commit b55847f2a3c973678eecded0ba6fbc6064ee82b4
author Harshit Aghera <haghera@nvidia.com> Tue May 27 14:53:56 2025 +0530
committer Aditya Kurdunkar <akurdunkar@nvidia.com> Wed Jun 18 17:11:51 2025 +0530
tree c46e928fd590e0b5f2b18990ee61cbac4da55e68
parent 128c91da0eccc3c1ea919020183bac37e9754ab3

nvidia-gpu: add voltage sensor

This commit introduces a voltage sensor for the GPU.

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/voltage_NVIDIA_GB200_GPU_0_Voltage_0
{
  "@odata.id": "/redfish/v1/Chassis/NVIDIA_GB200_1/Sensors/voltage_NVIDIA_GB200_GPU_0_Voltage_0",
  "@odata.type": "#Sensor.v1_2_0.Sensor",
  "Id": "voltage_NVIDIA_GB200_GPU_0_Voltage_0",
  "Name": "NVIDIA GB200 GPU 0 Voltage 0",
  "Reading": 0.735,
  "ReadingRangeMax": 1.8446744073709552e+16,
  "ReadingRangeMin": 0.0,
  "ReadingType": "Voltage",
  "ReadingUnits": "V",
  "Status": {
    "Health": "OK",
    "State": "Enabled"
  }
}%
```

Change-Id: I3d98f3d7c11221a42460c6f8420c927c1b1711b2
Signed-off-by: Harshit Aghera <haghera@nvidia.com>

src/nvidia-gpu/NvidiaGpuMctpVdm.cpp

diff --git a/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp b/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp
index b70e0c2..efde488 100644
--- a/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp
+++ b/src/nvidia-gpu/NvidiaGpuMctpVdm.cpp
@@ -349,5 +349,69 @@
 
     return 0;
 }
+
+int encodeGetVoltageRequest(uint8_t instanceId, uint8_t sensorId,
+                            std::span<uint8_t> buf)
+{
+    if (buf.size() < sizeof(GetVoltageRequest))
+    {
+        return EINVAL;
+    }
+
+    auto* msg = reinterpret_cast<GetVoltageRequest*>(buf.data());
+
+    ocp::accelerator_management::BindingPciVidInfo header{};
+    header.ocp_accelerator_management_msg_type =
+        static_cast<uint8_t>(ocp::accelerator_management::MessageType::REQUEST);
+    header.instance_id = instanceId &
+                         ocp::accelerator_management::instanceIdBitMask;
+    header.msg_type = static_cast<uint8_t>(MessageType::PLATFORM_ENVIRONMENTAL);
+
+    auto rc = packHeader(header, msg->hdr.msgHdr.hdr);
+
+    if (rc != 0)
+    {
+        return rc;
+    }
+
+    msg->hdr.command =
+        static_cast<uint8_t>(PlatformEnvironmentalCommands::GET_VOLTAGE);
+    msg->hdr.data_size = sizeof(sensorId);
+    msg->sensor_id = sensorId;
+
+    return 0;
+}
+
+int decodeGetVoltageResponse(std::span<const uint8_t> buf,
+                             ocp::accelerator_management::CompletionCode& cc,
+                             uint16_t& reasonCode, uint32_t& voltage)
+{
+    auto rc =
+        ocp::accelerator_management::decodeReasonCodeAndCC(buf, cc, reasonCode);
+
+    if (rc != 0 || cc != ocp::accelerator_management::CompletionCode::SUCCESS)
+    {
+        return rc;
+    }
+
+    if (buf.size() < sizeof(GetVoltageResponse))
+    {
+        return EINVAL;
+    }
+
+    const auto* response =
+        reinterpret_cast<const GetVoltageResponse*>(buf.data());
+
+    const uint16_t dataSize = le16toh(response->hdr.data_size);
+
+    if (dataSize != sizeof(uint32_t))
+    {
+        return EINVAL;
+    }
+
+    voltage = le32toh(response->voltage);
+
+    return 0;
+}
 // NOLINTEND(cppcoreguidelines-pro-type-reinterpret-cast)
 } // namespace gpu