Tasmota/lib/libesp32/berry_matter/src/embedded/Matter_IM.be
2024-06-16 18:42:23 +02:00

1290 lines
45 KiB
Plaintext

#
# Matter_IM.be - suppport for Matter Interaction Model
#
# Copyright (C) 2023 Stephan Hadinger & Theo Arends
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
import matter
#@ solidify:Matter_IM,weak
#################################################################################
# Matter_IM class
#################################################################################
class Matter_IM
var device
var subs_shop # subscriptions shop
var send_queue # list of IM_Message queued for sending as part of exchange-id
var read_request_solo # instance of ReadRequestMessage_solo to optimize single reads
var invoke_request_solo # instance of InvokeRequestMessage_solo to optimize single reads
var tlv_solo # instance of Matter_TLV_item for simple responses
def init(device)
self.device = device
self.send_queue = []
self.subs_shop = matter.IM_Subscription_Shop(self)
self.read_request_solo = matter.ReadRequestMessage_solo()
self.invoke_request_solo = matter.InvokeRequestMessage_solo()
self.tlv_solo = matter.TLV.Matter_TLV_item()
end
def process_incoming(msg)
var opcode = msg.opcode
# Fast-Track processing
# first pass is optimized for simple frequent messages and avoids complete decoding of TLV
if opcode == 0x02 # Read Request
var read_request_solo = self.read_request_solo.from_raw(msg.raw, msg.app_payload_idx)
if read_request_solo != nil
# log(f"MTR: process_incoming {read_request_solo=}")
return self.process_read_request_solo(msg, read_request_solo)
end
elif opcode == 0x08 # Invoke Request
var invoke_request_solo = self.invoke_request_solo.from_raw(msg.raw, msg.app_payload_idx)
# log(f"MTR: {invoke_request_solo=} {msg.raw[msg.app_payload_idx .. ].tohex()} {msg.app_payload_idx=} {msg.raw.tohex()}")
if invoke_request_solo != nil
return self.process_invoke_request_solo(msg, invoke_request_solo)
end
end
# log("MTR: received IM message " + matter.inspect(msg), 3)
var val = matter.TLV.parse(msg.raw, msg.app_payload_idx)
# log("MTR: IM TLV: " + str(val), 3)
# var InteractionModelRevision = val.findsubval(0xFF)
# log("MTR: InteractionModelRevision=" + (InteractionModelRevision != nil ? str(InteractionModelRevision) : "nil"), 4)
if opcode == 0x01 # Status Response
return self.process_status_response(msg, val)
elif opcode == 0x02 # Read Request
# self.send_ack_now(msg) # to improve latency, we don't automatically Ack on invoke request
return self.process_read_request(msg, val)
elif opcode == 0x03 # Subscribe Request
self.send_ack_now(msg)
return self.subscribe_request(msg, val)
elif opcode == 0x04 # Subscribe Response
return self.subscribe_response(msg, val)
elif opcode == 0x05 # Report Data
return self.report_data(msg, val)
elif opcode == 0x06 # Write Request
self.send_ack_now(msg)
return self.process_write_request(msg, val)
elif opcode == 0x07 # Write Response
return self.process_write_response(msg, val)
elif opcode == 0x08 # Invoke Request
# self.send_ack_now(msg) # to improve latency, we don't automatically Ack on invoke request
return self.process_invoke_request(msg, val)
elif opcode == 0x09 # Invoke Response
return self.process_invoke_response(msg, val)
elif opcode == 0x0A # Timed Request
return self.process_timed_request(msg, val)
end
return false
end
#############################################################
# check whether the ack received is of interest to any
# current exchange
#
# return `true` if handled
def process_incoming_ack(msg)
# check if there is an exchange_id interested in receiving this
var message = self.find_sendqueue_by_exchangeid(msg.exchange_id)
# log(format("MTR: process_incoming_ack exch=%i message=%i", msg.exchange_id, message != nil ? 1 : 0), 4)
if message
return message.ack_received(msg) # dispatch to IM_Message
end
return false
end
#############################################################
# send Ack response now and don't enqueue it
#
# returns `true` if packet could be sent
def send_ack_now(msg)
if msg == nil return end
msg.session._message_handler.send_encrypted_ack(msg, false #-not reliable-#)
end
#############################################################
# send enqueued responses
#
# self.send_queue is a list of <Matter_IM_Message>
#
def send_enqueued(responder)
var idx = 0
while idx < size(self.send_queue)
var message = self.send_queue[idx]
if !message.finish && message.ready
message.send_im(responder) # send message
end
if message.finish
log("MTR: remove IM message exch="+str(message.resp.exchange_id), 4)
self.send_queue.remove(idx)
else
idx += 1
end
end
end
#############################################################
# find in send_queue by exchangeid
#
def find_sendqueue_by_exchangeid(exchangeid)
if exchangeid == nil return nil end
var idx = 0
while idx < size(self.send_queue)
var message = self.send_queue[idx]
if message.get_exchangeid() == exchangeid
return message
end
idx += 1
end
return nil
end
#############################################################
# find in send_queue by exchangeid
#
def remove_sendqueue_by_exchangeid(exchangeid)
if exchangeid == nil return end
var idx = 0
while idx < size(self.send_queue)
if self.send_queue[idx].get_exchangeid() == exchangeid
self.send_queue.remove(idx)
else
idx += 1
end
end
end
#############################################################
# Remove any queued message that expired
#
def expire_sendqueue()
var idx = 0
while idx < size(self.send_queue)
var message = self.send_queue[idx]
if tasmota.time_reached(message.expiration)
message.reached_timeout()
self.send_queue.remove(idx)
else
idx += 1
end
end
return nil
end
#############################################################
# process IM 0x01 Status Response
#
# val is the TLV structure
# or raises an exception
# return true if we handled the response and ack, false instead
def process_status_response(msg, val)
var status = val.findsubval(0, 0xFF)
var message = self.find_sendqueue_by_exchangeid(msg.exchange_id)
if status == matter.SUCCESS
if message
return message.status_ok_received(msg) # re-arm the sending of next packets for the same exchange
else
log(format("MTR: >OK (%6i) exch=%i not found", msg.session.local_session_id, msg.exchange_id), 4) # don't show 'SUCCESS' to not overflow logs with non-information
end
else
# error
log(format("MTR: >Status ERROR = 0x%02X", status), 3)
if message
message.status_error_received(msg)
self.remove_sendqueue_by_exchangeid(msg.exchange_id)
end
end
return false # we did not ack the message, do it at higher level
end
#############################################################
# Inner code shared between read_attributes and subscribe_request
#
# query: `ReadRequestMessage` or `SubscribeRequestMessage`
def _inner_process_read_request(session, query, msg, no_log)
### Inner function to be iterated upon
# ret is the ReportDataMessage list to send back
# ctx is the context with endpoint/cluster/attribute
# direct is true if error is reported, false if error is silently ignored
#
# if `pi` is nil, just report the status for ctx.status
#
# should return true if answered, false if passing to next handler
def read_single_attribute(ret, pi, ctx, direct)
var TLV = matter.TLV
var attr_name
if tasmota.loglevel(3)
attr_name = matter.get_attribute_name(ctx.cluster, ctx.attribute)
attr_name = attr_name ? " (" + attr_name + ")" : ""
end
# Special case to report unsupported item, if pi==nil
ctx.status = nil # reset status, just in case
var res = (pi != nil) ? pi.read_attribute(session, ctx, self.tlv_solo) : nil
var found = true # stop expansion since we have a value
var a1_raw_or_list # contains either a bytes() buffer to append, or a list of bytes(), or nil
if res != nil
var res_str = ""
if !no_log
res_str = res.to_str_val() # get the value with anonymous tag before it is tagged, for logging
end
# check if too big to encode as a single packet
if (res.is_list || res.is_array) && res.encode_len() > matter.IM_ReportData.MAX_MESSAGE
# log(f"MTR: >>>>>> long response", 3)
a1_raw_or_list = [] # we return a list of block
var a1_raw = bytes(48)
var empty_list = TLV.Matter_TLV_array()
self.attributedata2raw(a1_raw, ctx, empty_list, false)
a1_raw_or_list.push(a1_raw)
# log(f"MTR: >>>>>> long response global DELETE {a1_raw.tohex()}", 3)
for elt:res.val
a1_raw = bytes(48)
# var list_item = TLV.Matter_TLV_array()
# list_item.val.push(elt)
self.attributedata2raw(a1_raw, ctx, elt, true #- add ListIndex:null -#)
# log(f"MTR: >>>>>> long response global ADD {a1_raw.tohex()}", 3)
a1_raw_or_list.push(a1_raw)
end
# log(f"MTR: >>>>>> long response global {a1_raw_or_list}", 3)
else
# normal encoding
# encode directly raw bytes()
a1_raw_or_list = bytes(48) # pre-reserve 48 bytes
self.attributedata2raw(a1_raw_or_list, ctx, res)
end
if tasmota.loglevel(3) && !no_log
log(f"MTR: >Read_Attr ({session.local_session_id:6i}) {ctx}{attr_name} - {res_str}", 3)
end
elif ctx.status != nil
if direct # we report an error only if a concrete direct read, not with wildcards
# encode directly raw bytes()
a1_raw_or_list = bytes(48) # pre-reserve 48 bytes
self.attributestatus2raw(a1_raw_or_list, ctx, ctx.status)
if tasmota.loglevel(3)
log(format("MTR: >Read_Attr (%6i) %s%s - STATUS: 0x%02X %s", session.local_session_id, str(ctx), attr_name, ctx.status, ctx.status == matter.UNSUPPORTED_ATTRIBUTE ? "UNSUPPORTED_ATTRIBUTE" : ""), 3)
end
end
else
if tasmota.loglevel(3) && !no_log
log(format("MTR: >Read_Attr (%6i) %s%s - IGNORED", session.local_session_id, str(ctx), attr_name), 3)
end
# ignore if content is nil and status is undefined
if direct
found = false
end
end
# a1_raw_or_list if either nil, bytes(), of list(bytes())
var idx = isinstance(a1_raw_or_list, list) ? 0 : nil # index in list, or nil if non-list
while a1_raw_or_list != nil
var elt = (idx == nil) ? a1_raw_or_list : a1_raw_or_list[idx] # dereference
if size(ret.attribute_reports) == 0
ret.attribute_reports.push(elt) # push raw binary instead of a TLV
else # already blocks present, see if we can add to the latest, or need to create a new block
var last_block = ret.attribute_reports[-1]
if size(last_block) + size(elt) <= matter.IM_ReportData.MAX_MESSAGE
# add to last block
last_block .. elt
else
ret.attribute_reports.push(elt) # push raw binary instead of a TLV
end
end
if idx == nil
a1_raw_or_list = nil # stop loop
else
idx += 1
if idx >= size(a1_raw_or_list)
a1_raw_or_list = nil # stop loop
end
end
end
# check if we still have enough room in last block
# if a1_raw_or_list # do we have bytes to add, and it's not zero size
# if size(ret.attribute_reports) == 0
# ret.attribute_reports.push(a1_raw_or_list) # push raw binary instead of a TLV
# else # already blocks present, see if we can add to the latest, or need to create a new block
# var last_block = ret.attribute_reports[-1]
# if size(last_block) + size(a1_raw_or_list) <= matter.IM_ReportData.MAX_MESSAGE
# # add to last block
# last_block .. a1_raw_or_list
# else
# ret.attribute_reports.push(a1_raw_or_list) # push raw binary instead of a TLV
# end
# end
# end
return found # return true if we had a match
end
var endpoints = self.device.get_active_endpoints()
# structure is `ReadRequestMessage` 10.6.2 p.558
var ctx = matter.Path()
ctx.msg = msg
var node_id = (msg != nil) ? msg.get_node_id() : nil
# prepare the response
var ret = matter.ReportDataMessage()
# ret.suppress_response = true
ret.attribute_reports = []
for q:query.attributes_requests
# need to do expansion here
ctx.endpoint = q.endpoint
ctx.cluster = q.cluster
ctx.attribute = q.attribute
ctx.fabric_filtered = query.fabric_filtered
ctx.status = matter.UNSUPPORTED_ATTRIBUTE #default error if returned `nil`
# expand endpoint
if ctx.endpoint == nil || ctx.cluster == nil || ctx.attribute == nil
# we need expansion, log first
if tasmota.loglevel(3)
if ctx.cluster != nil && ctx.attribute != nil
var attr_name = matter.get_attribute_name(ctx.cluster, ctx.attribute)
log(format("MTR: >Read_Attr (%6i) %s", session.local_session_id, str(ctx) + (attr_name ? " (" + attr_name + ")" : "")), 3)
else
log(format("MTR: >Read_Attr (%6i) %s", session.local_session_id, str(ctx)), 3)
end
end
end
# implement concrete expansion
self.device.process_attribute_expansion(ctx,
/ pi, ctx, direct -> read_single_attribute(ret, pi, ctx, direct)
)
end
# tasmota.log(f">>>: event_1")
var event_requests = query.event_requests
var event_filters = query.event_filters
var event_no_min = nil # do we have a filter for minimum event_no (int64 or nil)
if event_requests # if not `nil` and not empty list
# read event minimum
if event_filters
for filter: event_filters # filter is an instance of `EventFilterIB`
tasmota.log(f"MTR: EventFilter {filter=} {node_id=}", 3)
var filter_node = int64.toint64(filter.node) # nil or int64
if filter_node # there is a filter on node-id
if filter.node.tobytes() != node_id # the node id doesn't match
tasmota.log(f"MTR: node_id filter {filter_node.tobytes().tohex()} doesn't match {node_id.tohex()}")
continue
end
# specified minimum value
var new_event_no_min = int64.toint64(filter.event_min)
if (event_no_min != nil) || (event_no_min < new_event_no_min)
event_no_min = new_event_no_min
end
end
end
end
# event_no_min is either `nil` or has an `int64` value
ret.event_reports = []
for q: event_requests
# need to do expansion here
ctx.endpoint = q.endpoint
ctx.cluster = q.cluster
ctx.attribute = q.event
#TODO
tasmota.log(f"MTR: >Read_Event({session.local_session_id:%6i}) {ctx}", 3)
end
end
# tasmota.log("MTR: ReportDataMessage=" + str(ret), 3)
# tasmota.log("MTR: ReportDataMessageTLV=" + str(ret.to_TLV()), 3)
return ret
end
#############################################################
# path2raw
#
# Encodes endpoint/cluster/attribute as `AttributePathIB` elements
# Takes sub-tag
#
# 1 = AttributePathIB
# 0 = EnableTagCompression bool opt
# 1 = Node
# 2 = Endpoint
# 3 = Cluste
# 4 = Attribute
# 5 = ListIndex (opt)
#
# 3701 1 = LIST
# 2402 01 2 = 1U (U1)
# 2403 39 3 = 0x39U (U1)
# 2404 11 4 = 0x11U (U1)
# [OPTIONAL ListIndex]
# 3405 5 = NULL
# 18
def path2raw(raw, ctx, sub_tag, list_index_null)
# open struct
raw.add(0x37, 1) # add 37
raw.add(sub_tag, 1) # add sub_tag
# add endpoint
if ctx.endpoint <= 0xFF # endpoint is 16 bits max
raw.add(0x2402, -2) # add 2402
raw.add(ctx.endpoint, 1)
else
raw.add(0x2502, -2) # add 2502
raw.add(ctx.endpoint, 2)
end
# add cluster
if ctx.cluster <= 0xFF # cluster is 32 bits max
raw.add(0x2403, -2) # add 2403
raw.add(ctx.cluster, 1)
elif ctx.cluster <= 0xFFFF
raw.add(0x2503, -2) # add 2503
raw.add(ctx.cluster, 2)
else
raw.add(0x2603, -2) # add 2603
raw.add(ctx.cluster, 4)
end
# add attribute
if ctx.attribute <= 0xFF # cluster is 32 bits max
raw.add(0x2404, -2) # add 2404
raw.add(ctx.attribute, 1)
elif ctx.attribute <= 0xFFFF
raw.add(0x2504, -2) # add 2504
raw.add(ctx.attribute, 2)
else
raw.add(0x2604, -2) # add 2604
raw.add(ctx.attribute, 4)
end
# do we add ListIndex: null
if list_index_null
raw.add(0x3405, -2) # add 3405
end
# close
raw.add(0x18, 1) # add 18
end
#############################################################
# attributedata2raw
#
# generate a raw version of AttributeDataIB
#
# Typical answer
#
# AttributeReportIB
# 0 = AttributeStatusIB
# 1 = AttributeDataIB
# 0 = DataVersion U1
# 1 = AttributePathIB
# 0 = EnableTagCompression bool opt
# 1 = Node
# 2 = Endpoint
# 3 = Cluste
# 4 = Attribute
# 5 = ListIndex (opt)
# 2 = Data
#
# 153601.15350124000137012402012403392404111829021818.1824FF0118
# 15350124000137012402012403392404111829021818
# 1535012400013701
# 240201240339240411
# 1829021818
#
# 15
# 3501 1 = {}
# 2400 01 0 = 1U (U1)
# 3701 1 = LIST
# 2402 01 2 = 1U (U1)
# 2403 39 3 = 0x39U (U1)
# 2404 11 4 = 0x11U (U1)
# [OPTIONAL ListIndex]
# 3405 5 = NULL
#
# 18
# 2902 2 = True
# 18
# 18
def attributedata2raw(raw, ctx, val, list_index_null)
raw.add(0x15350124, -4) # add 15350124
raw.add(0x0001, -2) # add 0001
self.path2raw(raw, ctx, 0x01, list_index_null)
# add value with tag 2
val.tag_sub = 2
val.tlv2raw(raw)
# close 2 structs
raw.add(0x1818, -2)
end
#############################################################
# attributedata2raw
#
# generate a raw version of AttributeStatusIB
#
#
# Typical answer
#
# AttributeReportIB
# 0 = AttributeStatusIB
# 0 = AttributePathIB
# 0 = EnableTagCompression bool opt
# 1 = Node
# 2 = Endpoint
# 3 = Cluste
# 4 = Attribute
# 5 = ListIndex (opt)
# 1 = StatusIB
# 0 = Status (u1)
# 1 = ClusterStatus (u1)
# 1 = AttributeDataIB
#
# 15360115350037002402012403022404031835012400041818181824FF0118
# 153601 1535003700 - 2402012403022404031835012400041818181824FF0118
#
# 15
# 3601
#
# 15
# 3500 0 = struct
# 3700 0 = list
# 240201 2 = 1U endpoint
# 240302 3 = 2U cluster
# 240403 4 = 3U attribute
# 18
# 3501 1 = struct
# 240004 0 = 4U status
# 18
# 18
# 18
#
# 18
# 24FF01
# 18
def attributestatus2raw(raw, ctx, status)
raw.add(0x15, 1) # add 15
raw.add(0x3500, -2) # add 3500
self.path2raw(raw, ctx, 0x00)
raw.add(0x3501, -2) # add 3501 for status
# status
if ctx.status <= 255
raw.add(0x2400, -2) # add 2400
raw.add(ctx.status, 1)
else
raw.add(0x2500, -2) # add 2500
raw.add(ctx.status, 2)
end
# close
raw.add(0x1818, -2) # add 1818
raw.add(0x18, 1) # add 18
end
#############################################################
# attributedata2raw
#
# generate a raw version of InvokeResponseIB()
# Typical answer
#
# 1535013700240011240122240244183501240000181818
#
# 0 = CommandDataIB
# 0 = CommandPathIB
# 0 = endpoint u2
# 1 = cluster u4
# 2 = command u4
# 1 = <fields>
# 1 = CommandStatusIB
# 0 = CommandPathIB
# 0 = endpoint u2
# 1 = cluster u4
# 2 = command u4
# 1 = StatusIB
# 0 = status u1
# 1 = ClusterStatus u1
#
# 1535013700240011240122240244183501240000181818
# 15
# 3501 1 = struct
# 3700 0 = list
# 240011 0 = endpoint
# 240122 1 = cluster
# 240244 2 = command
# 18
# 3501 1 = struct
# 240000 0 = 0 (status)
# 18
# 18
# 18
#
# 1535003700240011240122240244182401031818
# 15
# 3500 0 = struct
# 3700 0 = list
# 240011 0 = endpoint
# 240122 1 = cluster
# 240244 2 = command
# 18
# 240103 1 = <field>
# 18
# 18
def invokeresponse2raw(raw, ctx, val)
raw.add(0x15, 1) # add 15
if val == nil
raw.add(0x3501, -2) # add 3500
else
raw.add(0x3500, -2) # add 3500
end
raw.add(0x3700, -2) # add 3700
# add endpoint
if ctx.endpoint <= 0xFF # endpoint is 16 bits max
raw.add(0x2400, -2) # add 2400
raw.add(ctx.endpoint, 1)
else
raw.add(0x2500, -2) # add 2500
raw.add(ctx.endpoint, 2)
end
# add cluster
if ctx.cluster <= 0xFF # cluster is 32 bits max
raw.add(0x2401, -2) # add 2401
raw.add(ctx.cluster, 1)
elif ctx.cluster <= 0xFFFF
raw.add(0x2501, -2) # add 2501
raw.add(ctx.cluster, 2)
else
raw.add(0x2601, -2) # add 2601
raw.add(ctx.cluster, 4)
end
# add attribute
if ctx.command <= 0xFF # cluster is 32 bits max
raw.add(0x2402, -2) # add 2402
raw.add(ctx.command, 1)
elif ctx.command <= 0xFFFF
raw.add(0x2502, -2) # add 2502
raw.add(ctx.command, 2)
else
raw.add(0x2602, -2) # add 2602
raw.add(ctx.command, 4)
end
raw.add(0x18, 1) # add 18
# either value or statuc
if val == nil
var status = ctx.status
if status == nil status = matter.SUCCESS end
raw.add(0x3501, -2) # add 3501
raw.add(0x2400, -2) # add 2400
raw.add(ctx.status, 1) # add status:1
raw.add(0x18, 1) # add 18
else
val.tag_sub = 1 # set sub_tag for reponse
val.tlv2raw(raw)
end
# close
raw.add(0x1818, -2) # add 1818
end
#############################################################
# process IM 0x02 Read Request
#
# val is the TLV structure
#
# returns `true` if processed, `false` if silently ignored,
# or raises an exception
def process_read_request(msg, val)
matter.profiler.log("read_request_start")
var query = matter.ReadRequestMessage().from_TLV(val)
if query.attributes_requests != nil
var ret = self._inner_process_read_request(msg.session, query, msg)
self.send_report_data(msg, ret)
end
return true # always consider succesful even if empty
end
#############################################################
# process IM 0x02 Read Request - solo attribute
#
# ctx is the decoded context (cluster/attribute)
#
# returns `true` if processed, `false` if silently ignored,
# or raises an exception
def process_read_request_solo(msg, ctx)
# prepare fallback error
ctx.status = matter.INVALID_ACTION
ctx.msg = msg
# find pi for this endpoint/cluster/attribute
var pi = self.device.resolve_attribute_read_solo(ctx)
var res = nil
var raw # this is the bytes() block we need to add to response (or nil)
if pi != nil
ctx.status = matter.UNSUPPORTED_ATTRIBUTE # new fallback error
res = pi.read_attribute(msg.session, ctx, self.tlv_solo)
end
matter.profiler.log("read_request_solo read done")
if res != nil
# check if the payload is a complex structure and too long to fit in a single response packet
if (res.is_list || res.is_array) && (res.encode_len() > matter.IM_ReportData.MAX_MESSAGE)
# revert to standard
# the attribute will be read again, but it's hard to avoid it
res = nil # indicated to GC that we don't need it again
log(f"MTR: Response to big, revert to non-solo", 3)
var val = matter.TLV.parse(msg.raw, msg.app_payload_idx)
return self.process_read_request(msg, val)
end
# encode directly raw bytes()
raw = bytes(48) # pre-reserve 48 bytes
raw.add(0x15, 1) # add 15
raw.add(0x3601, -2) # add 3601
self.attributedata2raw(raw, ctx, res)
# add suffix 1824FF0118
raw.add(0x1824FF01, -4) # add 1824FF01
raw.add(0x18, 1) # add 18
elif ctx.status != nil
# encode directly raw bytes()
raw = bytes(48) # pre-reserve 48 bytes
raw.add(0x15, 1) # add 15
raw.add(0x3601, -2) # add 3601
self.attributestatus2raw(raw, ctx, ctx.status)
# add suffix 1824FF0118
raw.add(0x1824FF01, -4) # add 1824FF01
raw.add(0x18, 1) # add 18
else
log(f"MTR: >Read_Attr ({msg.session.local_session_id:6i}) {ctx} - IGNORED", 3)
return false
end
# send packet
var resp = msg.build_response(0x05 #-Report Data-#, true)
var responder = self.device.message_handler
var msg_raw = msg.raw
msg_raw.clear()
resp.encode_frame(raw, msg_raw) # payload in cleartext
resp.encrypt()
if tasmota.loglevel(4)
log(format("MTR: <snd (%6i) id=%i exch=%i rack=%s", resp.session.local_session_id, resp.message_counter, resp.exchange_id, resp.ack_message_counter), 4)
end
responder.send_response_frame(resp)
# postpone lengthy operations after sending back response
matter.profiler.log("RESPONSE SENT")
var attr_name
if tasmota.loglevel(3)
attr_name = matter.get_attribute_name(ctx.cluster, ctx.attribute)
attr_name = attr_name ? " (" + attr_name + ")" : ""
end
if res != nil
if tasmota.loglevel(3)
var res_str = res.to_str_val() # get the value with anonymous tag before it is tagged, for logging
log(f"MTR: >Read_Attr1({msg.session.local_session_id:6i}) {ctx}{attr_name} - {res_str}", 3)
# log(f"MTR: {res.tlv2raw().tohex()}", 3)
end
# if matter.profiler.active && tasmota.loglevel(3)
# log(f"MTR: {raw=}", 3) # TODO remove before flight
# end
elif ctx.status != nil
var unsupported_attribute = (ctx.status == matter.UNSUPPORTED_ATTRIBUTE ? "UNSUPPORTED_ATTRIBUTE" : "")
if tasmota.loglevel(3)
log(f"MTR: >Read_Attr1({msg.session.local_session_id:6i}) {ctx}{attr_name} - STATUS: 0x{ctx.status:02X} {unsupported_attribute}", 3)
end
# if matter.profiler.active && tasmota.loglevel(3)
# log(f"MTR: {raw=}", 3) # TODO remove before flight
# end
else
if tasmota.loglevel(3)
log(f"MTR: >Read_Attr1({msg.session.local_session_id:6i}) {ctx}{attr_name} - IGNORED", 3)
end
end
return true # always consider succesful even if empty
end
#############################################################
# process IM 0x03 Subscribe Request
#
def subscribe_request(msg, val)
var query = matter.SubscribeRequestMessage().from_TLV(val)
if !query.keep_subscriptions
self.subs_shop.remove_by_session(msg.session) # if `keep_subscriptions`, kill all subscriptions from current session
end
# log("MTR: received SubscribeRequestMessage=" + str(query), 3)
var sub = self.subs_shop.new_subscription(msg.session, query)
# expand a string with all attributes requested
var attr_req = []
var ctx = matter.Path()
ctx.msg = msg
for q:query.attributes_requests
ctx.endpoint = q.endpoint
ctx.cluster = q.cluster
ctx.attribute = q.attribute
attr_req.push(str(ctx))
end
log(format("MTR: >Subscribe (%6i) %s (min=%i, max=%i, keep=%i) sub=%i fabric_filtered=%s attr_req=%s event_req=%s",
msg.session.local_session_id, attr_req.concat(" "), sub.min_interval, sub.max_interval, query.keep_subscriptions ? 1 : 0,
sub.subscription_id, query.fabric_filtered,
query.attributes_requests != nil ? size(query.attributes_requests) : "-",
query.event_requests != nil ? size(query.event_requests) : "-"), 3)
var ret = self._inner_process_read_request(msg.session, query, msg, !self.device.debug #-log only if debug enabled-#)
# ret is of type `Matter_ReportDataMessage`
ret.subscription_id = sub.subscription_id # enrich with subscription id TODO
self.send_subscribe_response(msg, ret, sub)
return true
end
#############################################################
# process IM 0x08 Invoke Request
#
# val is the TLV structure
# returns `true` if processed, `false` if silently ignored,
# or raises an exception
def process_invoke_request(msg, val)
# import debug
# structure is `ReadRequestMessage` 10.6.2 p.558
# log("MTR: IM:invoke_request processing start", 4)
matter.profiler.log("invoke_request_start")
var ctx = matter.Path()
ctx.msg = msg
var query = matter.InvokeRequestMessage().from_TLV(val)
if query.invoke_requests != nil
# prepare the response
var ret = matter.InvokeResponseMessage()
ret.suppress_response = false
ret.invoke_responses = []
for q:query.invoke_requests
ctx.endpoint = q.command_path.endpoint
ctx.cluster = q.command_path.cluster
ctx.command = q.command_path.command
ctx.status = matter.UNSUPPORTED_COMMAND #default error if returned `nil`
var cmd_name = matter.get_command_name(ctx.cluster, ctx.command)
var ctx_str = str(ctx) # keep string before invoking, it is modified by response
var res = self.device.invoke_request(msg.session, q.command_fields, ctx)
matter.profiler.log("COMMAND DONE")
var params_log = (ctx.log != nil) ? "(" + str(ctx.log) + ") " : ""
log(format("MTR: >Command (%6i) %s %s %s", msg.session.local_session_id, ctx_str, cmd_name ? cmd_name : "", params_log), 3)
# log("MTR: Perf/Command = " + str(debug.counters()), 4)
ctx.log = nil
var raw = bytes(32)
# var a1 = matter.InvokeResponseIB()
if res == true || ctx.status == matter.SUCCESS # special case, just respond ok
ctx.status = matter.SUCCESS
self.invokeresponse2raw(raw, ctx, nil)
ret.invoke_responses.push(raw)
if tasmota.loglevel(3)
log(f"MTR: <Replied ({msg.session.local_session_id:6i}) OK exch={msg.exchange_id:i}", 3)
end
elif res != nil
self.invokeresponse2raw(raw, ctx, res)
ret.invoke_responses.push(raw)
cmd_name = matter.get_command_name(ctx.cluster, ctx.command)
if !cmd_name cmd_name = "" end
if tasmota.loglevel(3)
log(f"MTR: <Replied ({msg.session.local_session_id:6i}) {ctx} {cmd_name}", 3)
end
elif ctx.status != nil
self.invokeresponse2raw(raw, ctx, nil)
ret.invoke_responses.push(raw)
if tasmota.loglevel(3)
log(f"MTR: <Replied ({msg.session.local_session_id:6i}) Status=0x{ctx.status:02X} exch={msg.exchange_id:i}", 3)
end
else
if tasmota.loglevel(3)
log(f"MTR: _Ignore ({msg.session.local_session_id:6i}) exch={msg.exchange_id:i}", 3)
end
# ignore if content is nil and status is undefined
end
end
if size(ret.invoke_responses) > 0
self.send_invoke_response(msg, ret)
else
return false # we don't send anything, hence the responder sends a simple packet ack
end
return true
end
end
#############################################################
# process IM 0x08 Invoke Request
#
# val is the TLV structure
# returns `true` if processed, `false` if silently ignored,
# or raises an exception
def process_invoke_request_solo(msg, ctx)
# import debug
matter.profiler.log("invoke_request_solo_start")
ctx.msg = msg
ctx.status = matter.UNSUPPORTED_COMMAND #default error if returned `nil`
var cmd_name = matter.get_command_name(ctx.cluster, ctx.command)
var ctx_str = str(ctx) # keep string before invoking, it is modified by response
var res = self.device.invoke_request(msg.session, ctx.command_fields, ctx)
matter.profiler.log("COMMAND DONE")
var params_log = (ctx.log != nil) ? "(" + str(ctx.log) + ") " : ""
if tasmota.loglevel(3)
log(format("MTR: >Command1 (%6i) %s %s %s", msg.session.local_session_id, ctx_str, cmd_name ? cmd_name : "", params_log), 3)
end
# log("MTR: Perf/Command = " + str(debug.counters()), 4)
ctx.log = nil
var raw = bytes(48)
# prefix 1528003601
raw.add(0x15280036, -4) # add 15280036
raw.add(0x01, 1) # add 01
if res == true || ctx.status == matter.SUCCESS # special case, just respond ok
ctx.status = matter.SUCCESS
self.invokeresponse2raw(raw, ctx, nil)
if tasmota.loglevel(3)
log(f"MTR: <Replied ({msg.session.local_session_id:6i}) OK exch={msg.exchange_id:i}", 3)
end
elif res != nil
self.invokeresponse2raw(raw, ctx, res)
if !cmd_name cmd_name = "" end
if tasmota.loglevel(3)
log(f"MTR: <Replied ({msg.session.local_session_id:6i}) {ctx} {cmd_name}", 3)
end
elif ctx.status != nil
self.invokeresponse2raw(raw, ctx, nil)
if tasmota.loglevel(3)
log(f"MTR: <Replied ({msg.session.local_session_id:6i}) Status=0x{ctx.status:02X} exch={msg.exchange_id:i}", 3)
end
else
if tasmota.loglevel(3)
log(f"MTR: _Ignore ({msg.session.local_session_id:6i}) exch={msg.exchange_id:i}", 3)
end
# ignore if content is nil and status is undefined
return false
end
# add suffix 1824FF0118
raw.add(0x1824FF01, -4) # add 1824FF01
raw.add(0x18, 1) # add 18
# log(f"MTR: raw={raw.tohex()}", 3)
var resp = msg.build_response(0x09 #-Invoke Response-#, true)
var responder = self.device.message_handler
var msg_raw = msg.raw
msg_raw.clear()
resp.encode_frame(raw, msg_raw) # payload in cleartext
resp.encrypt()
responder.send_response_frame(resp)
matter.profiler.log("RESPONSE SENT")
return true
end
#############################################################
# process IM 0x04 Subscribe Response
#
def subscribe_response(msg, val)
var query = matter.SubscribeResponseMessage().from_TLV(val)
# log("MTR: received SubscribeResponsetMessage=" + str(query), 4)
return false
end
#############################################################
# process IM 0x05 ReportData
#
def report_data(msg, val)
var query = matter.ReportDataMessage().from_TLV(val)
# log("MTR: received ReportDataMessage=" + str(query), 4)
return false
end
#############################################################
# process IM 0x06 Write Request
#
def process_write_request(msg, val)
var query = matter.WriteRequestMessage().from_TLV(val)
# log("MTR: received WriteRequestMessage=" + str(query), 3)
var suppress_response = query.suppress_response
# var timed_request = query.timed_request # TODO not supported
# var more_chunked_messages = query.more_chunked_messages # TODO not supported
var endpoints = self.device.get_active_endpoints()
### Inner function to be iterated upon
# ret is the ReportDataMessage list to send back
# ctx is the context with endpoint/cluster/attribute
# val is the TLV object containing the value
# direct is true if error is reported, false if error is silently ignored
#
# if `pi` is nil, just report the status for ctx.status
#
# should return true if answered, false if failed
def write_single_attribute(ret, pi, ctx, write_data, direct)
var attr_name = matter.get_attribute_name(ctx.cluster, ctx.attribute)
attr_name = attr_name ? " (" + attr_name + ")" : ""
# log(format("MTR: Read Attribute " + str(ctx) + (attr_name ? " (" + attr_name + ")" : ""), 2)
# Special case to report unsupported item, if pi==nil
ctx.status = matter.UNSUPPORTED_WRITE
var res = (pi != nil) ? pi.write_attribute(msg.session, ctx, write_data) : nil
if res ctx.status = matter.SUCCESS end # if the cb returns true, the request was processed
if ctx.status != nil
if direct
var a1 = matter.AttributeStatusIB()
a1.path = matter.AttributePathIB()
a1.status = matter.StatusIB()
a1.path.endpoint = ctx.endpoint
a1.path.cluster = ctx.cluster
a1.path.attribute = ctx.attribute
a1.status.status = ctx.status
ret.write_responses.push(a1)
log(format("MTR: Write_Attr %s%s - STATUS: 0x%02X %s", str(ctx), attr_name, ctx.status, ctx.status == matter.SUCCESS ? "SUCCESS" : ""), (ctx.endpoint != 0) ? 2 : 3)
return true
end
else
log(format("MTR: Write_Attr %s%s - IGNORED", str(ctx), attr_name), 3)
# ignore if content is nil and status is undefined
end
end
# structure is `ReadRequestMessage` 10.6.2 p.558
# log("MTR: IM:write_request processing start", 4)
var ctx = matter.Path()
ctx.msg = msg
if query.write_requests != nil
# prepare the response
var ret = matter.WriteResponseMessage()
# ret.suppress_response = true
ret.write_responses = []
for q:query.write_requests # q is AttributeDataIB
var path = q.path
var write_data = q.data
# need to do expansion here
ctx.endpoint = path.endpoint
ctx.cluster = path.cluster
ctx.attribute = path.attribute
ctx.status = matter.UNSUPPORTED_ATTRIBUTE #default error if returned `nil`
# return an error if the expansion is illegal
if ctx.cluster == nil || ctx.attribute == nil
# force INVALID_ACTION reporting
ctx.status = matter.INVALID_ACTION
write_single_attribute(ret, nil, ctx, nil, true)
continue
end
# expand endpoint
if tasmota.loglevel(3) && (ctx.endpoint == nil)
# we need expansion, log first
var attr_name = matter.get_attribute_name(ctx.cluster, ctx.attribute)
log("MTR: Write_Attr " + str(ctx) + (attr_name ? " (" + attr_name + ")" : ""), 3)
end
# implement concrete expansion
self.device.process_attribute_expansion(ctx,
/ pi, ctx, direct -> write_single_attribute(ret, pi, ctx, write_data, direct)
)
end
# log("MTR: ReportWriteMessage=" + str(ret), 4)
# log("MTR: ReportWriteMessageTLV=" + str(ret.to_TLV()), 3)
# send the reponse that may need to be chunked if too large to fit in a single UDP message
if !suppress_response
self.send_write_response(msg, ret)
end
end
return true
end
#############################################################
# process IM 0x07 Write Response
#
def process_write_response(msg, val)
var query = matter.WriteResponseMessage().from_TLV(val)
# log("MTR: received WriteResponseMessage=" + str(query), 4)
return false
end
#############################################################
# process IM 0x09 Invoke Response
#
def process_invoke_response(msg, val)
var query = matter.InvokeResponseMessage().from_TLV(val)
# log("MTR: received InvokeResponseMessage=" + str(query), 4)
return false
end
#############################################################
# process IM 0x0A Timed Request
#
def process_timed_request(msg, val)
var query = matter.TimedRequestMessage().from_TLV(val)
# log("MTR: received TimedRequestMessage=" + str(query), 3)
log(format("MTR: >Command (%6i) TimedRequest=%i", msg.session.local_session_id, query.timeout), 3)
# Send success status report
self.send_status(msg, matter.SUCCESS)
return true
end
#############################################################
# send regular update for data subscribed
#
def send_subscribe_update(sub)
var session = sub.session
# create a fake read request to feed to the ReportData
var fake_read = matter.ReadRequestMessage()
fake_read.fabric_filtered = false
fake_read.attributes_requests = []
for ctx: sub.updates
var p1 = matter.AttributePathIB()
p1.endpoint = ctx.endpoint
p1.cluster = ctx.cluster
p1.attribute = ctx.attribute
fake_read.attributes_requests.push(p1)
end
log(format("MTR: <Sub_Data (%6i) sub=%i", session.local_session_id, sub.subscription_id), 3)
sub.is_keep_alive = false # sending an actual data update
var ret = self._inner_process_read_request(session, fake_read, nil #-no msg-#)
ret.suppress_response = false
ret.subscription_id = sub.subscription_id
var report_data_msg = matter.IM_ReportDataSubscribed(session._message_handler, session, ret, sub)
self.send_queue.push(report_data_msg) # push message to queue
self.send_enqueued(session._message_handler) # and send queued messages now
end
#############################################################
# send regular update for data subscribed
#
def send_subscribe_heartbeat(sub)
var session = sub.session
log(format("MTR: <Sub_Alive (%6i) sub=%i", session.local_session_id, sub.subscription_id), 3)
sub.is_keep_alive = true # sending keep-alive
# prepare the response
var ret = matter.ReportDataMessage()
ret.suppress_response = true
ret.subscription_id = sub.subscription_id
var report_data_msg = matter.IM_SubscribedHeartbeat(session._message_handler, session, ret, sub)
self.send_queue.push(report_data_msg) # push message to queue
self.send_enqueued(session._message_handler) # and send queued messages now
end
#############################################################
# send_status
#
def send_status(msg, status)
self.send_queue.push(matter.IM_Status(msg, status))
end
#############################################################
# send_invoke_response
#
def send_invoke_response(msg, data)
self.send_queue.push(matter.IM_InvokeResponse(msg, data))
end
#############################################################
# send_invoke_response
#
def send_write_response(msg, data)
self.send_queue.push(matter.IM_WriteResponse(msg, data))
end
#############################################################
# send_report_data
#
def send_report_data(msg, data)
self.send_queue.push(matter.IM_ReportData(msg, data))
end
#############################################################
# send_report_data
#
def send_subscribe_response(msg, data, sub)
self.send_queue.push(matter.IM_SubscribeResponse(msg, data, sub))
end
#############################################################
# placeholder, nothing to run for now
def every_second()
self.expire_sendqueue()
end
#############################################################
# dispatch every 250ms click to sub-objects that need it
def every_250ms()
self.subs_shop.every_250ms()
end
end
matter.IM = Matter_IM
#-
# Unit tests
-#