Process isolation and clean shader

A few APIs and command line options are added to support process
isolation and clean shader.

Change-Id: I98ad3fc9fc7429799a21798b7fca1c307de7f403

amdsmi_cli/README.md

@@ -148,9 +148,9 @@ Command Modifiers:
 
 ```bash
 ~$ amd-smi static --help
-usage: amd-smi static [-h] [-g GPU [GPU ...] | -U CPU [CPU ...]] [-a] [-b] [-V] [-d] [-v]
-                      [-c] [-B] [-r] [-p] [-l] [-u] [-s] [-i] [--json | --csv]
-                      [--file FILE] [--loglevel LEVEL]
+usage: amd-smi static [-h] [-g GPU [GPU ...]] [-a] [-b] [-V] [-d] [-v] [-c] [-B] [-r] [-p]
+                      [-l] [-P] [-x] [-s] [-u] [--json | --csv] [--file FILE]
+                      [--loglevel LEVEL]
 
 If no GPU is specified, returns static information for all GPUs on the system.
 If no static argument is provided, all static information will be displayed.
@@ -179,6 +179,7 @@ Static Arguments:
   -r, --ras                Displays RAS features information
   -p, --partition          Partition information
   -l, --limit              All limit metric values (i.e. power and thermal limits)
+  -s, --process-isolation  The process isolation status
   -u, --numa               All numa node information
 
 CPU Arguments:
@@ -474,13 +475,13 @@ Command Modifiers:
 ```bash
 usage: amd-smi set [-h] (-g GPU [GPU ...] | -U CPU [CPU ...] | -O CORE [CORE ...]) [-f %]
                    [-l LEVEL] [-P SETPROFILE] [-d SCLKMAX] [-C PARTITION] [-M PARTITION]
-                   [-o WATTS] [-p POLICY] [--cpu-pwr-limit PWR_LIMIT]
+                   [-o WATTS] [-p POLICY] [-i STATUS] [--cpu-pwr-limit PWR_LIMIT]
                    [--cpu-xgmi-link-width MIN_WIDTH MAX_WIDTH]
                    [--cpu-lclk-dpm-level NBIOID MIN_DPM MAX_DPM] [--cpu-pwr-eff-mode MODE]
                    [--cpu-gmi3-link-width MIN_LW MAX_LW] [--cpu-pcie-link-rate LINK_RATE]
                    [--cpu-df-pstate-range MAX_PSTATE MIN_PSTATE] [--cpu-enable-apb]
                    [--cpu-disable-apb DF_PSTATE] [--soc-boost-limit BOOST_LIMIT]
-                   [--core-boost-limit BOOST_LIMIT] [--json | --csv] [--file FILE]
+                   [--core-boost-limit BOOST_LIMIT] [-c] [--json | --csv] [--file FILE]
                    [--loglevel LEVEL]
 
 A GPU must be specified to set a configuration.
@@ -514,6 +515,8 @@ Set Arguments:
   -o, --power-cap WATTS                        Set power capacity limit
   -p, --dpm-policy POLICY_ID                   Set the GPU DPM policy using policy id
   -x, --xgmi-plpd POLICY_ID                    Set the GPU XGMI per-link power down policy using policy id
+  -i, --process-isolation STATUS               Enable or disable the GPU process isolation: 0 for disable and 1 for enable.
+  -c, --clear-sram-data                        Clear the GPU SRAM data
 
 CPU Arguments:
   --cpu-pwr-limit PWR_LIMIT                    Set power limit for the given socket. Input parameter is power limit value.

amdsmi_cli/amdsmi_commands.py

@@ -245,7 +245,7 @@ class AMDSMICommands():
     def static_gpu(self, args, multiple_devices=False, gpu=None, asic=None, bus=None, vbios=None,
                         limit=None, driver=None, ras=None, board=None, numa=None, vram=None,
                         cache=None, partition=None, dfc_ucode=None, fb_info=None, num_vf=None,
-                        policy=None, xgmi_plpd=None):
+                        policy=None, xgmi_plpd=None, process_isolation=None):
         """Get Static information for target gpu
 
         Args:
@@ -270,6 +270,7 @@ class AMDSMICommands():
             num_vf (bool, optional): Value override for args.num_vf. Defaults to None.
             policy (bool, optional): Value override for args.policy. Defaults to None.
             xgmi_plpd (bool, optional): Value override for args.xgmi_plpd. Defaults to None.
+            process_isolation (bool, optional): Value override for args.process_isolation. Defaults to None.
         Returns:
             None: Print output via AMDSMILogger to destination
         """
@@ -306,8 +307,10 @@ class AMDSMICommands():
                 args.policy = policy
             if xgmi_plpd:
                 args.xgmi_plpd = xgmi_plpd
-            current_platform_args += ["ras", "limit", "partition", "policy", "xgmi_plpd"]
-            current_platform_values += [args.ras, args.limit, args.partition, args.policy, args.xgmi_plpd]
+            if process_isolation:
+                args.process_isolation = process_isolation
+            current_platform_args += ["ras", "limit", "partition", "policy", "xgmi_plpd", "process_isolation"]
+            current_platform_values += [args.ras, args.limit, args.partition, args.policy, args.xgmi_plpd, args.process_isolation]
 
         if self.helpers.is_linux() and not self.helpers.is_virtual_os():
             if numa:
@@ -643,6 +646,16 @@ class AMDSMICommands():
                     logging.debug("Failed to get xgmi_plpd info for gpu %s | %s", gpu_id, e.get_error_info())
 
                 static_dict['xgmi_plpd'] = policy_info
+        if 'process_isolation' in current_platform_args:
+            if args.process_isolation:
+                try:
+                    status = amdsmi_interface.amdsmi_get_gpu_process_isolation(args.gpu)
+                    status = "Enabled" if status else "Disabled"
+                except amdsmi_exception.AmdSmiLibraryException as e:
+                    status = "N/A"
+                    logging.debug("Failed to process isolation for gpu %s | %s", gpu_id, e.get_error_info())
+
+                static_dict['process_isolation'] = status
         if 'numa' in current_platform_args:
             if args.numa:
                 try:
@@ -779,7 +792,7 @@ class AMDSMICommands():
                 bus=None, vbios=None, limit=None, driver=None, ras=None,
                 board=None, numa=None, vram=None, cache=None, partition=None,
                 dfc_ucode=None, fb_info=None, num_vf=None, cpu=None,
-                interface_ver=None, policy=None, xgmi_plpd = None):
+                interface_ver=None, policy=None, xgmi_plpd = None, process_isolation=None):
         """Get Static information for target gpu and cpu
 
         Args:
@@ -804,6 +817,7 @@ class AMDSMICommands():
             interface_ver (bool, optional): Value override for args.interface_ver. Defaults to None
             policy (bool, optional): Value override for args.policy. Defaults to None.
             xgmi_plpd (bool, optional): Value override for args.xgmi_plpd. Defaults to None.
+            process_isolation (bool, optional): Value override for args.process_isolation. Defaults to None.
         Raises:
             IndexError: Index error if gpu list is empty
 
@@ -829,7 +843,8 @@ class AMDSMICommands():
         gpu_args_enabled = False
         gpu_attributes = ["asic", "bus", "vbios", "limit", "driver", "ras",
                           "board", "numa", "vram", "cache", "partition",
-                          "dfc_ucode", "fb_info", "num_vf", "policy", "xgmi_plpd"]
+                          "dfc_ucode", "fb_info", "num_vf", "policy", "xgmi_plpd",
+                          "process_isolation"]
         for attr in gpu_attributes:
             if hasattr(args, attr):
                 if getattr(args, attr):
@@ -859,7 +874,8 @@ class AMDSMICommands():
                 self.static_gpu(args, multiple_devices, gpu, asic,
                                     bus, vbios, limit, driver, ras,
                                     board, numa, vram, cache, partition,
-                                    dfc_ucode, fb_info, num_vf, policy)
+                                    dfc_ucode, fb_info, num_vf, policy,
+                                    process_isolation)
         elif self.helpers.is_amd_hsmp_initialized(): # Only CPU is initialized
             if args.cpu == None:
                 args.cpu = self.cpu_handles
@@ -873,7 +889,8 @@ class AMDSMICommands():
             self.static_gpu(args, multiple_devices, gpu, asic,
                                 bus, vbios, limit, driver, ras,
                                 board, numa, vram, cache, partition,
-                                dfc_ucode, fb_info, num_vf, policy, xgmi_plpd)
+                                dfc_ucode, fb_info, num_vf, policy, xgmi_plpd,
+                                process_isolation)
 
 
     def firmware(self, args, multiple_devices=False, gpu=None, fw_list=True):
@@ -3326,7 +3343,8 @@ class AMDSMICommands():
 
     def set_gpu(self, args, multiple_devices=False, gpu=None, fan=None, perf_level=None,
                   profile=None, perf_determinism=None, compute_partition=None,
-                  memory_partition=None, power_cap=None, dpm_policy=None, xgmi_plpd = None):
+                  memory_partition=None, power_cap=None, dpm_policy=None, xgmi_plpd = None,
+                  process_isolation=None, clear_sram_data = None):
         """Issue reset commands to target gpu(s)
 
         Args:
@@ -3342,7 +3360,8 @@ class AMDSMICommands():
             power_cap (int, optional): Value override for args.power_cap. Defaults to None.
             dpm_policy (int, optional): Value override for args.dpm_policy. Defaults to None.
             xgmi_plpd (int, optional): Value override for args.xgmi_plpd. Defaults to None.
-
+            process_isolation (int, optional): Value override for args.process_isolation. Defaults to None.
+            clear_sram_data (int, optional): Value override for args.clear_sram_data. Defaults to None.
         Raises:
             ValueError: Value error if no gpu value is provided
             IndexError: Index error if gpu list is empty
@@ -3371,6 +3390,10 @@ class AMDSMICommands():
             args.dpm_policy = dpm_policy
         if xgmi_plpd:
             args.xgmi_plpd = xgmi_plpd
+        if process_isolation:
+            args.process_isolation = process_isolation
+        if clear_sram_data:
+            args.clear_sram_data = clear_sram_data
         # Handle No GPU passed
         if args.gpu == None:
             raise ValueError('No GPU provided, specific GPU target(s) are needed')
@@ -3389,9 +3412,11 @@ class AMDSMICommands():
                     args.compute_partition,
                     args.memory_partition,
                     args.perf_determinism is not None,
-                    args.power_cap,
-                    args.dpm_policy,
-                    args.xgmi_plpd]):
+                    args.power_cap is not None,
+                    args.dpm_policy is not None,
+                    args.xgmi_plpd is not None,
+                    args.process_isolation is not None,
+                    args.clear_sram_data]):
             command = " ".join(sys.argv[1:])
             raise AmdSmiRequiredCommandException(command, self.logger.format)
 
@@ -3455,25 +3480,6 @@ class AMDSMICommands():
                     raise PermissionError('Command requires elevation') from e
                 raise ValueError(f"Unable to set memory partition to {args.memory_partition} on {gpu_string}") from e
             self.logger.store_output(args.gpu, 'memorypartition', f"Successfully set memory partition to {args.memory_partition}")
-
-        if args.dpm_policy:
-            try:
-                amdsmi_interface.amdsmi_set_dpm_policy(args.gpu, args.dpm_policy)
-            except amdsmi_exception.AmdSmiLibraryException as e:
-                if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
-                    raise PermissionError('Command requires elevation') from e
-                raise ValueError(f"Unable to set dpm policy to {args.dpm_policy} on {gpu_string}") from e
-            self.logger.store_output(args.gpu, 'dpmpolicy', f"Successfully set dpm policy to id {args.dpm_policy}")
-
-        if args.xgmi_plpd:
-            try:
-                amdsmi_interface.amdsmi_set_xgmi_plpd(args.gpu, args.xgmi_plpd)
-            except amdsmi_exception.AmdSmiLibraryException as e:
-                if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
-                    raise PermissionError('Command requires elevation') from e
-                raise ValueError(f"Unable to set XGMI policy to {args.xgmi_plpd} on {gpu_string}") from e
-            self.logger.store_output(args.gpu, 'xgmiplpd', f"Successfully set per-link power down policy to id {args.dpm_policy}")
-
         if isinstance(args.power_cap, int):
             try:
                 power_cap_info = amdsmi_interface.amdsmi_get_power_cap_info(args.gpu)
@@ -3499,6 +3505,48 @@ class AMDSMICommands():
                 if min_power_cap == 0:
                     min_power_cap = 1
                 self.logger.store_output(args.gpu, 'powercap', f"Power cap must be between {min_power_cap} and {max_power_cap}")
+        if isinstance(args.dpm_policy, int):
+            try:
+                amdsmi_interface.amdsmi_set_dpm_policy(args.gpu, args.dpm_policy)
+            except amdsmi_exception.AmdSmiLibraryException as e:
+                if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
+                    raise PermissionError('Command requires elevation') from e
+                raise ValueError(f"Unable to set dpm policy to {args.dpm_policy} on {gpu_string}") from e
+            self.logger.store_output(args.gpu, 'dpmpolicy', f"Successfully set dpm policy to id {args.dpm_policy}")
+        if isinstance(args.xgmi_plpd, int):
+            try:
+                amdsmi_interface.amdsmi_set_xgmi_plpd(args.gpu, args.xgmi_plpd)
+            except amdsmi_exception.AmdSmiLibraryException as e:
+                if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
+                    raise PermissionError('Command requires elevation') from e
+                raise ValueError(f"Unable to set XGMI policy to {args.xgmi_plpd} on {gpu_string}") from e
+            self.logger.store_output(args.gpu, 'xgmiplpd', f"Successfully set per-link power down policy to id {args.dpm_policy}")
+        if isinstance(args.process_isolation, int):
+            status_string = "Enabled" if args.process_isolation else "Disabled"
+            result = f"Requested process isolation to {status_string}" # This should not print out
+            try:
+                current_status = amdsmi_interface.amdsmi_get_gpu_process_isolation(args.gpu)
+                if current_status == args.process_isolation:
+                    result = f"Process isolation is already {status_string}"
+                else:
+                    amdsmi_interface.amdsmi_set_gpu_process_isolation(args.gpu, args.process_isolation)
+                    result = f"Successfully set process isolation to {status_string}"
+            except amdsmi_exception.AmdSmiLibraryException as e:
+                if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
+                    raise PermissionError('Command requires elevation') from e
+                raise ValueError(f"Unable to set process isolation to {status_string} on {gpu_string}") from e
+
+            self.logger.store_output(args.gpu, 'process_isolation', result)
+        if args.clear_sram_data:
+            try:
+                # Only 1 can be used for now.
+                amdsmi_interface.amdsmi_set_gpu_clear_sram_data(args.gpu, 1)
+                result = 'Successfully clear GPU SRAM data'
+            except amdsmi_exception.AmdSmiLibraryException as e:
+                if e.get_error_code() == amdsmi_interface.amdsmi_wrapper.AMDSMI_STATUS_NO_PERM:
+                    raise PermissionError('Command requires elevation') from e
+                raise ValueError(f"Unable to clear SRAM data on GPU {gpu_id}") from e
+            self.logger.store_output(args.gpu, 'clear_sram_data', result)
 
         if multiple_devices:
             self.logger.store_multiple_device_output()
@@ -3513,7 +3561,8 @@ class AMDSMICommands():
                   cpu=None, cpu_pwr_limit=None, cpu_xgmi_link_width=None, cpu_lclk_dpm_level=None,
                   cpu_pwr_eff_mode=None, cpu_gmi3_link_width=None, cpu_pcie_link_rate=None,
                   cpu_df_pstate_range=None, cpu_enable_apb=None, cpu_disable_apb=None,
-                  soc_boost_limit=None, core=None, core_boost_limit=None, dpm_policy=None, xgmi_plpd=None):
+                  soc_boost_limit=None, core=None, core_boost_limit=None, dpm_policy=None, xgmi_plpd=None,
+                  process_isolation=None, clear_sram_data=None):
         """Issue reset commands to target gpu(s)
 
         Args:
@@ -3544,7 +3593,8 @@ class AMDSMICommands():
             core_boost_limit (int, optional): Value override for args.core_boost_limit. Defaults to None
             dpm_policy (int, optional): Value override for args.dpm_policy. Defaults to None.
             xgmi_plpd (int, optional): Value override for args.xgmi_plpd. Defaults to None.
-
+            process_isolation (int, optional): Value override for args.process_isolation. Defaults to None.
+            clear_sram_data (int, optional): Value override for args.clear_sram_data. Defaults to None.
         Raises:
             ValueError: Value error if no gpu value is provided
             IndexError: Index error if gpu list is empty
@@ -3564,7 +3614,8 @@ class AMDSMICommands():
         # Check if a GPU argument has been set
         gpu_args_enabled = False
         gpu_attributes = ["fan", "perf_level", "profile", "perf_determinism", "compute_partition",
-                          "memory_partition", "power_cap", "dpm_policy", "xgmi_plpd"]
+                          "memory_partition", "power_cap", "dpm_policy", "xgmi_plpd", "process_isolation",
+                          "clear_sram_data"]
         for attr in gpu_attributes:
             if hasattr(args, attr):
                 if getattr(args, attr) is not None:
@@ -3620,7 +3671,8 @@ class AMDSMICommands():
                 self.logger.clear_multiple_devices_ouput()
                 self.set_gpu(args, multiple_devices, gpu, fan, perf_level,
                                 profile, perf_determinism, compute_partition,
-                                memory_partition, power_cap, dpm_policy, xgmi_plpd)
+                                memory_partition, power_cap, dpm_policy, xgmi_plpd,
+                                process_isolation, clear_sram_data)
         elif self.helpers.is_amd_hsmp_initialized(): # Only CPU is initialized
             if args.cpu == None and args.core == None:
                 raise ValueError('No CPU or CORE provided, specific target(s) are needed')
@@ -3639,7 +3691,8 @@ class AMDSMICommands():
             self.logger.clear_multiple_devices_ouput()
             self.set_gpu(args, multiple_devices, gpu, fan, perf_level,
                             profile, perf_determinism, compute_partition,
-                            memory_partition, power_cap, dpm_policy, xgmi_plpd)
+                            memory_partition, power_cap, dpm_policy, xgmi_plpd,
+                            process_isolation, clear_sram_data)
 
 
     def reset(self, args, multiple_devices=False, gpu=None, gpureset=None,
@@ -3660,7 +3713,6 @@ class AMDSMICommands():
             compute_partition (bool, optional): Value override for args.compute_partition. Defaults to None.
             memory_partition (bool, optional): Value override for args.memory_partition. Defaults to None.
             power_cap (int, optional): Value override for args.power_cap. Defaults to None.
-
         Raises:
             ValueError: Value error if no gpu value is provided
             IndexError: Index error if gpu list is empty

amdsmi_cli/amdsmi_parser.py

@@ -545,6 +545,7 @@ class AMDSMIParser(argparse.ArgumentParser):
         board_help = "All board information"
         dpm_policy_help = "The available DPM policy"
         xgmi_plpd_help = "The available XGMI per-link power down policy"
+        process_isolation_help = "The process isolation status"
 
         # Options arguments help text for Hypervisors and Baremetal
         ras_help = "Displays RAS features information"
@@ -586,6 +587,7 @@ class AMDSMIParser(argparse.ArgumentParser):
                 static_parser.add_argument('-l', '--limit', action='store_true', required=False, help=limit_help)
                 static_parser.add_argument('-P', '--policy', action='store_true', required=False, help=dpm_policy_help)
                 static_parser.add_argument('-x', '--xgmi-plpd', action='store_true', required=False, help=xgmi_plpd_help)
+                static_parser.add_argument('-R', '--process-isolation', action='store_true', required=False, help=process_isolation_help)
 
             if self.helpers.is_linux() and not self.helpers.is_virtual_os():
                 static_parser.add_argument('-u', '--numa', action='store_true', required=False, help=numa_help)
@@ -967,8 +969,9 @@ class AMDSMIParser(argparse.ArgumentParser):
         set_compute_partition_help = f"Set one of the following the compute partition modes:\n\t{compute_partition_choices_str}"
         set_memory_partition_help = f"Set one of the following the memory partition modes:\n\t{memory_partition_choices_str}"
         set_power_cap_help = "Set power capacity limit"
-        set_dpm_policy_help = f"Set the GPU DPM policy using policy id\n"
-        set_xgmi_plpd_help = f"Set the GPU XGMI per-link power down policy using policy id\n"
+        set_dpm_policy_help = "Set the GPU DPM policy using policy id\n"
+        set_xgmi_plpd_help = "Set the GPU XGMI per-link power down policy using policy id\n"
+        set_process_isolation_help = "Enable or disable the GPU process isolation: 0 for disable and 1 for enable.\n"
 
         # Help text for CPU set options
         set_cpu_pwr_limit_help = "Set power limit for the given socket. Input parameter is power limit value."
@@ -982,6 +985,7 @@ class AMDSMIParser(argparse.ArgumentParser):
         set_cpu_enable_apb_help = "Enables the DF p-state performance boost algorithm"
         set_cpu_disable_apb_help = "Disables the DF p-state performance boost algorithm. Input parameter is DFPstate (0-3)"
         set_soc_boost_limit_help = "Sets the boost limit for the given socket. Input parameter is socket BOOST_LIMIT value"
+        run_gpu_clear_sram_data_help = f"Clear the GPU SRAM data\n"
 
         # Help text for CPU Core set options
         set_core_boost_limit_help = "Sets the boost limit for the given core. Input parameter is core BOOST_LIMIT value"
@@ -1006,6 +1010,8 @@ class AMDSMIParser(argparse.ArgumentParser):
             set_value_parser.add_argument('-o', '--power-cap', action='store', type=self._positive_int, required=False, help=set_power_cap_help, metavar='WATTS')
             set_value_parser.add_argument('-p', '--dpm-policy', action='store', required=False,  type=self._not_negative_int, help=set_dpm_policy_help, metavar='POLICY_ID')
             set_value_parser.add_argument('-x', '--xgmi-plpd', action='store', required=False,  type=self._not_negative_int, help=set_xgmi_plpd_help, metavar='POLICY_ID')
+            set_value_parser.add_argument('-R', '--process-isolation', action='store', choices=[0,1], type=self._not_negative_int, required=False, help=set_process_isolation_help, metavar='STATUS')
+            set_value_parser.add_argument('-c', '--clear-sram-data', action='store_true', required=False, help=run_gpu_clear_sram_data_help)
 
         if self.helpers.is_amd_hsmp_initialized():
             # Optional CPU Args