cooker process: fire heartbeat event at regular time intervals

The intended usage is for recording current system statistics from
/proc in buildstats.bbclass during a build and for improving the
BB_DISKMON_DIRS implementation.

All other existing hooks are less suitable because they trigger at
unpredictable rates: too often can be handled by doing rate-limiting
in the event handler, but not often enough (for example, when there is
only one long-running task) cannot because the handler does not get
called at all.

The implementation of the new heartbeat event hooks into the cooker
process event queue. The process already wakes up every 0.1s, which is
often enough for the intentionally coarse 1s delay between
heartbeats. That value was chosen to keep the overhead low while still
being frequent enough for the intended usage.

If necessary, BB_HEARTBEAT_EVENT can be set to a float specifying
the delay in seconds between these heartbeat events.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

4 files changed, 39 insertions, 0 deletions

diff --git a/lib/bb/event.py b/lib/bb/event.py
index 6f1cb101f..cacbac8f5 100644
--- a/lib/bb/event.py
+++ b/lib/bb/event.py
@@ -48,6 +48,16 @@ class Event(object):
     def __init__(self):
         self.pid = worker_pid
 
+
+class HeartbeatEvent(Event):
+    """Triggered at regular time intervals of 10 seconds. Other events can fire much more often
+       (runQueueTaskStarted when there are many short tasks) or not at all for long periods
+       of time (again runQueueTaskStarted, when there is just one long-running task), so this
+       event is more suitable for doing some task-independent work occassionally."""
+    def __init__(self, time):
+        Event.__init__(self)
+        self.time = time
+
 Registered        = 10
 AlreadyRegistered = 14
 
diff --git a/lib/bb/server/process.py b/lib/bb/server/process.py
index 982fcf71c..1654faf92 100644
--- a/lib/bb/server/process.py
+++ b/lib/bb/server/process.py
@@ -92,6 +92,8 @@ class ProcessServer(Process, BaseImplServer):
         self.event = EventAdapter(event_queue)
         self.featurelist = featurelist
         self.quit = False
+        self.heartbeat_seconds = 1 # default, BB_HEARTBEAT_EVENT will be checked once we have a datastore.
+        self.next_heartbeat = time.time()
 
         self.quitin, self.quitout = Pipe()
         self.event_handle = multiprocessing.Value("i")
@@ -101,6 +103,14 @@ class ProcessServer(Process, BaseImplServer):
             self.event_queue.put(event)
         self.event_handle.value = bb.event.register_UIHhandler(self, True)
 
+        heartbeat_event = self.cooker.data.getVar('BB_HEARTBEAT_EVENT', True)
+        if heartbeat_event:
+            try:
+                self.heartbeat_seconds = float(heartbeat_event)
+            except:
+                # Throwing an exception here causes bitbake to hang.
+                # Just warn about the invalid setting and continue
+                bb.warn('Ignoring invalid BB_HEARTBEAT_EVENT=%s, must be a float specifying seconds.' % heartbeat_event)
         bb.cooker.server_main(self.cooker, self.main)
 
     def main(self):
@@ -160,6 +170,21 @@ class ProcessServer(Process, BaseImplServer):
                 del self._idlefuns[function]
                 self.quit = True
 
+        # Create new heartbeat event?
+        now = time.time()
+        if now >= self.next_heartbeat:
+            # We might have missed heartbeats. Just trigger once in
+            # that case and continue after the usual delay.
+            self.next_heartbeat += self.heartbeat_seconds
+            if self.next_heartbeat <= now:
+                self.next_heartbeat = now + self.heartbeat_seconds
+            heartbeat = bb.event.HeartbeatEvent(now)
+            bb.event.fire(heartbeat, self.cooker.data)
+        if nextsleep and now + nextsleep > self.next_heartbeat:
+            # Shorten timeout so that we we wake up in time for
+            # the heartbeat.
+            nextsleep = self.next_heartbeat - now
+
         if nextsleep is not None:
             select.select(fds,[],[],nextsleep)
 
diff --git a/lib/bb/ui/knotty.py b/lib/bb/ui/knotty.py
index 948f52769..48e1223c6 100644
--- a/lib/bb/ui/knotty.py
+++ b/lib/bb/ui/knotty.py
@@ -647,6 +647,7 @@ def main(server, eventHandler, params, tf = TerminalFilter):
                                   bb.event.OperationCompleted,
                                   bb.event.OperationProgress,
                                   bb.event.DiskFull,
+                                  bb.event.HeartbeatEvent,
                                   bb.build.TaskProgress)):
                 continue
 
diff --git a/lib/bb/ui/toasterui.py b/lib/bb/ui/toasterui.py
index b1b3684a8..17299026a 100644
--- a/lib/bb/ui/toasterui.py
+++ b/lib/bb/ui/toasterui.py
@@ -236,6 +236,9 @@ def main(server, eventHandler, params):
             # pylint: disable=protected-access
             # the code will look into the protected variables of the event; no easy way around this
 
+            if isinstance(event, bb.event.HeartbeatEvent):
+                continue
+
             if isinstance(event, bb.event.ParseStarted):
                 if not (build_log and build_log_file_path):
                     build_log, build_log_file_path = _open_build_log(log_dir)