Merge amd-staging into amd-mainline 20250310 (#176)

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

Author: dmitrii-galantsev

CMakeLists.txt

@@ -369,12 +369,10 @@ endif()
 
 ## Process the Rpm install/remove scripts to update the CPACK variables
 configure_file("${CMAKE_CURRENT_SOURCE_DIR}/RPM/post.in" RPM/post @ONLY)
-# configure_file("${CMAKE_CURRENT_SOURCE_DIR}/RPM/amd-smi-lib/preun.in" RPM/preun @ONLY)
 configure_file("${CMAKE_CURRENT_SOURCE_DIR}/RPM/preun.in" RPM/preun @ONLY)
 configure_file("${CMAKE_CURRENT_SOURCE_DIR}/RPM/postun.in" RPM/postun @ONLY)
 set(CPACK_RPM_POST_INSTALL_SCRIPT_FILE "${CMAKE_CURRENT_BINARY_DIR}/RPM/post")
 set(CPACK_RPM_PRE_UNINSTALL_SCRIPT_FILE "${CMAKE_CURRENT_BINARY_DIR}/RPM/preun")
-# set(CPACK_RPM_TESTS_PRE_UNINSTALL_SCRIPT_FILE "${CMAKE_CURRENT_BINARY_DIR}/RPM/amd-smi-lib-tests/preun")
 set(CPACK_RPM_POST_UNINSTALL_SCRIPT_FILE "${CMAKE_CURRENT_BINARY_DIR}/RPM/postun")
 
 #Set the names now using CPACK utility

RPM/amd-smi-lib-tests/preun.in

@@ -9,6 +9,9 @@ rm_leftovers() {
   rm -rf "$RPM_INSTALL_PREFIX0/@SHARE_INSTALL_PREFIX@/amdsmi.egg-info"
   rm -rf "$RPM_INSTALL_PREFIX0/@SHARE_INSTALL_PREFIX@/build"
 
+  # remove dist files (only applies to old setuptools versions like on RHEL8)
+  rm -rf "$RPM_INSTALL_PREFIX0/@SHARE_INSTALL_PREFIX@/dist"
+
   # remove leftover doc files
   if test -e "$RPM_INSTALL_PREFIX0/@SHARE_INSTALL_PREFIX@/../doc/amd_smi*"; then
     rm -rf "$RPM_INSTALL_PREFIX0/@SHARE_INSTALL_PREFIX@/../doc/amd_smi*"

RPM/amd-smi-lib/preun.in

@@ -29,10 +29,10 @@
 import time
 
 from _version import __version__
-from amdsmi import amdsmi_exception, amdsmi_interface
 from amdsmi_cli_exceptions import AmdSmiInvalidParameterException, AmdSmiRequiredCommandException
 from amdsmi_helpers import AMDSMIHelpers
 from amdsmi_logger import AMDSMILogger
+from amdsmi import amdsmi_exception, amdsmi_interface
 
 
 class AMDSMICommands():
@@ -387,6 +387,8 @@ def static_gpu(self, args, multiple_devices=False, gpu=None, asic=None, bus=None
             args.cache = cache
         if process_isolation:
             args.process_isolation = process_isolation
+        if partition:
+            args.partition = partition
         if clock:
             args.clock = clock
         # args.clock defaults to False so if it was overwritten to empty list, that indicates that it was given as an arguments but with an empty list
@@ -396,24 +398,22 @@ def static_gpu(self, args, multiple_devices=False, gpu=None, asic=None, bus=None
         # Store args that are applicable to the current platform
         current_platform_args = ["asic", "bus", "vbios", "driver", "ras",
                                  "vram", "cache", "board", "process_isolation",
-                                 "clock"]
+                                 "clock", "partition"]
         current_platform_values = [args.asic, args.bus, args.vbios, args.driver, args.ras,
                                    args.vram, args.cache, args.board, args.process_isolation,
-                                   args.clock]
+                                   args.clock, args.partition]
 
         self.helpers.check_required_groups()
 
         if self.helpers.is_linux() and self.helpers.is_baremetal():
-            if partition:
-                args.partition = partition
             if limit:
                 args.limit = limit
             if soc_pstate:
                 args.soc_pstate = soc_pstate
             if xgmi_plpd:
                 args.xgmi_plpd = xgmi_plpd
-            current_platform_args += ["ras", "limit", "partition", "soc_pstate", "xgmi_plpd"]
-            current_platform_values += [args.ras, args.limit, args.partition, args.soc_pstate, args.xgmi_plpd]
+            current_platform_args += ["ras", "limit", "soc_pstate", "xgmi_plpd"]
+            current_platform_values += [args.ras, args.limit, args.soc_pstate, args.xgmi_plpd]
 
         if self.helpers.is_linux() and not self.helpers.is_virtual_os():
             if numa:
@@ -4240,7 +4240,7 @@ def set_gpu(self, args, multiple_devices=False, gpu=None, fan=None, perf_level=N
                     if args.compute_partition in accelerator_profiles['profile_types']:
                         compute_partition = amdsmi_interface.AmdSmiComputePartitionType[args.compute_partition]
                         index = accelerator_profiles['profile_types'].index(args.compute_partition)
-                        attempted_to_set = f"Attempted to set accelerator partition to {args.compute_partition} (profile #{accelerator_profiles['profile_indices'][int(index)]} on {gpu_string}"
+                        attempted_to_set = f"Attempted to set accelerator partition to {args.compute_partition} (profile #{accelerator_profiles['profile_indices'][int(index)]}) on {gpu_string}"
                         amdsmi_interface.amdsmi_set_gpu_compute_partition(args.gpu, compute_partition)
                         self.logger.store_output(args.gpu, 'accelerator_partition', f"Successfully set accelerator partition to {args.compute_partition} (profile #{accelerator_profiles['profile_indices'][int(index)]})")
                     elif args.compute_partition in accelerator_profiles['profile_indices']:
@@ -4294,7 +4294,7 @@ def set_gpu(self, args, multiple_devices=False, gpu=None, fan=None, perf_level=N
 
                 threads = []
                 k140secs = 140
-                string_out = f"Updating memory partition for gpu {gpu_id}"
+                string_out = f"Updating memory partition for GPU: {gpu_id}"
                 timesToRetryRestartErr = 1
 
                 self.helpers.increment_set_count()
@@ -4305,9 +4305,9 @@ def set_gpu(self, args, multiple_devices=False, gpu=None, fan=None, perf_level=N
                 while timesToRetryRestartErr >= 0:
                     timesToRetryRestartErr -= 1
                     try:
-                        if showProgressBar: # only show reload warning on 1st set
+                        if showProgressBar:  # we want to overwrite the previous progress bar
                             t1 = multiprocessing.Process(target=self.helpers.showProgressbar,
-                                                        args=(string_out, k140secs,))
+                                                         args=(string_out, k140secs, True,))
                             threads.append(t1)
                             t1.start()
                         memory_partition = amdsmi_interface.AmdSmiMemoryPartitionType[args.memory_partition]
@@ -4342,7 +4342,7 @@ def set_gpu(self, args, multiple_devices=False, gpu=None, fan=None, perf_level=N
                             return
                         if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_AMDGPU_RESTART_ERR:
                             # Try again on a failure -> work around for not being able to close libdrm
-                            string_out = f"Trying again - Updating memory partition for gpu {gpu_id}"
+                            string_out = f"Trying again - Updating memory partition for GPU: {gpu_id}         "
                             for thread in threads:
                                 thread.terminate()
                                 thread.join()

RPM/preun.in

@@ -969,22 +969,26 @@ def get_pci_device_ids(self) -> Set[str]:
                 continue
         return pci_devices
 
-    def progressbar(self, it, prefix="", size=60, out=sys.stdout):
+    def progressbar(self, it, prefix="", size=60, out=sys.stdout, add_newline=False):
         count = len(it)
+        if (add_newline):
+            print("{}\n".format(prefix),end='\r', file=out, flush=False)
+        else:
+            print("{}".format(prefix),end='\r', file=out, flush=False)
         def show(j):
             x = int(size*j/count)
-            print("{}[{}{}] {}/{} secs remain".format(prefix, u"█"*x, "."*(size-x), j, count),
+            print("[{}{}] {}/{} secs remain".format(u"█"*x, "."*(size-x), j, count),
                     end='\r', file=out, flush=True)
         show(0)
         for i, item in enumerate(it):
             yield item
             show(i+1)
         print("\n\n", end='\r', flush=True, file=out)
 
-    def showProgressbar(self, title="", timeInSeconds=13):
+    def showProgressbar(self, title="", timeInSeconds=13, add_newline=False):
         if title != "":
-            title += ": "
-        for i in self.progressbar(range(timeInSeconds), title, 40):
+            title += " "
+        for i in self.progressbar(range(timeInSeconds), title, 40, add_newline=add_newline):
             time.sleep(1)
 
     def check_required_groups(self):

amdsmi_cli/amdsmi_commands.py

@@ -695,10 +695,10 @@ def _add_static_parser(self, subparsers: argparse._SubParsersAction, func):
             static_parser.add_argument('-R', '--process-isolation', action='store_true', required=False, help=process_isolation_help)
             static_parser.add_argument('-r', '--ras', action='store_true', required=False, help=ras_help)
             static_parser.add_argument('-C', '--clock', action='store', default=False, nargs='*', type=str, required=False, help=clock_help)
+            static_parser.add_argument('-p', '--partition', action='store_true', required=False, help=partition_help)
 
             # Options to display on Hypervisors and Baremetal
             if self.helpers.is_hypervisor() or self.helpers.is_baremetal():
-                static_parser.add_argument('-p', '--partition', action='store_true', required=False, help=partition_help)
                 static_parser.add_argument('-l', '--limit', action='store_true', required=False, help=limit_help)
                 static_parser.add_argument('-P', '--soc-pstate', action='store_true', required=False, help=soc_pstate_help)
                 static_parser.add_argument('-x', '--xgmi-plpd', action='store_true', required=False, help=xgmi_plpd_help)