# Copyright 2015-2017 ARM Limited # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # """Definitions of scheduler events registered by the FTrace class""" from trappy.base import Base from trappy.dynamic import register_ftrace_parser, register_dynamic_ftrace class SchedLoadAvgSchedGroup(Base): """Corresponds to Linux kernel trace event sched_load_avg_sched_group""" unique_word = "sched_load_avg_sg:" """The unique word that will be matched in a trace line""" _cpu_mask_column = "cpus" pivot = "cpus" """The Pivot along which the data is orthogonal""" def finalize_object(self): """This condition is necessary to force column 'cpus' to be printed as 8 digits w/ leading 0 """ if self._cpu_mask_column in self.data_frame.columns: dfr = self.data_frame[self._cpu_mask_column].apply('{:0>8}'.format) self.data_frame[self._cpu_mask_column] = dfr register_ftrace_parser(SchedLoadAvgSchedGroup, "sched") class SchedLoadAvgTask(Base): """Corresponds to Linux kernel trace event sched_load_avg_task""" unique_word = "sched_load_avg_task:" """The unique word that will be matched in a trace line""" pivot = "pid" """The Pivot along which the data is orthogonal""" def get_pids(self, key=""): """Returns a list of (comm, pid) that contain 'key' in their 'comm'.""" dfr = self.data_frame.drop_duplicates(subset=['comm', 'pid']) dfr = dfr.ix[:, ['comm', 'pid']] return dfr[dfr['comm'].str.contains(key)].values.tolist() register_ftrace_parser(SchedLoadAvgTask, "sched") # pylint doesn't like globals that are not ALL_CAPS # pylint: disable=invalid-name SchedLoadAvgCpu = register_dynamic_ftrace("SchedLoadAvgCpu", "sched_load_avg_cpu:", "sched", pivot="cpu") """Load and Utilization Signals for CPUs""" SchedContribScaleFactor = register_dynamic_ftrace("SchedContribScaleFactor", "sched_contrib_scale_f:", "sched") """Event to register tracing of contrib factor""" class SchedCpuCapacity(Base): """Corresponds to Linux kernel trace event sched/cpu_capacity""" unique_word = "cpu_capacity:" """The unique word that will be matched in a trace line""" pivot = "cpu" """The Pivot along which the data is orthogonal""" def finalize_object(self): """This renaming is necessary because our cpu related pivot is 'cpu' and not 'cpu_id'. Otherwise you cannot 'mix and match' with other classes """ self.data_frame.rename(columns={'cpu_id':'cpu'}, inplace=True) self.data_frame.rename(columns={'state' :'capacity'}, inplace=True) register_ftrace_parser(SchedCpuCapacity, "sched") SchedWakeup = register_dynamic_ftrace("SchedWakeup", "sched_wakeup:", "sched", parse_raw=True) """Register SchedWakeup Event""" SchedWakeupNew = register_dynamic_ftrace("SchedWakeupNew", "sched_wakeup_new:", "sched", parse_raw=True) """Register SchedWakeupNew Event""" # pylint: enable=invalid-name class SchedSwitch(Base): """Parse sched_switch""" unique_word = "sched_switch:" def __init__(self): super(SchedSwitch, self).__init__(parse_raw=True) def create_dataframe(self): self.data_array = [line.replace(" ==> ", " ", 1) for line in self.data_array] super(SchedSwitch, self).create_dataframe() register_ftrace_parser(SchedSwitch, "sched") class SchedCpuFrequency(Base): """Corresponds to Linux kernel trace event power/cpu_frequency""" unique_word = "cpu_frequency:" """The unique word that will be matched in a trace line""" pivot = "cpu" """The Pivot along which the data is orthogonal""" def finalize_object(self): """This renaming is necessary because our cpu related pivot is 'cpu' and not 'cpu_id'. Otherwise you cannot 'mix and match' with other classes """ self.data_frame.rename(columns={'cpu_id':'cpu'}, inplace=True) self.data_frame.rename(columns={'state' :'frequency'}, inplace=True) register_ftrace_parser(SchedCpuFrequency, "sched") register_dynamic_ftrace("SchedMigrateTask", "sched_migrate_task:", "sched")