi_action == NECP_ROUTE_RULE_QOS_MARKING && IFNET_IS_WIFI(ifp)) ||
            (route_rule->wired_action == NECP_ROUTE_RULE_QOS_MARKING && IFNET_IS_WIRED(ifp)) ||
-           (route_rule->expensive_action == NECP_ROUTE_RULE_QOS_MARKING && IFNET_IS_EXPENSIVE(ifp))) {
+           (route_rule->expensive_action == NECP_ROUTE_RULE_QOS_MARKING && IFNET_IS_EXPENSIVE(ifp)) ||
+           (route_rule->constrained_action == NECP_ROUTE_RULE_QOS_MARKING && IFNET_IS_CONSTRAINED(ifp))) {
                qos_marking = TRUE;
                if (necp_debug > 2) {
-                       NECPLOG(LOG_DEBUG, "QoS Marking: C:%d WF:%d W:%d E:%d for Rule %d Allowed %d",
+                       NECPLOG(LOG_DEBUG, "QoS Marking: C:%d WF:%d W:%d E:%d Cn:%d for Rule %d Allowed %d",
                            route_rule->cellular_action, route_rule->wifi_action, route_rule->wired_action,
-                           route_rule->expensive_action, route_rule_id, qos_marking);
+                           route_rule->expensive_action, route_rule->constrained_action, route_rule_id, qos_marking);
                }
                goto done;
        }
@@ -8630,6 +8854,22 @@ necp_route_is_allowed_inner(struct rtentry *route, struct ifnet *ifp, u_int32_t
                }
        }
 
+       if (IFNET_IS_CONSTRAINED(ifp)) {
+               if (route_rule->constrained_action == NECP_ROUTE_RULE_DENY_LQM_ABORT) {
+                       if (necp_route_is_lqm_abort(ifp, delegated_ifp)) {
+                               // Mark aggregate action as deny
+                               type_aggregate_action = NECP_ROUTE_RULE_DENY_INTERFACE;
+                       }
+               } else if (IS_NECP_ROUTE_RULE_ALLOW_OR_DENY(route_rule->constrained_action)) {
+                       if (type_aggregate_action == NECP_ROUTE_RULE_NONE ||
+                           (type_aggregate_action == NECP_ROUTE_RULE_ALLOW_INTERFACE &&
+                           route_rule->constrained_action == NECP_ROUTE_RULE_DENY_INTERFACE)) {
+                               // Deny wins if there is a conflict
+                               type_aggregate_action = route_rule->constrained_action;
+                       }
+               }
+       }
+
        if (type_aggregate_action != NECP_ROUTE_RULE_NONE) {
                if (necp_debug > 1) {
                        NECPLOG(LOG_DEBUG, "Route Allowed: C:%d WF:%d W:%d E:%d for Rule %d Allowed %d", route_rule->cellular_action, route_rule->wifi_action, route_rule->wired_action, route_rule->expensive_action, route_rule_id, ((type_aggregate_action == NECP_ROUTE_RULE_DENY_INTERFACE) ? FALSE : TRUE));
@@ -8726,7 +8966,8 @@ necp_socket_is_allowed_to_send_recv_internal(struct inpcb *inp, struct sockaddr
        u_int32_t route_rule_id = 0;
        struct rtentry *route = NULL;
        u_int32_t interface_type_denied = IFRTYPE_FUNCTIONAL_UNKNOWN;
-
+       necp_kernel_policy_result drop_dest_policy_result = NECP_KERNEL_POLICY_RESULT_NONE;
+       necp_drop_all_bypass_check_result_t drop_all_bypass = NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE;
        u_int32_t netagent_ids[NECP_MAX_NETAGENTS];
        memset(&netagent_ids, 0, sizeof(netagent_ids));
 
@@ -8746,10 +8987,14 @@ necp_socket_is_allowed_to_send_recv_internal(struct inpcb *inp, struct sockaddr
 
        route = inp->inp_route.ro_rt;
 
+       struct socket *so = inp->inp_socket;
+
+       u_int32_t drop_order = necp_process_drop_order(so->so_cred);
+
        // Don't lock. Possible race condition, but we don't want the performance hit.
        if (necp_kernel_socket_policies_count == 0 ||
            (!(inp->inp_flags2 & INP2_WANT_APP_POLICY) && necp_kernel_socket_policies_non_app_count == 0)) {
-               if (necp_drop_all_order > 0) {
+               if (necp_drop_all_order > 0 || drop_order > 0) {
                        if (necp_socket_bypass(override_local_addr, override_remote_addr, inp)) {
                                allowed_to_receive = TRUE;
                        } else {
@@ -8806,7 +9051,7 @@ necp_socket_is_allowed_to_send_recv_internal(struct inpcb *inp, struct sockaddr
 
        // Actually calculate policy result
        lck_rw_lock_shared(&necp_kernel_policy_lock);
-       necp_socket_fillout_info_locked(inp, override_local_addr, override_remote_addr, 0, &info);
+       necp_socket_fillout_info_locked(inp, override_local_addr, override_remote_addr, 0, drop_order, &info);
 
        flowhash = necp_socket_calc_flowhash_locked(&info);
        if (inp->inp_policyresult.policy_id != NECP_KERNEL_POLICY_ID_NONE &&
@@ -8834,7 +9079,16 @@ necp_socket_is_allowed_to_send_recv_internal(struct inpcb *inp, struct sockaddr
                goto done;
        }
 
-       struct necp_kernel_socket_policy *matched_policy = necp_socket_find_policy_match_with_info_locked(necp_kernel_socket_policies_map[NECP_SOCKET_MAP_APP_ID_TO_BUCKET(info.application_id)], &info, NULL, &route_rule_id, &service_action, &service, netagent_ids, NULL, NECP_MAX_NETAGENTS, NULL, 0, current_proc(), return_skip_policy_id);
+       u_int32_t route_rule_id_array[MAX_AGGREGATE_ROUTE_RULES];
+       size_t route_rule_id_array_count = 0;
+       struct necp_kernel_socket_policy *matched_policy = necp_socket_find_policy_match_with_info_locked(necp_kernel_socket_policies_map[NECP_SOCKET_MAP_APP_ID_TO_BUCKET(info.application_id)], &info, NULL, route_rule_id_array, &route_rule_id_array_count, MAX_AGGREGATE_ROUTE_RULES, &service_action, &service, netagent_ids, NULL, NECP_MAX_NETAGENTS, NULL, 0, current_proc(), return_skip_policy_id, inp->inp_route.ro_rt, &drop_dest_policy_result, &drop_all_bypass);
+
+       if (route_rule_id_array_count == 1) {
+               route_rule_id = route_rule_id_array[0];
+       } else if (route_rule_id_array_count > 1) {
+               route_rule_id = necp_create_aggregate_route_rule(route_rule_id_array);
+       }
+
        if (matched_policy != NULL) {
                if (matched_policy->result == NECP_KERNEL_POLICY_RESULT_DROP ||
                    matched_policy->result == NECP_KERNEL_POLICY_RESULT_SOCKET_DIVERT ||
@@ -8861,14 +9115,23 @@ necp_socket_is_allowed_to_send_recv_internal(struct inpcb *inp, struct sockaddr
                        NECPLOG(LOG_DEBUG, "Socket Send/Recv Policy: Policy %d Allowed %d", return_policy_id ? *return_policy_id : 0, allowed_to_receive);
                }
                goto done;
-       } else if (necp_drop_all_order > 0) {
-               allowed_to_receive = FALSE;
        } else {
-               if (return_policy_id) {
-                       *return_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+               bool drop_all = false;
+               if (necp_drop_all_order > 0 || info.drop_order > 0 || drop_dest_policy_result == NECP_KERNEL_POLICY_RESULT_DROP) {
+                       drop_all = true;
+                       if (drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_NONE) {
+                               drop_all_bypass = necp_check_drop_all_bypass_result(NULL);
+                       }
                }
-               if (return_route_rule_id) {
-                       *return_route_rule_id = route_rule_id;
+               if (drop_all && drop_all_bypass == NECP_DROP_ALL_BYPASS_CHECK_RESULT_FALSE) {
+                       allowed_to_receive = FALSE;
+               } else {
+                       if (return_policy_id) {
+                               *return_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+                       }
+                       if (return_route_rule_id) {
+                               *return_route_rule_id = route_rule_id;
+                       }
                }
        }
 
@@ -8915,10 +9178,13 @@ necp_socket_is_allowed_to_send_recv_v6(struct inpcb *inp, u_int16_t local_port,
 }
 
 bool
-necp_socket_is_allowed_to_send_recv(struct inpcb *inp, necp_kernel_policy_id *return_policy_id, u_int32_t *return_route_rule_id,
+necp_socket_is_allowed_to_send_recv(struct inpcb *inp, ifnet_t interface, necp_kernel_policy_id *return_policy_id,
+    u_int32_t *return_route_rule_id,
     necp_kernel_policy_id *return_skip_policy_id)
 {
-       return necp_socket_is_allowed_to_send_recv_internal(inp, NULL, NULL, NULL, return_policy_id, return_route_rule_id, return_skip_policy_id);
+       return necp_socket_is_allowed_to_send_recv_internal(inp, NULL, NULL, interface,
+                  return_policy_id, return_route_rule_id,
+                  return_skip_policy_id);
 }
 
 int
@@ -8946,8 +9212,18 @@ necp_mark_packet_from_socket(struct mbuf *packet, struct inpcb *inp, necp_kernel
        }
        packet->m_pkthdr.necp_mtag.necp_app_id = inp->inp_policyresult.app_id;
 
-       if (skip_policy_id != NECP_KERNEL_POLICY_ID_NONE) {
+       if (skip_policy_id != NECP_KERNEL_POLICY_ID_NONE &&
+           skip_policy_id != NECP_KERNEL_POLICY_ID_NO_MATCH) {
+               // Only mark the skip policy if it is a valid policy ID
                packet->m_pkthdr.necp_mtag.necp_skip_policy_id = skip_policy_id;
+       } else if (inp->inp_policyresult.results.filter_control_unit == NECP_FILTER_UNIT_NO_FILTER) {
+               // Overload the meaning of "NECP_KERNEL_POLICY_ID_NO_MATCH"
+               // to indicate that NECP_FILTER_UNIT_NO_FILTER was set
+               // See necp_get_skip_policy_id_from_packet() and
+               // necp_packet_should_skip_filters().
+               packet->m_pkthdr.necp_mtag.necp_skip_policy_id = NECP_KERNEL_POLICY_ID_NO_MATCH;
+       } else {
+               packet->m_pkthdr.necp_mtag.necp_skip_policy_id = NECP_KERNEL_POLICY_ID_NONE;
        }
 
        return 0;
@@ -9018,9 +9294,25 @@ necp_get_skip_policy_id_from_packet(struct mbuf *packet)
                return NECP_KERNEL_POLICY_ID_NONE;
        }
 
+       // Check for overloaded value. See necp_mark_packet_from_socket().
+       if (packet->m_pkthdr.necp_mtag.necp_skip_policy_id == NECP_KERNEL_POLICY_ID_NO_MATCH) {
+               return NECP_KERNEL_POLICY_ID_NONE;
+       }
+
        return packet->m_pkthdr.necp_mtag.necp_skip_policy_id;
 }
 
+bool
+necp_packet_should_skip_filters(struct mbuf *packet)
+{
+       if (packet == NULL || !(packet->m_flags & M_PKTHDR)) {
+               return false;
+       }
+
+       // Check for overloaded value. See necp_mark_packet_from_socket().
+       return packet->m_pkthdr.necp_mtag.necp_skip_policy_id == NECP_KERNEL_POLICY_ID_NO_MATCH;
+}
+
 u_int32_t
 necp_get_last_interface_index_from_packet(struct mbuf *packet)
 {
@@ -9273,7 +9565,7 @@ necp_addr_is_loopback(struct sockaddr *address)
 }
 
 static bool
-necp_is_loopback(struct sockaddr *local_addr, struct sockaddr *remote_addr, struct inpcb *inp, struct mbuf *packet)
+necp_is_loopback(struct sockaddr *local_addr, struct sockaddr *remote_addr, struct inpcb *inp, struct mbuf *packet, u_int32_t bound_interface_index)
 {
        // Note: This function only checks for the loopback addresses.
        // In the future, we may want to expand to also allow any traffic
@@ -9303,6 +9595,8 @@ necp_is_loopback(struct sockaddr *local_addr, struct sockaddr *remote_addr, stru
                                return TRUE;
                        }
                }
+       } else if (bound_interface_index != IFSCOPE_NONE && lo_ifp->if_index == bound_interface_index) {
+               return TRUE;
        }
 
        if (packet != NULL) {
@@ -9332,7 +9626,17 @@ static bool
 necp_is_intcoproc(struct inpcb *inp, struct mbuf *packet)
 {
        if (inp != NULL) {
-               return sflt_permission_check(inp) ? true : false;
+               if (!(inp->inp_vflag & INP_IPV6)) {
+                       return false;
+               }
+               if (INP_INTCOPROC_ALLOWED(inp)) {
+                       return true;
+               }
+               if ((inp->inp_flags & INP_BOUND_IF) &&
+                   IFNET_IS_INTCOPROC(inp->inp_boundifp)) {
+                       return true;
+               }
+               return false;
        }
        if (packet != NULL) {
                struct ip6_hdr *ip6 = mtod(packet, struct ip6_hdr *);
@@ -9346,3 +9650,176 @@ necp_is_intcoproc(struct inpcb *inp, struct mbuf *packet)
 
        return false;
 }
+
+static bool
+necp_address_matches_drop_dest_policy(union necp_sockaddr_union *sau, u_int32_t session_order)
+{
+       char dest_str[MAX_IPv6_STR_LEN];
+
+       if (necp_drop_dest_debug > 0) {
+               if (sau->sa.sa_family == AF_INET) {
+                       (void) inet_ntop(AF_INET, &sau->sin.sin_addr, dest_str, sizeof(dest_str));
+               } else if (sau->sa.sa_family == AF_INET6) {
+                       (void) inet_ntop(AF_INET6, &sau->sin6.sin6_addr, dest_str, sizeof(dest_str));
+               } else {
+                       dest_str[0] = 0;
+               }
+       }
+       for (u_int32_t i = 0; i < necp_drop_dest_policy.entry_count; i++) {
+               struct necp_drop_dest_entry *necp_drop_dest_entry = &necp_drop_dest_policy.entries[i];
+               struct necp_policy_condition_addr *npca = &necp_drop_dest_entry->cond_addr;
+
+               if (session_order >= necp_drop_dest_entry->order && necp_is_addr_in_subnet(&sau->sa, &npca->address.sa, npca->prefix)) {
+                       if (necp_drop_dest_debug > 0) {
+                               char subnet_str[MAX_IPv6_STR_LEN];
+                               struct proc *p = current_proc();
+                               pid_t pid = proc_pid(p);
+
+                               if (sau->sa.sa_family == AF_INET) {
+                                       (void) inet_ntop(AF_INET, &npca->address.sin, subnet_str, sizeof(subnet_str));
+                                       os_log(OS_LOG_DEFAULT, "%s (process %s:%u) %s matches %s/%u", __func__, proc_best_name(p), pid, dest_str, subnet_str, npca->prefix);
+                               } else if (sau->sa.sa_family == AF_INET6) {
+                                       (void) inet_ntop(AF_INET6, &npca->address.sin6, subnet_str, sizeof(subnet_str));
+                                       os_log(OS_LOG_DEFAULT, "%s (process %s:%u) %s matches %s/%u", __func__, proc_best_name(p), pid, dest_str, subnet_str, npca->prefix);
+                               }
+                       }
+                       return true;
+               }
+       }
+       if (necp_drop_dest_debug > 1) {
+               struct proc *p = current_proc();
+               pid_t pid = proc_pid(p);
+
+               os_log(OS_LOG_DEFAULT, "%s (process %s:%u) %s no match", __func__, proc_best_name(p), pid, dest_str);
+       }
+       return false;
+}
+
+static int
+sysctl_handle_necp_drop_dest_level SYSCTL_HANDLER_ARGS
+{
+#pragma unused(arg1, arg2, oidp)
+       int changed = 0;
+       int error = 0;
+       struct necp_drop_dest_policy tmp_drop_dest_policy;
+       struct proc *p = current_proc();
+       pid_t pid = proc_pid(p);
+
+       if (req->newptr != USER_ADDR_NULL && proc_suser(current_proc()) != 0 &&
+           priv_check_cred(kauth_cred_get(), PRIV_NET_PRIVILEGED_NECP_POLICIES, 0) != 0) {
+               NECPLOG(LOG_ERR, "%s (process %s:%u) not permitted", __func__, proc_best_name(p), pid);
+               return EPERM;
+       }
+       if (req->newptr != USER_ADDR_NULL && req->newlen != sizeof(struct necp_drop_dest_policy)) {
+               NECPLOG(LOG_ERR, "%s (process %s:%u) bad newlen %lu", __func__, proc_best_name(p), pid, req->newlen);
+               return EINVAL;
+       }
+
+       memcpy(&tmp_drop_dest_policy, &necp_drop_dest_policy, sizeof(struct necp_drop_dest_policy));
+       error = sysctl_io_opaque(req, &tmp_drop_dest_policy, sizeof(struct necp_drop_dest_policy), &changed);
+       if (error != 0) {
+               NECPLOG(LOG_ERR, "%s (process %s:%u) sysctl_io_opaque() error %d", __func__, proc_best_name(p), pid, error);
+               return error;
+       }
+       if (changed == 0 || req->newptr == USER_ADDR_NULL) {
+               return error;
+       }
+
+       //
+       // Validate the passed parameters
+       //
+       if (tmp_drop_dest_policy.entry_count >= MAX_NECP_DROP_DEST_LEVEL_ADDRS) {
+               NECPLOG(LOG_ERR, "%s (process %s:%u) bad entry_count %u", __func__, proc_best_name(p), pid, tmp_drop_dest_policy.entry_count);
+               return EINVAL;
+       }
+       for (u_int32_t i = 0; i < tmp_drop_dest_policy.entry_count; i++) {
+               struct necp_drop_dest_entry *tmp_drop_dest_entry = &tmp_drop_dest_policy.entries[i];
+               struct necp_policy_condition_addr *npca = &tmp_drop_dest_entry->cond_addr;
+
+               switch (tmp_drop_dest_entry->level) {
+               case NECP_SESSION_PRIORITY_UNKNOWN:
+                       if (tmp_drop_dest_policy.entry_count != 0) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) NECP_SESSION_PRIORITY_UNKNOWN bad entry_count %u", __func__, proc_best_name(p), pid, tmp_drop_dest_policy.entry_count);
+                               return EINVAL;
+                       }
+                       break;
+               case NECP_SESSION_PRIORITY_CONTROL:
+               case NECP_SESSION_PRIORITY_PRIVILEGED_TUNNEL:
+               case NECP_SESSION_PRIORITY_HIGH:
+               case NECP_SESSION_PRIORITY_DEFAULT:
+               case NECP_SESSION_PRIORITY_LOW:
+                       if (tmp_drop_dest_policy.entry_count == 0) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) priority %u entry_count 0", __func__, proc_best_name(p), pid, tmp_drop_dest_entry->level);
+                               return EINVAL;
+                       }
+                       break;
+               default: {
+                       NECPLOG(LOG_ERR, "%s (process %s:%u) bad level %u", __func__, proc_best_name(p), pid, tmp_drop_dest_entry->level);
+                       return EINVAL;
+               }
+               }
+
+               switch (npca->address.sa.sa_family) {
+               case AF_INET: {
+                       if (npca->prefix > 32) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET bad prefix %u", __func__, proc_best_name(p), pid, npca->prefix);
+                               return EINVAL;
+                       }
+                       if (npca->address.sin.sin_len != sizeof(struct sockaddr_in)) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET bad sin_len %u", __func__, proc_best_name(p), pid, npca->address.sin.sin_len);
+                               return EINVAL;
+                       }
+                       if (npca->address.sin.sin_port != 0) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET bad sin_port %u, not zero", __func__, proc_best_name(p), pid, npca->address.sin.sin_port);
+                               return EINVAL;
+                       }
+                       break;
+               }
+               case AF_INET6: {
+                       if (npca->prefix > 128) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET6 bad prefix %u", __func__, proc_best_name(p), pid, npca->prefix);
+                               return EINVAL;
+                       }
+                       if (npca->address.sin6.sin6_len != sizeof(struct sockaddr_in6)) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET6 bad sin6_len %u", __func__, proc_best_name(p), pid, npca->address.sin6.sin6_len);
+                               return EINVAL;
+                       }
+                       if (npca->address.sin6.sin6_port != 0) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET6 bad sin6_port %u, not zero", __func__, proc_best_name(p), pid, npca->address.sin6.sin6_port);
+                               return EINVAL;
+                       }
+                       if (npca->address.sin6.sin6_flowinfo != 0) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET6 bad sin6_flowinfo %u, not zero", __func__, proc_best_name(p), pid, npca->address.sin6.sin6_flowinfo);
+                               return EINVAL;
+                       }
+                       if (npca->address.sin6.sin6_scope_id != 0) {
+                               NECPLOG(LOG_ERR, "%s (process %s:%u) AF_INET6 bad sin6_scope_id %u, not zero", __func__, proc_best_name(p), pid, npca->address.sin6.sin6_scope_id);
+                               return EINVAL;
+                       }
+                       break;
+               }
+               default: {
+                       return EINVAL;
+               }
+               }
+       }
+
+       //
+       // Commit the changed policy
+       //
+       lck_rw_lock_exclusive(&necp_kernel_policy_lock);
+       memset(&necp_drop_dest_policy, 0, sizeof(struct necp_drop_dest_policy));
+
+       necp_drop_dest_policy.entry_count = tmp_drop_dest_policy.entry_count;
+       for (u_int32_t i = 0; i < tmp_drop_dest_policy.entry_count; i++) {
+               struct necp_drop_dest_entry *tmp_drop_dest_entry = &tmp_drop_dest_policy.entries[i];
+               struct necp_drop_dest_entry *necp_drop_dest_entry = &necp_drop_dest_policy.entries[i];
+
+               memcpy(necp_drop_dest_entry, tmp_drop_dest_entry, sizeof(struct necp_drop_dest_entry));
+
+               necp_drop_dest_entry->order = necp_get_first_order_for_priority(necp_drop_dest_entry->level);
+       }
+       lck_rw_done(&necp_kernel_policy_lock);
+
+       return 0;
+}