Merge amd-staging into amd-mainline 20251028 (#793)

Signed-off-by: Maisam Arif <[email protected]>

Author: marifamd

CHANGELOG.md

@@ -8,7 +8,7 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
 
 ### Added
 
-- **Added the following API's to amdsmi_interface.py**.  
+- **Added the following C API's to amdsmi_interface.py**.  
   - amdsmi_get_cpu_handle()
   - amdsmi_get_esmi_err_msg()
   - amdsmi_get_gpu_event_notification()
@@ -24,6 +24,37 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
   - The entry `policies` is added to the end of the dictionary to match API definition.
   - The entry `plpds` is marked for deprecation as it has the same information as `policies`.
 
+- **Added pcie levels to `amd-smi static --bus` command**.  
+  - The static --bus option has been updated to include the range of pcie levels that one may set a device to.
+  - Levels are a 2-tuple composed of the PCIE speed and bandwidth.
+
+  ```console
+  $ amd-smi static --bus
+  GPU: 0
+  BUS:
+  BDF: 0000:43:00.0
+  MAX_PCIE_WIDTH: 16
+  MAX_PCIE_SPEED: 16 GT/s
+  PCIE_LEVELS:
+    0: (2.5 GT/s, 1)
+    1: (5.0 GT/s, 4)
+    2: (16.0 GT/s, 16)
+  PCIE_INTERFACE_VERSION: Gen 4
+  SLOT_TYPE: CEM
+  ```
+
+- **Added evicted_time metric for kfd processes**.  
+  - Time that queues are evicted on a GPU in milliseconds
+  - Added to CLI in `amd-smi monitor -q` and `amd-smi process`
+  - Added to C API and Python API:
+    - amdsmi_get_gpu_process_list()
+    - amdsmi_get_gpu_compute_process_info()
+    - amdsmi_get_gpu_compute_process_info_by_pid()
+
+- **Added new VRAM types to  `amdsmi_vram_type_t`**.
+  - `amd-smi static --vram` & `amdsmi_get_gpu_vram_info()` now support the following types:
+  - DDR5, LPDDR4, LPDDR5, and HBM3E
+
 ### Changed
 
 - N/A
@@ -1211,9 +1242,9 @@ Updated `amdsmi_get_gpu_metrics_info()` and structure `amdsmi_gpu_metrics_t` to
     GPU: 0
         CLK_LEVEL: Successfully changed sclk perf level(s) to 5, 6
 
-    GPU: 1
-        CLK_LEVEL: level(s) 5, 6 is/are greater than performance levels supported for device
-    ```
+GPU: 1
+    CLK_LEVEL: clock level(s) 5, 6 is/are greater than sclk frequency levels supported for device GPU ID: 1 BDF:0000:46:00.0
+```
 
 - **Added new command `amd-smi static -C/--clock`**.  
   - This new command displays the clock frequency performance levels for the selected GPUs and clocks.

DEBIAN/postinst.in

@@ -176,15 +176,20 @@ do_install_amdsmi_python_lib() {
   export PIP_BREAK_SYSTEM_PACKAGES="$PREVIOUS_PIP_BREAK_SYSTEM_PACKAGES"
 
   # only try to activate argcomplete if such command exists
+  # python3-argcomplete is recommended but optional, we handle its absence gracefully
   if command -v activate-global-python-argcomplete &>/dev/null; then
-    activate-global-python-argcomplete
+    activate-global-python-argcomplete 2>/dev/null || {
+      echo "[INFO] Bash completion activation skipped. You can manually enable it with: activate-global-python-argcomplete"
+    }
   else
     # try older argcomplete3 version
     if command -v activate-global-python-argcomplete3 &>/dev/null; then
-      activate-global-python-argcomplete3
+      activate-global-python-argcomplete3 2>/dev/null || {
+        echo "[INFO] Bash completion activation skipped. You can manually enable it with: activate-global-python-argcomplete3"
+      }
     else
-      echo "[WARNING] Could not find argcomplete or argcomplete3. "\
-        "Argument completion will not work..."
+      echo "[WARNING] Could not find argcomplete activation command. "\
+        "Argument completion will not work. Install python3-argcomplete package to enable it."
     fi
   fi
 }

README.md

@@ -30,7 +30,7 @@ for more information.
 
 The following are required to install and use the AMD SMI library through its language interfaces and CLI.
 
-* `amdgpu` driver must be loaded for [`amdsmi_init()`](./docs/how-to/amdsmi-cpp-lib#hello-amd-smi) to work.
+* `amdgpu` driver must be loaded for [`amdsmi_init()`](./docs/how-to/amdsmi-cpp-lib#hello-amd-smi) to work. Refer to the [Instinct documentation](https://instinct.docs.amd.com/projects/amdgpu-docs/en/latest/install/detailed-install/prerequisites.html) for installation instructions.
 * Export `LD_LIBRARY_PATH` to the `amdsmi` installation directory.
 
   ```bash
@@ -51,51 +51,6 @@ sudo python3 -m pip install more_itertools
 
 * Go version 1.20 or greater
 
-## Install amdgpu driver and AMD SMI with ROCm
-
-1. Get the `amdgpu-install` installer following the instructions for your Linux distribution at
-   [Installation via AMDGPU installer](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/amdgpu-install.html#installation).
-
-2. Use `amdgpu-install` to install the `amdgpu` driver and ROCm packages with
-   AMD SMI included.
-
-   ``` shell
-   sudo amdgpu-install --usecase=rocm
-   ```
-
-   The `amdgpu-install --usecase=rocm` option triggers both an `amdgpu` driver
-   update and AMD SMI packages to be installed on your device.
-
-3. Verify your installation.
-
-   ```shell
-   amd-smi --help
-   ```
-
-## Install AMD SMI without ROCm
-
-The following are example steps to install the AMD SMI libraries and CLI tool on
-Ubuntu 22.04.
-
-1. Install the library.
-
-   ```shell
-   sudo apt install amd-smi-lib
-   ```
-
-2. Add the installation directory to your PATH. If installed with ROCm, ignore
-   this step.
-
-   ```shell
-   export PATH="${PATH:+${PATH}:}~/opt/rocm/bin"
-   ```
-
-3. Verify your installation.
-
-   ```shell
-   amd-smi --help
-   ```
-
 ## AMD SMI basic usage
 
 ### C++ library
@@ -175,6 +130,10 @@ during development; earlier versions are not guaranteed to work.
 * libdrm-dev (for Ubuntu and Debian)
 * libdrm-devel (for RPM-based distributions)
 
+To build the tests, the following additional dependency is required:
+
+* Google Test (libgtest-dev for Ubuntu/Debian, gtest-devel for RPM-based distributions)
+
 In order to build the AMD SMI Python package, the following components are required:
 
 * Python (3.6.8 or later)

RPM/post.in

@@ -178,15 +178,20 @@ do_install_amdsmi_python_lib() {
   export PIP_BREAK_SYSTEM_PACKAGES="$PREVIOUS_PIP_BREAK_SYSTEM_PACKAGES"
 
   # only try to activate argcomplete if such command exists
+  # python3-argcomplete is recommended but optional, we handle its absence gracefully
   if command -v activate-global-python-argcomplete &>/dev/null; then
-    activate-global-python-argcomplete
+    activate-global-python-argcomplete 2>/dev/null || {
+      echo "[INFO] Bash completion activation skipped. You can manually enable it with: activate-global-python-argcomplete"
+    }
   else
     # try older argcomplete3 version
     if command -v activate-global-python-argcomplete3 &>/dev/null; then
-      activate-global-python-argcomplete3
+      activate-global-python-argcomplete3 2>/dev/null || {
+        echo "[INFO] Bash completion activation skipped. You can manually enable it with: activate-global-python-argcomplete3"
+      }
     else
-      echo "[WARNING] Could not find argcomplete or argcomplete3. "\
-        "Argument completion will not work..."
+      echo "[WARNING] Could not find argcomplete activation command. "\
+        "Argument completion will not work. Install python3-argcomplete package to enable it."
     fi
   fi
 }