diff options
| author | Gregory Nutt <gnutt@nuttx.org> | 2014-08-12 13:50:18 -0600 |
|---|---|---|
| committer | Gregory Nutt <gnutt@nuttx.org> | 2014-08-12 13:50:18 -0600 |
| commit | de88ac15376fb0d4bd844051b60035463a1f7bbc (patch) | |
| tree | 653489bff5d79a8116a0e122fadedc813f4c945f | |
| parent | 33da3d498b9d27edc0a6fdf9aa7f9933a8045a25 (diff) | |
| download | nuttx-de88ac15376fb0d4bd844051b60035463a1f7bbc.tar.gz nuttx-de88ac15376fb0d4bd844051b60035463a1f7bbc.tar.bz2 nuttx-de88ac15376fb0d4bd844051b60035463a1f7bbc.zip | |
Updated comments
| -rw-r--r-- | nuttx/sched/sched/sched_timerexpiration.c | 18 |
1 files changed, 14 insertions, 4 deletions
diff --git a/nuttx/sched/sched/sched_timerexpiration.c b/nuttx/sched/sched/sched_timerexpiration.c index c4eeaafbb..b7702804d 100644 --- a/nuttx/sched/sched/sched_timerexpiration.c +++ b/nuttx/sched/sched/sched_timerexpiration.c @@ -66,7 +66,7 @@ * that the read-to-run list is modified in order to muck with timers. * * The kludge/work-around is simple to keep the timer running all of the - * time with an interval of no more than the timeslice interface. If we + * time with an interval of no more than the timeslice interval. If we * this, then there is really no need to do anything when on context * switches. */ @@ -684,12 +684,22 @@ void sched_timer_resume(void) * - If the watchdog at the head of the expiration list changes (or if its * delay changes. This can occur as a consequence of the actions of * wd_start() or wd_cancel(). - * - Any context switch occurs, changing the task at the head of the - * ready-to-run list. The task at the head of list may require - * different timeslice processing (or no timeslice at all). * - When pre-emption is re-enabled. A previous time slice may have * expired while pre-emption was enabled and now needs to be executed. * + * In the original design, it was also planned that sched_timer_reasses() + * be called whenever there was a change at the head of the ready-to-run + * list. That call was intended to establish a new time-slice for the + * newly activated task or to stop the timer if time-slicing is no longer + * needed. However, it turns out that that solution is too fragile: The + * system is too vulnerable at the time that the read-to-run list is + * modified in order to muck with timers. + * + * The kludge/work-around is simple to keep the timer running all of the + * time with an interval of no more than the timeslice interval. If we + * this, then there is really no need to do anything when on context + * switches. + * * Input Parameters: * None * |