Merge amd-staging into amd-master 20231121

Change-Id: I400dfcdbf7fd1afcb020805342a4389038ce3917
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>

python_smi_tools/README.md

@@ -1,6 +1,4 @@
-## Synopsis
-
-Radeon Open Compute Platform - System Management Interface - Command Line tool.
+## Radeon Open Compute (ROCm) - System Management Interface - Command Line Tool
 
 This tool acts as a command line interface for manipulating
 and monitoring the amdgpu kernel, and is intended to replace

python_smi_tools/rocm_smi.py

@@ -29,10 +29,12 @@ from rsmiBindings import *
 # Major version - Increment when backwards-compatibility breaks
 # Minor version - Increment when adding a new feature, set to 0 when major is incremented
 # Patch version - Increment when adding a fix, set to 0 when minor is incremented
-SMI_MAJ = 1
-SMI_MIN = 5
+# Hash  version - Shortened commit hash. Print here and not with lib for consistency with amd-smi
+SMI_MAJ = 2
+SMI_MIN = 0
 SMI_PAT = 0
-__version__ = '%s.%s.%s' % (SMI_MAJ, SMI_MIN, SMI_PAT)
+# SMI_HASH is provided by rsmiBindings
+__version__ = '%s.%s.%s+%s' % (SMI_MAJ, SMI_MIN, SMI_PAT, SMI_HASH)
 
 # Set to 1 if an error occurs
 RETCODE = 0
@@ -828,23 +830,20 @@ def printTableRow(space, displayString, v_delim=" "):
 
 def checkIfSecondaryDie(device):
     """ Checks if GCD(die) is the secondary die in a MCM.
+    MI200 device specific feature check.
+    The secondary dies lacks power management features.
 
-    Secondary dies lack power management features.
-    TODO: switch to more robust way to check for primary/secondary die, when implemented in Kernel and rocm_smi_lib.
     @param device: The device to check
     """
-    power_cap = c_uint64()
-    # secondary die can currently be determined by checking if all power1_* (power cap) values are equal to zero.
-    ret = rocmsmi.rsmi_dev_power_cap_get(device, 0, byref(power_cap))
-    if not (rsmi_ret_ok(ret, None, 'get_power_cap', False) and power_cap.value == 0):
-        return False
-    ret = rocmsmi.rsmi_dev_power_cap_default_get(device, byref(power_cap))
-    if not (rsmi_ret_ok(ret, None, 'get_power_cap_default', False) and power_cap.value == 0):
-        return False
-    ret = rocmsmi.rsmi_dev_power_ave_get(device, 0, byref(power_cap))
-    if not (rsmi_ret_ok(ret, None, 'get_power_avg', False) and power_cap.value == 0):
-        return False
-    return True
+    energy_count = c_uint64()
+    counter_resoution = c_float()
+    timestamp = c_uint64()
+
+    # secondary die can be determined by checking if energy counter == 0
+    ret = rocmsmi.rsmi_dev_energy_count_get(device, byref(energy_count), byref(counter_resoution), byref(timestamp))
+    if (rsmi_ret_ok(ret, None, 'energy_count_secondary_die_check', silent=False)) and (energy_count.value == 0):
+        return True
+    return False
 
 def resetClocks(deviceList):
     """ Reset clocks to default

python_smi_tools/rsmiBindings.py.in

@@ -55,6 +55,8 @@ dv_id = c_uint64()
 # GPU ID
 gpu_id = c_uint32(0)
 
+SMI_HASH = '@PKG_VERSION_HASH@'
+
 
 # Policy enums
 RSMI_MAX_NUM_FREQUENCIES = 33