Fix ZGP handling in ZNP

zigpy_znp/commands/zdo.py

@@ -12,7 +12,6 @@
 
 import zigpy_znp.types as t
 
-
 class SecurityEntry(t.FixedList, item_type=t.uint8_t, length=5):
     pass
 
@@ -797,10 +796,10 @@ class ZDO(t.CommandsBase, subsystem=t.Subsystem.ZDO):
         0x49,
         req_schema=(
             t.Param("Endpoint", t.uint8_t, "Endpoint to look for"),
-            # this parameter does not make sense
-            t.Param("Groups", t.uint16_t, "List to hold group IDs"),
         ),
-        rsp_schema=(t.Param("Groups", GroupIdList, "List of Group IDs"),),
+        rsp_schema=(
+            t.Param("Groups", GroupIdList, "List of Group IDs"),
+        ),
     )
 
     # handle the ZDO extension find group message
@@ -811,7 +810,11 @@ class ZDO(t.CommandsBase, subsystem=t.Subsystem.ZDO):
             t.Param("Endpoint", t.uint8_t, "Endpoint to look for"),
             t.Param("GroupId", t.GroupId, "ID to look for group"),
         ),
-        rsp_schema=(t.Param("Group", t.Bytes, "Group information"),),
+        rsp_schema=(
+            t.Param("Status", t.Status, "Status is either Success (0) or Failure (1)"),
+            t.Param("GroupId", t.GroupId, "Found Group ID"),
+            t.Param("GroupName", t.Bytes, "Group name"),
+        ),
     )
 
     # handle the ZDO extension add group message

zigpy_znp/commands/zgp.py

@@ -97,7 +97,7 @@ class ZGP(t.CommandsBase, subsystem=t.Subsystem.ZGP):
     # This message provides a mechanism for dGP stub to request security data from the
     # Green Power EndPoint in the host processor
     SecReq = t.CommandDef(
-        t.CommandType.AREQ,
+        t.CommandType.SREQ,
         0x03,
         req_schema=(
             t.Param(
@@ -127,14 +127,15 @@ class ZGP(t.CommandsBase, subsystem=t.Subsystem.ZGP):
                 "Handle", t.uint8_t, "dGP stub handle to match req to confirmation"
             ),
         ),
+        rsp_schema=t.STATUS_SCHEMA,
     )
 
     # This message provides a mechanism for identifying and conveying a received
     # GPDF to the Green Power EndPoint in the host processor
     DataInd = t.CommandDef(
         t.CommandType.AREQ,
         0x04,
-        req_schema=(
+        rsp_schema=(
             t.Param("Status", t.uint8_t, "The status code as returned by the dGP stub"),
             t.Param(
                 "RSSI",

zigpy_znp/zigbee/application.py

@@ -416,6 +416,14 @@ def _bind_callbacks(self) -> None:
             self.on_intentionally_unhandled_message,
         )
 
+        self._znp.callback_for_response(
+            c.ZGP.DataInd.Callback(partial=True),
+            self.on_zgp_data_ind
+        )
+
+    def on_zgp_data_ind(self, msg: c.ZGP.DataInd.Callback) -> None:
+        pass
+
     def on_intentionally_unhandled_message(self, msg: t.CommandBase) -> None:
         """
         Some commands are unhandled but frequently sent by devices on the network. To
@@ -691,6 +699,9 @@ def _find_endpoint(self, dst_ep: int, profile: int, cluster: int) -> int:
 
         if dst_ep == ZDO_ENDPOINT:
             return ZDO_ENDPOINT
+
+        if profile == zigpy.profiles.zgp.PROFILE_ID:
+            return zigpy.profiles.zgp.GREENPOWER_ENDPOINT_ID
 
         # Newer Z-Stack releases ignore profiles and will work properly with endpoint 1
         if (

zigpy_znp/zigbee/device.py

@@ -1,19 +1,48 @@
 from __future__ import annotations
 
 import logging
+from typing import Any, Coroutine
+from zigpy.zcl.foundation import Status as ZCLStatus, Status
 
 import zigpy.zdo
+import zigpy.endpoint
 import zigpy.device
 import zigpy.application
+from zigpy_znp.api import ZNP
+import zigpy_znp.commands as c
+import zigpy_znp.types as t
 
 LOGGER = logging.getLogger(__name__)
 
+class ZNPEndpoint(zigpy.endpoint.Endpoint):
+    async def add_to_group(self, grp_id: int, name: str | None = None) -> ZCLStatus:
+        znp: ZNP = self.device.application._znp
+        if name is None:
+            name = ""
 
+        result = await znp.request(c.ZDO.ExtFindGroup.Req(Endpoint=self.endpoint_id, GroupId=grp_id))
+        if result.Status == t.Status.FAILURE:
+            result = await znp.request(c.ZDO.ExtAddGroup.Req(Endpoint=self.endpoint_id, GroupId=grp_id, GroupName=t.CharacterString(name)))
+        if result.Status == t.Status.FAILURE:
+            return ZCLStatus.FAILURE
+        group = self.device.application.groups.add_group(grp_id, name)
+        group.add_member(self)
+        return ZCLStatus.SUCCESS
+
+    async def remove_from_group(self, grp_id: int) -> ZCLStatus:
+        znp: ZNP = self.device.application._znp
+        result = await znp.request(c.ZDO.ExtRemoveGroup.Req(Endpoint=self.endpoint_id, GroupId=grp_id))
+        if result.Status == t.Status.FAILURE:
+            return ZCLStatus.FAILURE
+        if grp_id in self.device.application.groups:
+            self.device.application.groups[grp_id].remove_member(self)
+        return ZCLStatus.SUCCESS
+
+
 class ZNPCoordinator(zigpy.device.Device):
     """
     Coordinator zigpy device that keeps track of our endpoints and clusters.
     """
-
     @property
     def manufacturer(self):
         return "Texas Instruments"
@@ -22,6 +51,11 @@ def manufacturer(self):
     def model(self):
         return "Coordinator"
 
+    def add_endpoint(self, endpoint_id) -> zigpy.endpoint.Endpoint:
+        ep = ZNPEndpoint(self, endpoint_id)
+        self.endpoints[endpoint_id] = ep
+        return ep
+
     def request(
         self,
         profile,