commit ba138dae62cfd96571372a3e22317dc57ab72c80
author Harshit Aghera <haghera@nvidia.com> Mon May 05 12:26:35 2025 +0530
committer Ed Tanous <ed@tanous.net> Thu Jul 10 15:01:22 2025 +0000
tree 1b00cc22de6359dba7b387943d9f60fcbd681f1c
parent 4ecdfaaaf39039bd6d73ee67aa44830672675a81

nvidia-gpu: add TLimit sensor

This commit introduces a new thermal limit (TLimit) sensor for the GPU,
enhancing the existing temperature monitoring capabilities.

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/+/79422

```
$ curl -s -k -u 'root:0penBmc' https://10.137.203.137/redfish/v1/Chassis/NVIDIA_GB200_1/Sensors/temperature_NVIDIA_GB200_GPU_0_TEMP_1
{
  "@odata.id": "/redfish/v1/Chassis/NVIDIA_GB200_1/Sensors/temperature_NVIDIA_GB200_GPU_0_TEMP_1",
  "@odata.type": "#Sensor.v1_2_0.Sensor",
  "Id": "temperature_NVIDIA_GB200_GPU_0_TEMP_1",
  "Name": "NVIDIA GB200 GPU 0 TEMP 1",
  "Reading": 57.3984375,
  "ReadingRangeMax": 127.0,
  "ReadingRangeMin": -128.0,
  "ReadingType": "Temperature",
  "ReadingUnits": "Cel",
  "Status": {
    "Health": "OK",
    "State": "Enabled"
  }
}%
```

Change-Id: Ib8e0ef93a4acbb8870671665b098fb61d0205cb2
Signed-off-by: Harshit Aghera <haghera@nvidia.com>

src/nvidia-gpu/NvidiaGpuDevice.cpp

diff --git a/src/nvidia-gpu/NvidiaGpuDevice.cpp b/src/nvidia-gpu/NvidiaGpuDevice.cpp
index cd39b56..ab476da 100644
--- a/src/nvidia-gpu/NvidiaGpuDevice.cpp
+++ b/src/nvidia-gpu/NvidiaGpuDevice.cpp
@@ -42,8 +42,12 @@
 void GpuDevice::makeSensors()
 {
     tempSensor = std::make_shared<NvidiaGpuTempSensor>(
-        conn, mctpRequester, name + "_TEMP_0", path, eid, objectServer,
-        std::vector<thresholds::Threshold>{});
+        conn, mctpRequester, name + "_TEMP_0", path, eid, gpuTempSensorId,
+        objectServer, std::vector<thresholds::Threshold>{});
+
+    tLimitSensor = std::make_shared<NvidiaGpuTempSensor>(
+        conn, mctpRequester, name + "_TEMP_1", path, eid, gpuTLimitSensorId,
+        objectServer, std::vector<thresholds::Threshold>{});
 
     lg2::info("Added GPU {NAME} Sensors with chassis path: {PATH}.", "NAME",
               name, "PATH", path);
@@ -54,6 +58,7 @@
 void GpuDevice::read()
 {
     tempSensor->update();
+    tLimitSensor->update();
 
     waitTimer.expires_after(std::chrono::milliseconds(sensorPollMs));
     waitTimer.async_wait([this](const boost::system::error_code& ec) {

src/nvidia-gpu/NvidiaGpuDevice.hpp

diff --git a/src/nvidia-gpu/NvidiaGpuDevice.hpp b/src/nvidia-gpu/NvidiaGpuDevice.hpp
index 3653928..78b3e4a 100644
--- a/src/nvidia-gpu/NvidiaGpuDevice.hpp
+++ b/src/nvidia-gpu/NvidiaGpuDevice.hpp
@@ -53,6 +53,7 @@
     sdbusplus::asio::object_server& objectServer;
 
     std::shared_ptr<NvidiaGpuTempSensor> tempSensor;
+    std::shared_ptr<NvidiaGpuTempSensor> tLimitSensor;
 
     SensorConfigs configs;
 

src/nvidia-gpu/NvidiaGpuSensor.cpp

diff --git a/src/nvidia-gpu/NvidiaGpuSensor.cpp b/src/nvidia-gpu/NvidiaGpuSensor.cpp
index 1626545..3853048 100644
--- a/src/nvidia-gpu/NvidiaGpuSensor.cpp
+++ b/src/nvidia-gpu/NvidiaGpuSensor.cpp
@@ -31,20 +31,19 @@
 
 using namespace std::literals;
 
-constexpr uint8_t gpuTempSensorId{0};
 static constexpr double gpuTempSensorMaxReading = 127;
 static constexpr double gpuTempSensorMinReading = -128;
 
 NvidiaGpuTempSensor::NvidiaGpuTempSensor(
     std::shared_ptr<sdbusplus::asio::connection>& conn,
     mctp::MctpRequester& mctpRequester, const std::string& name,
-    const std::string& sensorConfiguration, const uint8_t eid,
+    const std::string& sensorConfiguration, const uint8_t eid, uint8_t sensorId,
     sdbusplus::asio::object_server& objectServer,
     std::vector<thresholds::Threshold>&& thresholdData) :
     Sensor(escapeName(name), std::move(thresholdData), sensorConfiguration,
            "temperature", false, true, gpuTempSensorMaxReading,
            gpuTempSensorMinReading, conn),
-    eid(eid), sensorId{gpuTempSensorId}, mctpRequester(mctpRequester),
+    eid(eid), sensorId{sensorId}, mctpRequester(mctpRequester),
     objectServer(objectServer)
 {
     std::string dbusPath =

src/nvidia-gpu/NvidiaGpuSensor.hpp

diff --git a/src/nvidia-gpu/NvidiaGpuSensor.hpp b/src/nvidia-gpu/NvidiaGpuSensor.hpp
index 25fe069..1e21f4d 100644
--- a/src/nvidia-gpu/NvidiaGpuSensor.hpp
+++ b/src/nvidia-gpu/NvidiaGpuSensor.hpp
@@ -20,17 +20,20 @@
 #include <string>
 #include <vector>
 
+constexpr uint8_t gpuTempSensorId{0};
+constexpr uint8_t gpuTLimitSensorId{2};
+
 struct NvidiaGpuTempSensor :
     public Sensor,
     public std::enable_shared_from_this<NvidiaGpuTempSensor>
 {
   public:
-    NvidiaGpuTempSensor(std::shared_ptr<sdbusplus::asio::connection>& conn,
-                        mctp::MctpRequester& mctpRequester,
-                        const std::string& name,
-                        const std::string& sensorConfiguration, uint8_t eid,
-                        sdbusplus::asio::object_server& objectServer,
-                        std::vector<thresholds::Threshold>&& thresholdData);
+    NvidiaGpuTempSensor(
+        std::shared_ptr<sdbusplus::asio::connection>& conn,
+        mctp::MctpRequester& mctpRequester, const std::string& name,
+        const std::string& sensorConfiguration, uint8_t eid, uint8_t sensorId,
+        sdbusplus::asio::object_server& objectServer,
+        std::vector<thresholds::Threshold>&& thresholdData);
 
     ~NvidiaGpuTempSensor() override;