gpu : add support for communication to the endpoint
The commit uses MCTP VDM protocol to read temperature sensor value from
the gpu.
The MCTP VDM protocol is an extension of the OCP Accelerator Management
Interface specification -
'''
https://www.opencompute.org/documents/ocp-gpu-accelerator-management-interfaces-v1-pdf
'''
Tested.
Build an image for gb200nvl-obmc machine with the following patches
cherry picked. This patches are needed to enable the mctp stack.
https://gerrit.openbmc.org/c/openbmc/openbmc/+/79312
https://gerrit.openbmc.org/c/openbmc/openbmc/+/79410
https://gerrit.openbmc.org/c/openbmc/openbmc/+/79422
Copy the configuration file on gb200nvl-obmc machine and restart the
entity-manager service.
```
root@gb200nvl-obmc:~# rm -rf /var/configuration/
root@gb200nvl-obmc:~# systemctl restart xyz.openbmc_project.EntityManager.service
```
Copy the gpusensor app and run it.
```
root@gb200nvl-obmc:~# ./gpusensor
```
The app is detecting entity-manager configuration on gb200nvl-obmc
machine. The app is also able to detect all the endpoints from the mctp
service dbus tree. The app is reading temperature sensor value from gpu
correctly and the temperature sensor is also present on redfish.
```
$ curl -k -u 'root:0penBmc' https://10.137.203.137/redfish/v1/Chassis/NVIDIA_GB200_1/Sensors/temperature_NVIDIA_GB200_GPU
{
"@odata.id": "/redfish/v1/Chassis/NVIDIA_GB200_1/Sensors/temperature_NVIDIA_GB200_GPU",
"@odata.type": "#Sensor.v1_2_0.Sensor",
"Id": "temperature_NVIDIA_GB200_GPU",
"Name": "NVIDIA GB200 GPU",
"Reading": 36.4375,
"ReadingRangeMax": 127.0,
"ReadingRangeMin": -128.0,
"ReadingType": "Temperature",
"ReadingUnits": "Cel",
"Status": {
"Health": "OK",
"State": "Enabled"
}
}%
root@gb200nvl-obmc:~# busctl tree xyz.openbmc_project.GpuSensor
└─ /xyz
└─ /xyz/openbmc_project
└─ /xyz/openbmc_project/sensors
└─ /xyz/openbmc_project/sensors/temperature
└─ /xyz/openbmc_project/sensors/temperature/NVIDIA_GB200_GPU
root@gb200nvl-obmc:~# busctl introspect xyz.openbmc_project.GpuSensor /xyz/openbmc_project/sensors/temperature/NVIDIA_GB200_GPU
NAME TYPE SIGNATURE RESULT/VALUE FLAGS
org.freedesktop.DBus.Introspectable interface - - -
.Introspect method - s -
org.freedesktop.DBus.Peer interface - - -
.GetMachineId method - s -
.Ping method - - -
org.freedesktop.DBus.Properties interface - - -
.Get method ss v -
.GetAll method s a{sv} -
.Set method ssv - -
.PropertiesChanged signal sa{sv}as - -
xyz.openbmc_project.Association.Definitions interface - - -
.Associations property a(sss) 1 "chassis" "all_sensors" "/xyz/openbmc… emits-change
xyz.openbmc_project.Sensor.Value interface - - -
.MaxValue property d 127 emits-change
.MinValue property d -128 emits-change
.Unit property s "xyz.openbmc_project.Sensor.Value.Unit.… emits-change
.Value property d 36.3125 emits-change writable
xyz.openbmc_project.Sensor.ValueMutability interface - - -
.Mutable property b true emits-change
xyz.openbmc_project.State.Decorator.Availability interface - - -
.Available property b true emits-change writable
xyz.openbmc_project.State.Decorator.OperationalStatus interface - - -
.Functional property b true emits-change
```
Change-Id: Ied938b9e5c19751ee283b4b948e16c905c78fb48
Signed-off-by: Harshit Aghera <haghera@nvidia.com>
diff --git a/src/gpu/GpuSensor.hpp b/src/gpu/GpuSensor.hpp
index 01eb23f..7c70d55 100644
--- a/src/gpu/GpuSensor.hpp
+++ b/src/gpu/GpuSensor.hpp
@@ -5,6 +5,7 @@
#pragma once
+#include "MctpRequester.hpp"
#include "Thresholds.hpp"
#include "sensor.hpp"
@@ -15,6 +16,7 @@
#include <sdbusplus/asio/object_server.hpp>
#include <sdbusplus/message.hpp>
+#include <chrono>
#include <cstdint>
#include <map>
#include <memory>
@@ -68,10 +70,12 @@
* @param verbose Whether to enable verbose logging
*/
GpuTempSensor(std::shared_ptr<sdbusplus::asio::connection>& conn,
- boost::asio::io_context& io, const std::string& name,
+ boost::asio::io_context& io,
+ mctp::MctpRequester& mctpRequester, const std::string& name,
const std::string& sensorConfiguration,
sdbusplus::asio::object_server& objectServer,
- std::vector<thresholds::Threshold>&& thresholdData);
+ std::vector<thresholds::Threshold>&& thresholdData,
+ std::chrono::milliseconds pollRate);
/**
* @brief Destructor
@@ -87,11 +91,21 @@
private:
/**
+ * @brief Read the current temperature value from the GPU
+ */
+ void read();
+
+ /**
* @brief Initialize the sensor
*/
void init();
/**
+ * @brief Update the sensor reading
+ */
+ void update();
+
+ /**
* @brief Discover available GPUs on the system
*/
void discoverGpus();
@@ -113,6 +127,26 @@
*/
void processEndpointConfigs(const boost::system::error_code& ec,
const GpuSensorConfigMap& configs);
+ /**
+ * @brief Process a discovered GPU endpoint
+ * @param eid The endpoint ID of the discovered GPU
+ */
+ void processGpuEndpoint(uint8_t eid);
+
+ /**
+ * @brief MCTP endpoint ID
+ */
+ uint8_t eid{};
+
+ /**
+ * @brief The sensor ID
+ */
+ uint8_t sensorId;
+
+ /**
+ * @brief How often to poll the sensor in milliseconds
+ */
+ std::chrono::milliseconds sensorPollMs;
/**
* @brief Timer for scheduling sensor reads
@@ -120,6 +154,11 @@
boost::asio::steady_timer waitTimer;
/**
+ * @brief Reference to the MCTP requester for communication
+ */
+ mctp::MctpRequester& mctpRequester;
+
+ /**
* @brief D-Bus connection
*/
std::shared_ptr<sdbusplus::asio::connection> conn;
@@ -136,12 +175,14 @@
* @param objectServer D-Bus object server
* @param sensors Map to store created sensors
* @param dbusConnection D-Bus connection
+ * @param mctpRequester MCTP requester for GPU communication
*/
void createSensors(
boost::asio::io_context& io, sdbusplus::asio::object_server& objectServer,
boost::container::flat_map<std::string, std::shared_ptr<GpuTempSensor>>&
sensors,
- std::shared_ptr<sdbusplus::asio::connection>& dbusConnection);
+ std::shared_ptr<sdbusplus::asio::connection>& dbusConnection,
+ mctp::MctpRequester& mctpRequester);
/**
* @brief Handle D-Bus interface removal events