Merge amd-staging into amd-master 20231214

Signed-off-by: guanyu12 <guanyu12@amd.com>
Change-Id: Iebd82680b7ed56abf84ad71a92a267a90a488aa6


[ROCm/rocm_smi_lib commit: 4b17a34716]

projects/rocm-smi-lib/.github/CODEOWNERS

@@ -0,0 +1 @@
+* @bill-shuzhou-liu @dmitrii-galantsev @charis-poag-amd @oliveiradan

projects/rocm-smi-lib/src/rocm_smi.cc

@@ -4836,15 +4836,11 @@ rsmi_dev_memory_partition_set(uint32_t dv_ind,
          << devInfoTypesStrings.at(amd::smi::kDevMemoryPartition)
          << " | Cause: device board name does not support this action"
          << " | Returning = "
-         << getRSMIStatusString(RSMI_STATUS_INVALID_ARGS) << " |";
+         << getRSMIStatusString(RSMI_STATUS_NOT_SUPPORTED) << " |";
       LOG_ERROR(ss);
     return RSMI_STATUS_NOT_SUPPORTED;
   }
 
-  std::string newMemoryPartition
-              = mapRSMIToStringMemoryPartitionTypes.at(memory_partition);
-  std::string currentMemoryPartition;
-
   switch (memory_partition) {
     case RSMI_MEMORY_PARTITION_NPS1:
     case RSMI_MEMORY_PARTITION_NPS2:
@@ -4865,6 +4861,9 @@ rsmi_dev_memory_partition_set(uint32_t dv_ind,
       LOG_ERROR(ss);
       return RSMI_STATUS_INVALID_ARGS;
   }
+  std::string newMemoryPartition
+              = mapRSMIToStringMemoryPartitionTypes.at(memory_partition);
+  std::string currentMemoryPartition;
 
   // do nothing if memory_partition is the current mode
   rsmi_status_t ret_get = get_memory_partition(dv_ind, currentMemoryPartition);
@@ -4907,6 +4906,9 @@ rsmi_dev_memory_partition_set(uint32_t dv_ind,
 
   if (amd::smi::ErrnoToRsmiStatus(ret) != RSMI_STATUS_SUCCESS) {
     rsmi_status_t err = amd::smi::ErrnoToRsmiStatus(ret);
+    if (ret == EACCES) {
+      err = RSMI_STATUS_NOT_SUPPORTED;  // already verified permissions
+    }
     ss << __PRETTY_FUNCTION__
        << " | ======= end ======= "
        << " | Fail "

projects/rocm-smi-lib/tests/rocm_smi_test/functional/computepartition_read_write.cc

@@ -237,8 +237,10 @@ void TestComputePartitionReadWrite::Run(void) {
     }
 
     // Verify api support checking functionality is working
-    err = rsmi_dev_compute_partition_set(dv_ind,
-      RSMI_COMPUTE_PARTITION_INVALID);
+    rsmi_compute_partition_type_t breakMe;
+    err = rsmi_dev_compute_partition_set(dv_ind, breakMe);
+    std::cout << "\t**rsmi_dev_compute_partition_set(null ptr): "
+              << amd::smi::getRSMIStatusString(err, false) << "\n";
     ASSERT_TRUE((err == RSMI_STATUS_INVALID_ARGS) ||
                 (err == RSMI_STATUS_NOT_SUPPORTED) ||
                 (err == RSMI_STATUS_PERMISSION));

projects/rocm-smi-lib/tests/rocm_smi_test/functional/memorypartition_read_write.cc

@@ -53,6 +53,7 @@
 #include "gtest/gtest.h"
 #include "rocm_smi/rocm_smi.h"
 #include "rocm_smi_test/functional/memorypartition_read_write.h"
+#include "rocm_smi/rocm_smi_utils.h"
 #include "rocm_smi_test/test_common.h"
 
 TestMemoryPartitionReadWrite::TestMemoryPartitionReadWrite() : TestBase() {
@@ -123,6 +124,7 @@ void TestMemoryPartitionReadWrite::Run(void) {
   }
 
   for (uint32_t dv_ind = 0; dv_ind < num_monitor_devs(); ++dv_ind) {
+    bool wasSetSuccess = false;
     if (dv_ind != 0) {
       IF_VERB(STANDARD) {
         std::cout << std::endl;
@@ -197,8 +199,10 @@ void TestMemoryPartitionReadWrite::Run(void) {
     /* rsmi_dev_memory_partition_set(...) */
     /******************************/
     // Verify api support checking functionality is working
-    rsmi_memory_partition_type new_memory_partition;
+    rsmi_memory_partition_type new_memory_partition = {};
     err = rsmi_dev_memory_partition_set(dv_ind, new_memory_partition);
+    std::cout << "\t**rsmi_dev_memory_partition_set(null ptr): "
+              << amd::smi::getRSMIStatusString(err, false) << "\n";
     // Note: new_memory_partition is not set
     ASSERT_TRUE((err == RSMI_STATUS_INVALID_ARGS) ||
                 (err == RSMI_STATUS_NOT_SUPPORTED));
@@ -264,7 +268,18 @@ void TestMemoryPartitionReadWrite::Run(void) {
                   << memoryPartitionString(new_memory_partition) << std::endl;
       }
       ret = rsmi_dev_memory_partition_set(dv_ind, new_memory_partition);
-      CHK_ERR_ASRT(ret)
+      if (ret == RSMI_STATUS_NOT_SUPPORTED) {
+        IF_VERB(STANDARD) {
+          std::cout << "\t**" <<  ": "
+                    << "Not supported on this machine" << std::endl;
+        }
+        break;
+      } else {
+        CHK_ERR_ASRT(ret)
+      }
+      if (ret != RSMI_STATUS_SUCCESS) {  // do not continue trying to reset
+        wasSetSuccess = true;
+      }
 
       ret = rsmi_dev_memory_partition_get(dv_ind, current_memory_partition,
                                           255);
@@ -301,7 +316,7 @@ void TestMemoryPartitionReadWrite::Run(void) {
                 << "Current memory partition: " << current_memory_partition
                 << std::endl;
     }
-    if (wasResetSuccess) {
+    if (wasResetSuccess && wasSetSuccess) {
       ASSERT_STRNE(oldMode.c_str(), current_memory_partition);
       IF_VERB(STANDARD) {
       std::cout << "\t**"