** ''org.eclipse.tracecompass.tmf.ui.views.statesystem.TmfStateSystemExplorer.java''
** ''org.eclipse.tracecompass.analysis.os.linux.ui.views.cpuusage.CpuUsageComposite.java''
+== Timing Analysis Views and Viewers ==
+
+Trace Compass provides base implementations for timing views and viewers for generating Latency Tables, Scatter Charts, Density Graphs and Statistics Tables. They are well integrated with various Trace Compass features such as reading traces and time synchronization with other views. They also handle mouse events for navigating the trace and view, zooming or presenting detailed information at mouse position. The code can be found in the Analysis Timing plug-in ''org.eclipse.tracecompass.analysis.timing.ui''. See below for a list of relevant java packages:
+
+* Latency Table
+** ''org.eclipse.tracecompass.analysis.timing.ui.views.segmentstore.table'': Base classes for Latency Tables
+* Scatter Chart
+** ''org.eclipse.tracecompass.tmf.ui.views.tmfChartView.java'': Common base classes for X-Y-Chart viewers based on SWTChart
+** ''org.eclipse.tracecompass.analysis.timing.ui.views.segmentstore.scatter'': Base classes for Scatter Charts
+* Density Graph
+** ''org.eclipse.tracecompass.analysis.timing.ui.views.segmentstore.density'': Base classes for Density Graphs
+* Statistics Table
+** ''org.eclipse.tracecompass.analysis.timing.ui.views.segmentstore.statistics'': Base classes for Statistics Tables
+
+Several features in Trace Compass are using this framework and can be used as example for further development:
+
+* Latency Table
+** ''org.eclipse.tracecompass.internal.analysis.os.linux.ui.views.latency.SystemCallLatencyView.java''
+** ''org.eclipse.tracecompass.internal.tmf.analysis.xml.ui.views.latency.PatternLatencyTableView.java''
+* Scatter Chart
+** ''org.eclipse.tracecompass.internal.analysis.os.linux.ui.views.latency.SystemCallLatencyScatterView.java''
+** ''org.eclipse.tracecompass.internal.tmf.analysis.xml.ui.views.latency.PatternScatterGraphView.java''
+* Density Graph
+** ''org.eclipse.tracecompass.internal.analysis.os.linux.ui.views.latency.SystemCallLatencyDensityView.java''
+** ''org.eclipse.tracecompass.internal.tmf.analysis.xml.ui.views.latency.PatternDensityView.java''
+
+* Statistics Table
+** ''org.eclipse.tracecompass.internal.analysis.os.linux.ui.views.latency.statistics.SystemCallLatencyStatisticsView.java''
+** ''org.eclipse.tracecompass.internal.tmf.analysis.xml.ui.views.latency.PatternStatisticsView.java''
+
= Component Interaction =
TMF provides a mechanism for different components to interact with each other using signals. The signals can carry information that is specific to each signal.
Received by views that analyze packet streams.
+=== TmfStartAnalysisSignal ===
+
+''Purpose''
+
+This signal is used to indicate that an analysis has started.
+
+''Senders''
+
+Sent by an analysis module when it starts to execute the analyis.
+
+''Receivers''
+
+Received by components that need to be notified of the start of an analysis
+or that need to receive the analysis module.
+
+=== TmfCpuSelectedSignal ===
+
+''Purpose''
+
+This signal is used to indicate that the user has selected a CPU core.
+
+''Senders''
+
+Sent by any component that allows the user to select a CPU.
+
+''Receivers''
+
+Received by viewers that show information specific to a selected CPU.
+
+=== TmfThreadSelectedSignal ===
+
+''Purpose''
+
+This signal is used to indicate that the user has selected a thread.
+
+''Senders''
+
+Sent by any component that allows the user to select a thread.
+
+''Receivers''
+
+Received by viewers that show information specific to a selected thread.
+
+=== TmfTraceSynchronizedSignal ===
+
+''Purpose''
+
+This signal is used to indicate that trace synchronization has been completed.
+
+''Senders''
+
+Sent by the experiment after trace synchronization.
+
+''Receivers''
+
+Received by any component that needs to be notified of trace synchronization.
+
+=== TmfMarkerEventSourceUpdatedSignal ===
+
+''Purpose''
+
+This signal is used to indicate that a marker event source has been updated.
+
+''Senders''
+
+Sent by a component that has triggered a change in a marker event source.
+
+''Receivers''
+
+Received by any component that needs to refresh the markers due to the change in marker event source.
+
== Debugging ==
TMF has built-in Eclipse tracing support for the debugging of signal interaction between components. To enable it, open the '''Run/Debug Configuration...''' dialog, select a configuration, click the '''Tracing''' tab, select the plug-in '''org.eclipse.tracecompass.tmf.core''', and check the '''signal''' item.
they will have been written by the same person. In other cases, it's possible to
use getSubAttributes() to explore the organization of the attribute tree first.
+===== optQuarkAbsolute()/optQuarkRelative() =====
+
+These two methods are similar to their counterparts getQuarkAbsolute() and
+getQuarkRelative(). The only difference is that if the referenced attribute does
+not exist, the value ITmfStateSystem#INVALID_ATTRIBUTE (-2) is returned instead
+of throwing an exception.
+
+These methods should be used when the presence of the referenced attribute is
+known to be optional, to avoid the performance cost of generating exceptions.
+
+===== getQuarks() =====
+
+This method (with or without a starting node quark) takes an attribute path
+array which may contain wildcard "*" or parent ".." elements, and returns the
+list of matching attribute quarks. If no matching attribute is found, an empty
+list is returned.
+
===== waitUntilBuilt() =====
This is a simple method used to block the caller until the construction phase of
==== AttributeNotFoundException ====
-This is thrown by getQuarkRelative() and getQuarkAbsolute() (but not byt the
+This is thrown by getQuarkRelative() and getQuarkAbsolute() (but not by the
-AndAdd versions!) when passing an attribute path that is not present in the
state system. This is to ensure that no new attribute is created when using
these versions of the methods.
<pre>
private void createFrame() {
//...
- start.setTime(new TmfTimestamp(1000, -3));
- syn1.setTime(new TmfTimestamp(1005, -3));
- synReturn1.setTime(new TmfTimestamp(1050, -3));
- asyn1.setStartTime(new TmfTimestamp(1060, -3));
- asyn1.setEndTime(new TmfTimestamp(1070, -3));
- asynReturn1.setStartTime(new TmfTimestamp(1060, -3));
- asynReturn1.setEndTime(new TmfTimestamp(1070, -3));
+ start.setTime(TmfTimestamp.create(1000, -3));
+ syn1.setTime(TmfTimestamp.create(1005, -3));
+ synReturn1.setTime(TmfTimestamp.create(1050, -3));
+ asyn1.setStartTime(TmfTimestamp.create(1060, -3));
+ asyn1.setEndTime(TmfTimestamp.create(1070, -3));
+ asynReturn1.setStartTime(TmfTimestamp.create(1060, -3));
+ asynReturn1.setEndTime(TmfTimestamp.create(1070, -3));
//...
}
</pre>
There are other helper files that format given events for views, they are simpler and the architecture does not depend on them.
=== Limitations ===
-For the moment live trace reading is not supported, there are no sources of traces to test on.
+For the moment live CTF trace reading is not supported.
= Event matching and trace synchronization =
@Override
public Iterable<TmfAnalysisRequirement> getAnalysisRequirements() {
- // initialize the requirement: domain and events
- TmfAnalysisRequirement domainReq = new TmfAnalysisRequirement(SessionConfigStrings.CONFIG_ELEMENT_DOMAIN);
- domainReq.addValue(SessionConfigStrings.CONFIG_DOMAIN_TYPE_KERNEL, ValuePriorityLevel.MANDATORY);
-
- List<String> requiredEvents = ImmutableList.of("sched_switch", "sched_wakeup");
- TmfAnalysisRequirement eventReq = new TmfAnalysisRequirement(SessionConfigStrings.CONFIG_ELEMENT_EVENT,
- requiredEvents, ValuePriorityLevel.MANDATORY);
+ // initialize the requirement: events
+ Set<@NonNull String> requiredEvents = ImmutableSet.of("sched_switch", "sched_wakeup");
+ TmfAbstractAnalysisRequirement eventsReq = new TmfAnalysisEventRequirement(requiredEvents, PriorityLevel.MANDATORY);
- return ImmutableList.of(domainReq, eventReq);
+ return ImmutableList.of(eventsReq);
}
@Override
=== Analysis requirement provider API ===
-A requirement defines the needs of an analysis. For example, an analysis could need an event named ''"sched_switch"'' in order to be properly executed. The requirements are represented by the class '''TmfAnalysisRequirement'''. Since '''IAnalysisModule''' extends the '''IAnalysisRequirementProvider''' interface, all analysis modules must provide their requirements. If the analysis module extends '''TmfAbstractAnalysisModule''', it has the choice between overriding the requirements getter ('''IAnalysisRequirementProvider#getAnalysisRequirements()''') or not, since the abstract class returns an empty collection by default (no requirements).
+A requirement defines the needs of an analysis. For example, an analysis could need an event named ''"sched_switch"'' in order to be properly executed. The requirements are represented by extending the class '''TmfAbstractAnalysisRequirement'''. Since '''IAnalysisModule''' extends the '''IAnalysisRequirementProvider''' interface, all analysis modules must provide their requirements. If the analysis module extends '''TmfAbstractAnalysisModule''', it has the choice between overriding the requirements getter ('''IAnalysisRequirementProvider#getAnalysisRequirements()''') or not, since the abstract class returns an empty collection by default (no requirements).
=== Requirement values ===
-When instantiating a requirement, the developer needs to specify a type to which all the values added to the requirement will be linked. In the earlier example, there would be an ''"event"'' or ''"eventName"'' type. The type is represented by a string, like all values added to the requirement object. With an 'event' type requirement, a trace generator like the LTTng Control could automatically enable the required events. This is possible by calling the '''TmfAnalysisRequirementHelper''' class. Another point we have to take into consideration is the priority level of each value added to the requirement object. The enum '''TmfAnalysisRequirement#ValuePriorityLevel''' gives the choice between '''ValuePriorityLevel#MANDATORY''' and '''ValuePriorityLevel#OPTIONAL'''. That way, we can tell if an analysis can run without a value or not. To add values, one must call '''TmfAnalysisRequirement#addValue()'''.
+Each concrete analysis requirement class will define how a requirement is verified on a given trace.
+When creating a requirement, the developer will specify a set of values for that class.
+With an 'event' type requirement, a trace generator like the LTTng Control could automatically
+enable the required events.
+Another point we have to take into consideration is the priority level when creating a requirement object.
+The enum '''TmfAbstractAnalysisRequirement#PriorityLevel''' gives the choice
+between '''PriorityLevel#OPTIONAL''', '''PriorityLevel#ALL_OR_NOTHING''',
+'''PriorityLevel#AT_LEAST_ONE''' or '''PriorityLevel#MANDATORY'''. That way, we
+can tell if an analysis can run without a value or not.
+
+
+To create a requirement one has the choice to extend the abstract class
+'''TmfAbstractAnalysisRequirement''' or use the existing implementations:
+'''TmfAnalysisEventRequirement''' (will verify the presence of events identified by name),
+'''TmfAnalysisEventFieldRequirement''' (will verify the presence of fields for some or all events) or
+'''TmfCompositeAnalysisRequirement''' (join requirements together with one of the priority levels).
Moreover, information can be added to requirements. That way, the developer can explicitly give help details at the requirement level instead of at the analysis level (which would just be a general help text). To add information to a requirement, the method '''TmfAnalysisRequirement#addInformation()''' must be called. Adding information is not mandatory.
=== Example of providing requirements ===
-In this example, we will implement a method that initializes a requirement object and return it in the '''IAnalysisRequirementProvider#getAnalysisRequirements()''' getter. The example method will return a set with two requirements. The first one will indicate the events needed by a specific analysis and the last one will tell on what domain type the analysis applies. In the event type requirement, we will indicate that the analysis needs a mandatory event and an optional one.
+In this example, we will implement a method that initializes a requirement object
+and return it in the '''IAnalysisRequirementProvider#getAnalysisRequirements()'''
+getter. The example method will return a set with three requirements.
+The first one will indicate a mandatory event needed by a specific analysis,
+the second one will tell an optional event name and the third will indicate
+mandatory event fields for the given event type.
+
+Note that in LTTng event contexts are considered as event fields. Using the
+'''TmfAnalysisEventFieldRequirement''' it's possible to define requirements
+on event contexts (see 3rd requirement in example below).
<pre>
-@Override
-public Iterable<TmfAnalysisRequirement> getAnalysisRequirements() {
- Set<TmfAnalysisRequirement> requirements = new HashSet<>();
+ @Override
+ public @NonNull Iterable<@NonNull TmfAbstractAnalysisRequirement> getAnalysisRequirements() {
- /* Create requirements of type 'event' and 'domain' */
- TmfAnalysisRequirement eventRequirement = new TmfAnalysisRequirement("event");
- TmfAnalysisRequirement domainRequirement = new TmfAnalysisRequirement("domain");
+ /* Requirement on event name */
+ Set<@NonNull String> requiredEvents = ImmutableSet.of("sched_wakeup");
+ TmfAbstractAnalysisRequirement eventsReq1 = new TmfAnalysisEventRequirement(requiredEvents, PriorityLevel.MANDATORY);
- /* Add the values */
- domainRequirement.addValue("kernel", TmfAnalysisRequirement.ValuePriorityLevel.MANDATORY);
- eventRequirement.addValue("sched_switch", TmfAnalysisRequirement.ValuePriorityLevel.MANDATORY);
- eventRequirement.addValue("sched_wakeup", TmfAnalysisRequirement.ValuePriorityLevel.OPTIONAL);
+ requiredEvents = ImmutableSet.of("sched_switch");
+ TmfAbstractAnalysisRequirement eventsReq2 = new TmfAnalysisEventRequirement(requiredEvents, PriorityLevel.OPTIONAL);
- /* An information about the events */
- eventRequirement.addInformation("The event sched_wakeup is optional because it's not properly handled by this analysis yet.");
+ /* An information about the events */
+ eventsReq2.addInformation("The event sched_wakeup is optional because it's not properly handled by this analysis yet.");
- /* Add them to the set */
- requirements.add(domainRequirement);
- requirements.add(eventRequirement);
+ /* Requirement on event fields */
+ Set<@NonNull String> requiredEventFields = ImmutableSet.of("context._procname", "context._ip");
+ TmfAbstractAnalysisRequirement eventFieldRequirement = new TmfAnalysisEventFieldRequirement(
+ "event name",
+ requiredEventFields,
+ PriorityLevel.MANDATORY);
- return requirements;
-}
+ Set<TmfAbstractAnalysisRequirement> requirements = ImmutableSet.of(eventsReq1, eventsReq2, eventFieldRequirement);
+ return requirements;
+ }
</pre>
* Add SWTBOT tests to org.eclipse.tracecompass.tmf.pcap.ui
* Add a Raw Viewer, similar to Wireshark. We could use the “Show Raw” in the event editor to do that.
* Externalize strings in org.eclipse.tracecompass.pcap.core. At the moment, all the strings are hardcoded. It would be good to externalize them all.
+
+= Markers =
+
+Markers are annotations that are defined with a time range, a color, a category and an optional label. The markers are displayed in the time graph of any view that extends ''AbstractTimeGraphView''. The markers are drawn as a line or a region (in case the time range duration is not zero) of the given color, which can have an alpha value to use transparency. The markers can be drawn in the foreground (above time graph states) or in the background (below time graph states). An optional label can be drawn in the the time scale area.
+
+The developer can add trace-specific markers and/or view-specific markers.
+
+== Trace-specific markers ==
+
+Trace-specific markers can be added by registering an ''IAdapterFactory'' with the TmfTraceAdapterManager. The adapter factory must provide adapters of the ''IMarkerEventSource'' class for a given ''ITmfTrace'' object. The adapter factory can be registered for traces of a certain class (which will include sub-classes of the given class) or it can be registered for traces of a certain trace type id (as defined in the ''org.eclipse.linuxtools.tmf.core.tracetype'' extension point).
+
+The adapter factory can be registered in the ''Activator'' of the plug-in that introduces it, in the ''start()'' method, and unregistered in the ''stop()'' method.
+
+It is recommended to extend the ''AbstractTmfTraceAdapterFactory'' class when creating the adapter factory. This will ensure that a single instance of the adapter is created for a specific trace and reused by all components that need the adapter, and that the adapter is disposed when the trace is closed.
+
+The adapter implementing the ''IMarkerEventSource'' interface must provide two methods:
+
+* ''getMarkerCategories()'' returns a list of category names which will be displayed to the user, who can then enable or disable markers on a per-category basis.
+
+* ''getMarkerList()'' returns a list of markers instances of class ''IMarkerEvent'' for the given category and time range. The resolution can be used to limit the number of markers returned for the current zoom level, and the progress monitor can be checked for early cancellation of the marker computation.
+
+The trace-specific markers for a particular trace will appear in all views extending ''AbstractTimeGraphView'' when that trace (or an experiment containing that trace) is selected.
+
+An example of a trace-specific markers implementation can be seen by examining classes ''LostEventsMarkerEventSourceFactory'', ''LostEventsMarkerEventSource'' and ''Activator'' in the ''org.eclipse.tracecompass.tmf.ui'' plug-in.
+
+== View-specific markers ==
+
+View-specific markers can by added in sub-classes of ''AbstractTimeGraphView'' by implementing the following two methods:
+
+* ''getViewMarkerCategories()'' returns a list of category names which will be displayed to the user, who can then enable or disable markers on a per-category basis.
+
+* ''getViewMarkerList()'' returns a list of markers instances of class ''IMarkerEvent'' for the given time range. The resolution can be used to limit the number of markers returned for the current zoom level, and the progress monitor can be checked for early cancellation of the marker computation.
+
+= Virtual Machine Analysis =
+
+Virtualized environment are becoming more popular and understanding them can be challenging as machines share resources (CPU, disks, memory, etc), but from their point of view, they are running on bare metal. Tracing all the machines (guests and hosts) in a virtualized environment allows information to be correlated between all the nodes to better understand the system. See the User documentation for more info on this analysis.
+
+The virtual machine analysis has been implemented in the following plugins:
+
+* '''org.eclipse.tracecompass.lttng2.kernel.core''' contains the virtual machine analysis itself, the model of the virtualized environment, as well as its implementation for different hypervisors.
+* '''org.eclipse.tracecompass.lttng2.kernel.ui''' contains the views for the analysis.
+
+== Adding support for an hypervisor ==
+
+Supporting a new hypervisor in Trace Compass requires implementing the model for this new hypervisor. The following sections will describe for each part of the model what has to be considered, what information we need to have, etc. Note that each hypervisor will require some work and investigation. The information might already be available as a single tracepoint for some, while other may require many tracepoints. It is also possible that some will require to add tracepoints, either to the kernel, or the hypervisor code itself, in which case a userspace trace (LTTng UST) might be necessary to get all the information.
+
+=== Virtual CPU analysis ===
+
+This analysis tracks the state of the virtual CPUs in conjunction with the physical CPU it is running on. For this, we need the following information:
+
+* A way to link a virtual CPU on a guest with a process on the host, such that it is possible to determine when the virtual CPU is preempted on the host. If trace data does not provide this information, some hypervisors have a command line option to dump that information. Manually feeding that information to the analysis is not supported now though.
+* A way to differentiate between hypervisor mode and normal mode for the virtual CPU. A virtual CPU usually runs within a process on the host, but sometimes that process may need to run hypervisor-specific code. That is called '''hypervisor mode'''. During that time, no code from the guest itself is run. Typically, the process is running on the host (not preempted), but from the guest's point of view, the virtual CPU should be preempted.
+
+A model implementation for a new hypervisor will need to implement class '''IVirtualMachineModel''', that can be found in package '''org.eclipse.tracecompass.internal.lttng2.kernel.core.analysis.vm.model'''. See the javadoc in the class itself for more information on what each method does.
+
+= JUL Logging =
+
+Logging can be quite useful to debug a class, see its interactions with other components and understand the behavior of the system. TraceCompass uses JUL to log various events in the code, which can then be used to model and analyze the system's workflow. Here are some guidelines to use logging efficiently in Trace Compass. See the User Documentation for instructions on how to enable logging and obtain traces.
+
+=== Use a static logger for each class ===
+
+Each class should define and use their own static logger like this:
+
+ private static final Logger LOGGER = TraceCompassLog.getLogger(StateSystem.class);
+
+It is then easy to filter the components to log by their full class name. The ''TraceCompassLog#getLogger'' method is a wrapper for ''java.util.logging.Logger#getLogger'', but the Trace Compass's logging initialization (overriding the default's ConsoleHandler and INFO level for the org.eclipse.tracecompass namespace when logging is not enabled) is done in the static initializer of this class. Using the wrapper method ensures that this code is called and the user will not see Console message all over the place.
+
+'''Note on abstract classes''': It is debatable whether to use a static logger with the abstract class name or a logger with the concrete class's name.
+
+In the former case, logging for this class uses the classes's own namespace, but it is impossible to discriminate logging statement by concrete classes unless the concrete class name is added as parameter to the statement (when necessary).
+
+The latter case has the advantage that one can log only the concrete class and see all that goes on in the abstract class as well, but the concrete class may be in another namespace and will not benefit from the ''TraceCompassLog'' logging initialization and the user will see console logging happen.
+
+Both methods have their advantages and there is no clear good answer.
+
+=== Use a message supplier ===
+
+A logging statement, to be meaningful, will usually log a string that contains data from the context and will thus do string concatenation. This has a non-negligible overhead. To avoid having to do the costly string concatenation when the statement is not logged, java provides method taking a ''Supplier<String>'' as argument and that method should be used for all logging statements
+
+ LOGGER.info(() -> "[Component:Action] param1=" + myParam1 + ", param2=" + myParam2);
+
+=== Choose the appropriate log level ===
+
+The available log levels for JUL are SEVERE, WARNING, INFO, CONFIG, FINE, FINER, FINEST. The default level when not specified is INFO.
+
+* As a rule of thumb, enabling all INFO level statements should have a near zero impact on the execution, so log parameters that require some computations, or methods that are called very often should not be logged at INFO level.
+* CONFIG level should provide more detailed information than the INFO level, but still not impact the execution too much. It should be possible for a component to use up to CONFIG level statements and make meaningful analyses using the timestamps of the events.
+* FINE, FINER and FINEST are for statements that will not be used with the timestamps. Enabling them may have a visible effect on the performance of Trace Compass. They will typically be used with a purpose in mind, like debugging a component or getting data on caches for examples.
+
+=== Log message format ===
+
+JUL logging will produce trace data and unless one wants to visually parse a trace one event at a time, it will typically be used with an analysis to produce a result. To do so, the log messages should have a format that can then be associated with a trace type.
+
+Third party plugins provide a custom trace parser and LTTng trace type for JUL statements that use the following format
+
+ [EventName:MayContainSemiColon] paramName1=paramValue1, paramName2=paramValue2
projects / deliverable / tracecompass.git / blobdiff