Modified code to support 1.95.0-SS-17-release

device/pensando/arm64-elba-asic-flash128-r0/pre_reboot_hook

@@ -21,6 +21,12 @@ LOG_MSG "Dropping caches..."
 LOG_MSG 3 > /proc/sys/vm/drop_caches
 LOG_MSG "Cache cleared."
 
+sleep 2
 LOG_MSG "Pre-reboot flush and cache clear complete."
+
+sleep 2
+LOG_MSG "Writing pre-shutdown success to CPLD reg"
+docker exec "$(cat /host/dpu-docker-info/name)" /nic/bin/cpldapp -w 0xd 200
+
 LOG_MSG "Platform specific pre-shutdown is successful"
 

device/pensando/arm64-elba-asic-flash128-r0/system_health_monitoring_config.json

@@ -2,7 +2,7 @@
     "services_to_ignore": ["vlanmgrd", "vxlanmgrd"],
     "devices_to_ignore": ["psu","fan"],
     "user_defined_checkers": ["fetch_dpu_status"],
-    "polling_interval": 60,
+    "polling_interval": 20,
     "led_color": {
         "fault": "amber",
         "normal": "green",

files/dsc/dpu.init

@@ -41,7 +41,8 @@ function start_polaris()
     mkdir -p $HOST_DIR_POLARIS/obfl
     mkdir -p $HOST_DIR_POLARIS/data
     mkdir -p $HOST_DIR_POLARIS/share
-    mount -t tmpfs -o size=20M sharetmps $HOST_DIR_POLARIS/share
+    mount -t tmpfs -o size=20M tmpfs $HOST_DIR_POLARIS/share
+    mount -t tmpfs -o size=5M tmpfs $HOST_DIR_POLARIS/external
     mkdir -p $HOST_DIR_POLARIS/dsc
     mount -t tmpfs -o size=5M dsctmps $HOST_DIR_POLARIS/dsc
     mkdir -p $HOST_DIR_POLARIS/mnt
@@ -61,7 +62,7 @@ function start_polaris()
 
     docker ps -a --format "{{.ID}}\t{{.Image}}" | grep "$IMAGE_NAME:$TAG" | awk '{print $1}' | xargs -I {} docker rm {}
 
-    docker run -v $HOST_DIR_POLARIS/update:/update -v $HOST_DIR_POLARIS/sysconfig/config0:/sysconfig/config0 -v $HOST_DIR_POLARIS/sysconfig/config1:/sysconfig/config1 -v $HOST_DIR_POLARIS/obfl/a:/obfl -v $HOST_DIR_POLARIS/obfl:/var/log/obfl -v $HOST_DIR_POLARIS/data:/data -v $HOST_DIR_POLARIS/share:/share -v $HOST_DIR_POLARIS/share:/external -v $HOST_DIR_POLARIS/mnt/a:/ro -v /dev:/dev -v /sys:/sys --net=host --name=$CONTAINER_NAME_POLARIS --privileged $IMAGE_NAME:$TAG &
+    docker run -v $HOST_DIR_POLARIS/update:/update -v $HOST_DIR_POLARIS/sysconfig/config0:/sysconfig/config0 -v $HOST_DIR_POLARIS/sysconfig/config1:/sysconfig/config1 -v $HOST_DIR_POLARIS/obfl/a:/obfl -v $HOST_DIR_POLARIS/obfl:/var/log/obfl -v $HOST_DIR_POLARIS/data:/data -v $HOST_DIR_POLARIS/share:/share -v $HOST_DIR_POLARIS/external:/external -v $HOST_DIR_POLARIS/mnt/a:/ro -v /dev:/dev -v /sys:/sys --net=host --name=$CONTAINER_NAME_POLARIS --privileged $IMAGE_NAME:$TAG &
 
     if [ -f /boot/first_boot ]; then
         device="/usr/share/sonic/device"
@@ -74,6 +75,9 @@ function start_polaris()
     sleep 5
     INTERFACE="eth0-midplane"
     if ip link show "$INTERFACE" &> /dev/null; then
+        echo "dhclient -r $INTERFACE"
+        /usr/sbin/dhclient -r $INTERFACE
+        sleep 1
         echo "dhclient $INTERFACE"
         /usr/sbin/dhclient $INTERFACE
     fi

files/image_config/midplane-network/midplane-network-dpu.network

@@ -3,3 +3,6 @@ Name=eth0-midplane
 
 [Network]
 DHCP=yes
+
+[DHCPv4]
+ClientIdentifier=mac

platform/checkout/pensando.ini

@@ -1,4 +1,4 @@
 [module]
 repo=git@github.com:pensando/pensando-sonic-artifacts.git
-ref=1.87.0-SS-15-release
+ref=1.95.0-SS-17-release
 path=platform/pensando/pensando-sonic-artifacts

platform/pensando/one-image.mk

@@ -9,10 +9,21 @@ $(SONIC_ONE_IMAGE)_INSTALLS += $(DPU_MODULE)
 $(SONIC_ONE_IMAGE)_INSTALLS += $(IONIC_MODULE)
 $(SONIC_ONE_IMAGE)_INSTALLS += $(PENSANDO_DPU_PLATFORM_MODULE)
 
+
+DISABLED_DOCKERS = $(DOCKER_SFLOW) $(DOCKER_MGMT_FRAMEWORK) $(DOCKER_NAT) $(DOCKER_TEAMD) $(DOCKER_ROUTER_ADVERTISER) $(DOCKER_MUX) $(DOCKER_SNMP)
+DISABLED_PACKAGES_LOCAL = $(DOCKER_DHCP_RELAY) $(DOCKER_MACSEC)
+DISABLED_FEATURE_FLAGS = INCLUDE_SFLOW INCLUDE_MGMT_FRAMEWORK INCLUDE_NAT INCLUDE_MACSEC INCLUDE_TEAMD INCLUDE_ROUTER_ADVERTISER INCLUDE_MUX
+$(info Disabling the following docker images: $(DISABLED_DOCKERS))
+$(info Disabling the following packages:  $(DISABLED_PACKAGES_LOCAL))
+$(info Disabling the following feauture flags:  $(DISABLED_FEATURE_FLAGS))
+
+SONIC_PACKAGES_LOCAL := $(filter-out $(DISABLED_PACKAGES_LOCAL), $(SONIC_PACKAGES_LOCAL))
+$(foreach feature, $(DISABLED_FEATURE_FLAGS), $(eval override $(feature)=n ))
+
 ifeq ($(INSTALL_DEBUG_TOOLS),y)
 $(SONIC_ONE_IMAGE)_DOCKERS += $(SONIC_INSTALL_DOCKER_DBG_IMAGES)
-$(SONIC_ONE_IMAGE)_DOCKERS += $(filter-out $(patsubst %-$(DBG_IMAGE_MARK).gz,%.gz, $(SONIC_INSTALL_DOCKER_DBG_IMAGES)), $(SONIC_INSTALL_DOCKER_IMAGES))
+$(SONIC_ONE_IMAGE)_DOCKERS += $(filter-out $(patsubst %-$(DBG_IMAGE_MARK).gz,%.gz, $(SONIC_INSTALL_DOCKER_DBG_IMAGES)), $(filter-out $(DISABLED_DOCKERS), $(SONIC_INSTALL_DOCKER_IMAGES)))
 else
-$(SONIC_ONE_IMAGE)_DOCKERS = $(SONIC_INSTALL_DOCKER_IMAGES)
+$(SONIC_ONE_IMAGE)_DOCKERS = $(filter-out $(DISABLED_DOCKERS), $(SONIC_INSTALL_DOCKER_IMAGES))
 endif
 SONIC_INSTALLERS += $(SONIC_ONE_IMAGE)

platform/pensando/sonic-platform-modules-dpu/dpu/service/dpu-db-util.service

@@ -5,8 +5,8 @@ After=dpu-platform-init.service
 [Service]
 ExecStart=/usr/bin/python3 /usr/local/bin/dpu_db_util.py
 Restart=always
-StandardOutput=syslog+console
-StandardError=syslog+console
+StandardOutput=journal+console
+StandardError=journal+console
 
 [Install]
 WantedBy=multi-user.target

platform/pensando/sonic-platform-modules-dpu/dpu/utils/dpu_db_util.py

@@ -6,18 +6,15 @@
 
 import sys
 import signal
-import subprocess
 import threading
 import time
 from datetime import datetime
-import json
 import docker
 import redis
 import syslog
 from sonic_py_common import daemon_base, logger, syslogger
-import multiprocessing
 import grpc
-from concurrent import futures
+import queue
 
 SYSLOG_IDENTIFIER = 'dpu-db-utild'
 logger_instance = syslogger.SysLogger(SYSLOG_IDENTIFIER)
@@ -29,7 +26,6 @@ def log_err(msg, also_print_to_console=False):
     logger_instance.log_error(msg, also_print_to_console)
 
 try:
-    from health_checker.manager import HealthCheckerManager
     from sonic_py_common import daemon_base
     import sonic_platform
     from sonic_platform.chassis import Chassis
@@ -84,11 +80,12 @@ def __init__(self, chassis, db):
         super(EventHandler, self).__init__(SYSLOG_IDENTIFIER)
 
         # operd attributes
-        self.events = []
-        self.events.extend(CRITICAL_EVENTS)
-        self.events.extend(NETWORK_EVENTS)
-        self.event_thread = None
-        self.event_stop = False
+        self.event_types = []
+        self.event_types.extend(CRITICAL_EVENTS)
+        self.event_types.extend(NETWORK_EVENTS)
+        self.request_queue = queue.Queue()
+        self.thread = None
+        self.stop_event = threading.Event()
 
         # dpu state db update related attributes
         try:
@@ -242,25 +239,13 @@ def _update_dpu_control_plane_db(self):
         except Exception as e:
             log_err(f'Failed to populate dpu control plane entries due to {e}')
 
-    def _getGrpcEventMessage(self):
-        while True:
-            if self.event_stop:
-                return
-            grpcmsg = oper_pb2.OperInfoRequest()
-            spec = grpcmsg.Request.add()
-            spec.InfoType = oper_pb2.OPER_INFO_TYPE_EVENT
-            spec.Action = oper_pb2.OPER_INFO_OP_SUBSCRIBE
-            for event in self.events:
-                spec.EventFilter.Types.append(event)
-            yield grpcmsg
-
     def _process_event(self, event):
         global g_count
 
         try:
-            event_type = event.EventInfo.Type
-            event_description = event.EventInfo.Description
-            event_message = event.EventInfo.Message
+            event_type = event.Type
+            event_description = event.Description
+            event_message = event.Message
         except Exception as e:
             log_err(f"Failed to process event due to {e}")
             return
@@ -301,29 +286,60 @@ def _process_event(self, event):
         except Exception as e:
             log_err(f"Failed to update dpu state db control plane fields due to {e}")
 
+    def _build_event_request(self):
+        """Builds a single gRPC subscription request."""
+        request = oper_pb2.OperInfoRequest()
+        spec = request.Request.add()
+        spec.InfoType = oper_pb2.OPER_INFO_TYPE_EVENT
+        spec.Action = oper_pb2.OPER_INFO_OP_SUBSCRIBE
+        for t in self.event_types:
+            spec.EventFilter.Types.append(t)
+        return request
+
+    def _request_generator(self):
+        """Generator that yields the gRPC request once."""
+        yield self._build_event_request()
+        while not self.stop_event.is_set():
+            time.sleep(1)
+
     def _event_listener(self):
-        channel_addr = "{}:{}".format(LOCALHOST,str(EVENT_PORT))
+        channel_addr = f"{LOCALHOST}:{EVENT_PORT}"
+        log_info(f"Connecting to gRPC server at {channel_addr}")
         channel = grpc.insecure_channel(channel_addr)
         stub = oper_pb2_grpc.OperSvcStub(channel)
-        resp = stub.OperInfoSubscribe(self._getGrpcEventMessage())
-        time.sleep(1)
-        for event in resp:
-            self._process_event(event)
+
+        try:
+            response_stream = stub.OperInfoSubscribe(self._request_generator())
+            for response in response_stream:
+                if self.stop_event.is_set():
+                    break
+                if response.Status != 0:
+                    log_info(f"Received non-OK status: {response.Status}")
+                    continue
+                event = response.EventInfo
+                try:
+                    self._process_event(event)
+                except Exception as e:
+                    log_err(f"_process_event() raised exception: {e}")
+        except grpc.RpcError as e:
+            log_err(f"gRPC error: {e}")
+        finally:
+            log_info("Event listener thread exiting")
 
     def start(self):
-        # spawn operd listener thread
-        self.event_stop = False
-        if (self.event_thread == None) or (not self.event_thread.is_alive()):
-            self.event_thread = threading.Thread(target=self._event_listener)
-            self.event_thread.daemon = True
-            self.event_thread.start()
+        if self.thread is None or not self.thread.is_alive():
+            self.stop_event.clear()
+            self.thread = threading.Thread(target=self._event_listener, daemon=True)
+            self.thread.start()
+            log_info("Event listener thread started")
 
     def stop(self):
-        self.log_warning("Stopping event listener thread")
-        self.event_stop = True
-        if self.event_thread is not None:
-            self.event_thread.join()
-            self.event_thread = None
+        log_info("Stopping event listener thread")
+        self.stop_event.set()
+        if self.thread is not None:
+            self.thread.join(timeout=2)
+            log_info("Event listener thread stopped")
+            self.thread = None
 
 #
 # Daemon =======================================================================

platform/pensando/sonic-platform-modules-dpu/dpu/utils/fetch_dpu_status

@@ -7,11 +7,7 @@
 #
 #############################################################################
 
-import sys
-import subprocess
-from datetime import datetime
-import json
-import syslog
+from datetime import datetime, timezone, timedelta
 import docker
 import redis
 from sonic_py_common import syslogger
@@ -26,9 +22,7 @@ def log_err(msg, also_print_to_console=False):
     logger_instance.log_error(msg, also_print_to_console)
 
 try:
-    from swsscommon import swsscommon
     from health_checker.manager import HealthCheckerManager
-    from sonic_py_common import daemon_base
     from sonic_platform.chassis import Chassis
     from sonic_platform.helper import APIHelper
 except Exception as e:
@@ -42,10 +36,13 @@ STATE_DB = 6
 REDIS_LOCALHOST_SERVER_PORT = 6379
 REDIS_LOCALHOST_SERVER_IP = '127.0.0.1'
 NOT_AVAILABLE = 'N/A'
-
+STARTUP_CONTAINER_LIST = ['swss', 'syncd', 'database']
+CONTAINERS_TO_IGNORE = ['macsec']
 try:
     apiHelper = APIHelper()
     chassis = Chassis()
+    dpu_docker_name = apiHelper.get_dpu_docker_container_name()
+    STARTUP_CONTAINER_LIST.append(dpu_docker_name)
 except Exception as e:
     log_err(f'failed to fetch dpu docker name due to {e}')
 
@@ -67,6 +64,14 @@ def bool_to_link_status(status):
     else:
         return "down"
 
+def parse_docker_time(ts: str) -> datetime:
+    ts = ts.rstrip("Z")
+    if '.' in ts:
+        base, fraction = ts.split('.')
+        ts = f"{base}.{fraction[:6]}"
+    dt = datetime.strptime(ts, "%Y-%m-%dT%H:%M:%S.%f")
+    return dt.replace(tzinfo=timezone.utc)
+
 class DPUHealthUpdater():
 
     def __init__(self, chassis):
@@ -117,18 +122,31 @@ class DPUHealthUpdater():
             client = docker.from_env()
             containers = client.containers.list(all=True)
             all_container_status = True
+            startup_container_status = True
+            swss_container_uptime = 0
             container_not_running = []
             container_restarting = []
             reason = ""
             for container in containers:
                 container_name = container.name
                 container_status = container.status
+                if container_name in CONTAINERS_TO_IGNORE:
+                    continue
                 if container_status == 'restarting':
+                    if container_name in STARTUP_CONTAINER_LIST:
+                        startup_container_status &= False
                     all_container_status &= False
                     container_restarting.append(container_name)
                 elif container_status == 'exited':
+                    if container_name in STARTUP_CONTAINER_LIST:
+                        startup_container_status &= False
                     all_container_status &= False
                     container_not_running.append(container_name)
+                if container_name == "swss" and startup_container_status:
+                    swss_container_uptime = datetime.now(timezone.utc) - parse_docker_time(container.attrs['State']['StartedAt'])
+            if startup_container_status and swss_container_uptime < timedelta(minutes=2):
+                reason = "Early stage containers (swss, syncd, database) are up and running"
+                return startup_container_status, reason
             if container_not_running:
                 reason += "Container not running : " + ', '.join(container_not_running)
             if container_restarting:

platform/pensando/sonic-platform-modules-dpu/sonic_platform/chassis.py

@@ -23,7 +23,7 @@
 
 NUM_THERMAL = 2
 NUM_VOLTAGE_SENSORS = 6
-NUM_CURRENT_SENSORS = 3
+NUM_CURRENT_SENSORS = 6
 HOST_REBOOT_CAUSE_PATH = "/host/reboot-cause/"
 REBOOT_CAUSE_FILE = "reboot-cause.txt"
 HOST_CHK_CMD = "docker > /dev/null 2>&1"
@@ -140,7 +140,7 @@ def __initialize_thermals(self):
         global NUM_THERMAL
         board_id = self._api_helper.get_board_id()
         if board_id == self._api_helper.mtfuji_board_id:
-            NUM_THERMAL = 5
+            NUM_THERMAL = 8
         if Thermal._thermals_available():
             for index in range(0, NUM_THERMAL):
                 thermal = Thermal(index)

platform/pensando/sonic-platform-modules-dpu/sonic_platform/sensor.py

@@ -29,9 +29,12 @@
 
 # [ Sensor-Name, sysfs, low_threshold, high_threshold, critical_low, critical_high]
 CURRENT_SENSOR_MAPPING = [
-    ["Current sensor 1", "/sys/class/hwmon/hwmon0/curr1_input", "0", "15100", NOT_AVAILABLE, "30000"],
-    ["Current sensor 2", "/sys/class/hwmon/hwmon1/curr1_input", "0", "13800", NOT_AVAILABLE, "30000"],
-    ["Current sensor 3", "/sys/class/hwmon/hwmon2/curr1_input", "0", "79100", NOT_AVAILABLE, "30000"]
+    ["VP0P85_VDD_DDR_DPU0", "/sys/bus/i2c/devices/0-0044/hwmon/hwmon2/curr1_input", "0", "15100", NOT_AVAILABLE, "30000"],
+    ["VP1P2_DDR_VDDQ_DPU0", "/sys/bus/i2c/devices/0-0044/hwmon/hwmon2/curr2_input", "0", "13800", NOT_AVAILABLE, "30000"],
+    ["VP0P75_VDD_CORE_DPU0 1", "/sys/bus/i2c/devices/0-0055/hwmon/hwmon1/curr1_input", "0", "25000", NOT_AVAILABLE, "30000"],
+    ["VP0P75_VDD_CORE_DPU0 2", "/sys/bus/i2c/devices/0-0055/hwmon/hwmon1/curr2_input", "0", "25000", NOT_AVAILABLE, "30000"],
+    ["VP0P75_VDD_CORE_DPU0 3", "/sys/bus/i2c/devices/0-0066/hwmon/hwmon0/curr1_input", "0", "25000", NOT_AVAILABLE, "30000"],
+    ["VP0P85_VDD_ARM_DPU0", "/sys/bus/i2c/devices/0-0066/hwmon/hwmon0/curr2_input", "0", "29100", NOT_AVAILABLE, "30000"],
 ]
 
 class VoltageSensor(SensorBase):
@@ -207,3 +210,4 @@ def get_low_critical_threshold(self):
         if value == NOT_AVAILABLE:
             return NOT_AVAILABLE
         return float(value)
+

platform/pensando/sonic-platform-modules-dpu/sonic_platform/thermal.py

@@ -31,9 +31,12 @@ class Thermal(ThermalBase):
     SENSOR_MAPPING_MTFUJI = [
         ["Die temperature", "/sys/class/hwmon/hwmon0/temp2_input", 1, 110, -10, 130],
         ["Board temperature", "/sys/class/hwmon/hwmon0/temp1_input", 1, 110, -10, 130],
-        ["Thermal sensor 1", "/sys/class/hwmon/hwmon0/temp1_input", 1, 110, -10, 130],
-        ["Thermal sensor 2", "/sys/class/hwmon/hwmon1/temp1_input", 1, 110, -10, 130],
-        ["Thermal sensor 3", "/sys/class/hwmon/hwmon2/temp1_input", 1, 110, -10, 130]
+        ["VP0P85_VDD_DDR_DPU0", "/sys/bus/i2c/devices/0-0044/hwmon/hwmon2/temp2_input", 1, 110, -10, 130],
+        ["VP1P2_DDR_VDDQ_DPU0", "/sys/bus/i2c/devices/0-0044/hwmon/hwmon2/temp3_input", 1, 110, -10, 130],
+        ["VP0P75_VDD_CORE_DPU0 1", "/sys/bus/i2c/devices/0-0055/hwmon/hwmon1/temp2_input", 1, 110, -10, 130],
+        ["VP0P75_VDD_CORE_DPU0 2", "/sys/bus/i2c/devices/0-0055/hwmon/hwmon1/temp3_input", 1, 110, -10, 130],
+        ["VP0P75_VDD_CORE_DPU0 3", "/sys/bus/i2c/devices/0-0066/hwmon/hwmon0/temp2_input", 1, 110, -10, 130],
+        ["VP0P85_VDD_ARM_DPU0", "/sys/bus/i2c/devices/0-0066/hwmon/hwmon0/temp3_input", 1, 110, -10, 130],
     ]
 
     @classmethod
@@ -43,7 +46,7 @@ def _thermals_available(cls):
         apiHelper = APIHelper()
         g_board_id = apiHelper.get_board_id()
         temp_hwmon = '/sys/bus/i2c/devices/i2c-0/0-004c/hwmon'
-        if g_board_id == self._api_helper.mtfuji_board_id:
+        if g_board_id == apiHelper.mtfuji_board_id:
             temp_hwmon = '/sys/class/hwmon/hwmon0/temp1_input'
         if os.path.exists(temp_hwmon):
             return True