Merge amd-staging into amd-mainline 20250308 (#162)

Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>

Author: marifamd

.github/workflows/amdsmi-build.yml

@@ -13,6 +13,7 @@ jobs:
     runs-on:
       - self-hosted
       - ${{ vars.RUNNER_TYPE }}
+    continue-on-error: true
     strategy:
       matrix:
         os:
@@ -171,6 +172,7 @@ jobs:
     runs-on:
       - self-hosted
       - ${{ vars.RUNNER_TYPE }}
+    continue-on-error: true
     strategy:
       matrix:
         os:

CHANGELOG.md

@@ -8,6 +8,36 @@ Full documentation for amd_smi_lib is available at [https://rocm.docs.amd.com/pr
 
 ### Added
 
+- **Added enumeration mapping `amdsmi_get_gpu_enumeration_info()` to Python & C APIs.**  
+  - Enumeration mapping consists of `amdsmi_enumeration_info_t`
+
+    ```shell
+    typedef struct {
+        uint32_t drm_render; // the render node under /sys/class/drm/renderD*
+        uint32_t drm_card;   // the graphic card device under /sys/class/drm/card*
+        uint32_t hsa_id;     // the HSA enumeration ID
+        uint32_t hip_id;     // the HIP enumeration ID
+        char hip_uuid[AMDSMI_MAX_STRING_LENGTH];  // the HIP unique identifer
+    } amdsmi_enumeration_info_t;
+    ```
+
+  - The mapping is also enabled in the CLI interface via `amd-smi list -e`
+
+    ```shell
+    $ amd-smi list -e
+    GPU: 0
+        BDF: 0000:23:00.0
+        UUID: XXXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX
+        KFD_ID: 45412
+        NODE_ID: 1
+        PARTITION_ID: 0
+        RENDER: renderD128
+        CARD: card0
+        HSA_ID: 1
+        HIP_ID: 0
+        HIP_UUID: GPU-XXXXXXXXXXXXXXXX
+    ```
+
 - **Added dynamic virtualization mode detection**.  
   - Added new C and Python API `amdsmi_get_gpu_virtualization_mode_info`
   - Added new C and Python enum `amdsmi_virtualization_mode_t`
@@ -414,7 +444,9 @@ Updated structure `amdsmi_vram_info_t`:
 
 ### Upcoming changes
 
-- **Deprication of the `AMDSMI_LIB_VERSION_YEAR` enum and API fields.**  
+- **Deprication in ROCm 7.0 of the `AMDSMI_LIB_VERSION_YEAR` enum and API fields.**  
+
+- **Deprication in ROCm 7.0 of the `pasid` field within struct `amdsmi_process_info_t`**  
 
 ### Known issues
 

README.md

@@ -26,10 +26,23 @@ for more information.
 * [Install the AMD SMI library and CLI tool](https://rocm.docs.amd.com/projects/amdsmi/en/latest/install/install.html)
 
 ## Requirements
-The following are required to install and use the AMD SMI libraries and CLI tool.
 
-* Python 3.6.8+ (64-bit)
+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.
+* Export `LD_LIBRARY_PATH` to the `amdsmi` installation directory.
+
+  ```bash
+  export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/rocm/lib:/opt/rocm/lib64
+  ```
+
+### Python interface and CLI tool prerequisites
+
+* Python 3.6.8+ (64-bit)
+
+### Go API prerequisites
+
+* Go version 1.20 or greater
 
 ## Install amdgpu driver and AMD SMI with ROCm
 
@@ -97,6 +110,17 @@ Refer to the [user guide](https://rocm.docs.amd.com/projects/amdsmi/en/latest/ho
 detailed [Python API reference](https://rocm.docs.amd.com/projects/amdsmi/en/latest/reference/amdsmi-py-api.html) in the
 ROCm documentation portal.
 
+### Go library
+
+The AMD SMI Go interface provides a simple
+[API](https://rocm.docs.amd.com/projects/amdsmi/en/latest/reference/amdsmi-go-lib.html)
+for AMD hardware management. It streamlines hardware monitoring and control
+while leveraging Golang's features.
+
+Refer to the [user guide](https://rocm.docs.amd.com/projects/amdsmi/en/latest/how-to/amdsmi-go-lib.html) and the
+[Go API reference](https://rocm.docs.amd.com/projects/amdsmi/en/latest/reference/amdsmi-go-api.html) in the
+ROCm documentation portal.
+
 ### CLI tool
 
 A versatile command line tool for managing and monitoring AMD hardware. You can use `amd-smi` for:

amdsmi_cli/amdsmi_commands.py

@@ -175,7 +175,6 @@ def version(self, args, gpu_version=None, cpu_version=None):
         elif self.logger.is_json_format() or self.logger.is_csv_format():
             self.logger.print_output()
 
-
     def list(self, args, multiple_devices=False, gpu=None):
         """List information for target gpu
 
@@ -214,12 +213,20 @@ def list(self, args, multiple_devices=False, gpu=None):
             bdf = amdsmi_interface.amdsmi_get_gpu_device_bdf(args.gpu)
         except amdsmi_exception.AmdSmiLibraryException as e:
             bdf = e.get_error_info()
-
         try:
             uuid = amdsmi_interface.amdsmi_get_gpu_device_uuid(args.gpu)
         except amdsmi_exception.AmdSmiLibraryException as e:
             uuid = e.get_error_info()
 
+        try:
+            enumeration_info = amdsmi_interface.amdsmi_get_gpu_enumeration_info(args.gpu)
+        except:
+            enumeration_info = {"drm_render": "N/A",
+                                "drm_card": "N/A",
+                                "hip_id": "N/A",
+                                "hip_uuid": "N/A",
+                                "hsa_id": "N/A"}
+
         try:
             kfd_info = amdsmi_interface.amdsmi_get_gpu_kfd_info(args.gpu)
             kfd_id = kfd_info['kfd_id']
@@ -233,15 +240,26 @@ def list(self, args, multiple_devices=False, gpu=None):
         if self.logger.is_csv_format():
             self.logger.store_output(args.gpu, 'gpu_bdf', bdf)
             self.logger.store_output(args.gpu, 'gpu_uuid', uuid)
-            self.logger.store_output(args.gpu, 'kfd_id', kfd_id)
-            self.logger.store_output(args.gpu, 'node_id', node_id)
-            self.logger.store_output(args.gpu, 'partition_id', partition_id)
         else:
             self.logger.store_output(args.gpu, 'bdf', bdf)
             self.logger.store_output(args.gpu, 'uuid', uuid)
-            self.logger.store_output(args.gpu, 'kfd_id', kfd_id)
-            self.logger.store_output(args.gpu, 'node_id', node_id)
-            self.logger.store_output(args.gpu, 'partition_id', partition_id)
+
+        self.logger.store_output(args.gpu, 'kfd_id', kfd_id)
+        self.logger.store_output(args.gpu, 'node_id', node_id)
+        self.logger.store_output(args.gpu, 'partition_id', partition_id)
+
+        if args.e:
+            if enumeration_info['drm_render'] == "N/A":
+                self.logger.store_output(args.gpu, 'render', enumeration_info['drm_render'])
+            else:
+                self.logger.store_output(args.gpu, 'render', f"renderD{enumeration_info['drm_render']}")
+            if enumeration_info['drm_card'] == "N/A":
+                self.logger.store_output(args.gpu, 'card', enumeration_info['drm_card'])
+            else:
+                self.logger.store_output(args.gpu, 'card', f"card{enumeration_info['drm_card']}")
+            self.logger.store_output(args.gpu, 'hsa_id', enumeration_info['hsa_id'])
+            self.logger.store_output(args.gpu, 'hip_id', enumeration_info['hip_id'])
+            self.logger.store_output(args.gpu, 'hip_uuid', enumeration_info['hip_uuid'])
 
         if multiple_devices:
             self.logger.store_multiple_device_output()

amdsmi_cli/amdsmi_parser.py

@@ -626,6 +626,8 @@ def _add_list_parser(self, subparsers: argparse._SubParsersAction, func):
 
         # Create list subparser
         list_parser = subparsers.add_parser('list', help=list_help, description=list_subcommand_help)
+        # Create -e subparser
+        list_parser.add_argument("-e", action="store_true", help="Enumeration mapping to other features.\n    Lists the BDF, UUID, KFD_ID, CARD, RENDER, HIP_ID, HIP_UUID and HSA_ID for each GPU.")
         list_parser.formatter_class=lambda prog: AMDSMISubparserHelpFormatter(prog)
         list_parser.set_defaults(func=func)