MINOR: sched: add TASK_F_WANTS_TIME to make the scheduler update the call date

Currently tasks being profiled have th_ctx->sched_call_date set to the
current nanosecond in monotonic time. But there's no other way to have
this, despite the scheduler being capable of it. Let's just declare a
new task flag, TASK_F_WANTS_TIME, that makes the scheduler take the time
just before calling the handler. This way, a task that needs nanosecond
resolution on the call date will be able to be called with an up-to-date
date without having to abuse now_mono_time() if not needed. In addition,
if CLOCK_MONOTONIC is not supported (now_mono_time() always returns 0),
the date is set to the most recently known now_ns, which is guaranteed
to be atomic and is only updated once per poll loop.

This date can be more conveniently retrieved using task_mono_time().

This can be useful, e.g. for pacing. The code was slightly adjusted so
as to merge the common parts between the profiling case and this one.

Author: wtarreau

include/haproxy/task-t.h

@@ -60,11 +60,13 @@
 #define TASK_F_USR1       0x00010000  /* preserved user flag 1, application-specific, def:0 */
 #define TASK_F_UEVT1      0x00020000  /* one-shot user event type 1, application specific, def:0 */
 #define TASK_F_UEVT2      0x00040000  /* one-shot user event type 2, application specific, def:0 */
-/* unused: 0x80000..0x80000000 */
+#define TASK_F_WANTS_TIME 0x00080000  /* task/tasklet wants th_ctx->sched_call_date to be set */
+/* unused: 0x100000..0x80000000 */
 
 /* These flags are persistent across scheduler calls */
 #define TASK_PERSISTENT   (TASK_SELF_WAKING | TASK_KILLED | \
-                           TASK_HEAVY | TASK_F_TASKLET | TASK_F_USR1)
+                           TASK_HEAVY | TASK_F_TASKLET | TASK_F_USR1 | \
+                           TASK_F_WANTS_TIME)
 
 /* This function is used to report state in debugging tools. Please reflect
  * below any single-bit flag addition above in the same order via the

include/haproxy/task.h

@@ -193,6 +193,12 @@ static inline int thread_has_tasks(void)
 		(int)!MT_LIST_ISEMPTY(&th_ctx->shared_tasklet_list));
 }
 
+/* returns the most recent known date of the task's call from the scheduler */
+static inline uint64_t task_mono_time(void)
+{
+	return th_ctx->sched_call_date;
+}
+
 /* puts the task <t> in run queue with reason flags <f>, and returns <t> */
 /* This will put the task in the local runqueue if the task is only runnable
  * by the current thread, in the global runqueue otherwies. With DEBUG_TASK,

src/task.c

@@ -567,17 +567,24 @@ unsigned int run_tasks_from_lists(unsigned int budgets[])
 		t->calls++;
 
 		th_ctx->sched_wake_date = t->wake_date;
-		if (th_ctx->sched_wake_date) {
-			uint32_t now_ns = now_mono_time();
-			uint32_t lat = now_ns - th_ctx->sched_wake_date;
+		if (th_ctx->sched_wake_date || (t->state & TASK_F_WANTS_TIME)) {
+			/* take the most accurate clock we have, either
+			 * mono_time() or last now_ns (monotonic but only
+			 * incremented once per poll loop).
+			 */
+			th_ctx->sched_call_date = now_mono_time();
+			if (unlikely(!th_ctx->sched_call_date))
+				th_ctx->sched_call_date = now_ns;
+		}
 
+		if (th_ctx->sched_wake_date) {
 			t->wake_date = 0;
-			th_ctx->sched_call_date = now_ns;
 			profile_entry = sched_activity_entry(sched_activity, t->process, t->caller);
 			th_ctx->sched_profile_entry = profile_entry;
-			HA_ATOMIC_ADD(&profile_entry->lat_time, lat);
+			HA_ATOMIC_ADD(&profile_entry->lat_time, (uint32_t)(th_ctx->sched_call_date - th_ctx->sched_wake_date));
 			HA_ATOMIC_INC(&profile_entry->calls);
 		}
+
 		__ha_barrier_store();
 
 		th_ctx->current = t;