Upstream/NoCheatPlus
1
0
Fork 0
mirror of https://github.com/NoCheatPlus/NoCheatPlus.git synced 2024-06-26 06:14:42 +02:00
Code Issues Projects Releases Wiki Activity

Fix heavyLoad constant. Include TickTask.getLag.

Browse Source 
The original idea had been to just count an abstract value for load, so
the number of permission checks per tick stays limited, however that
doesn't scale with the CPU power and other.
This commit is contained in:
asofold 2018-02-07 23:44:46 +01:00
parent 3d8eb00981
commit 4bfc4f2cd2
1 changed files with 29 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
NCPCore/src/main/java/fr/neatmonster/nocheatplus/players/PlayerData.java
View File

@ -35,6 +35,7 @@ import fr.neatmonster.nocheatplus.permissions.PermissionNode;
import fr.neatmonster.nocheatplus.permissions.PermissionPolicy.FetchingPolicy;
import fr.neatmonster.nocheatplus.permissions.PermissionRegistry;
import fr.neatmonster.nocheatplus.permissions.RegisteredPermission;
import fr.neatmonster.nocheatplus.utilities.TickTask;
import fr.neatmonster.nocheatplus.utilities.ds.corw.DualSet;
import fr.neatmonster.nocheatplus.utilities.ds.count.ActionFrequency;
import fr.neatmonster.nocheatplus.utilities.ds.map.HashMapLOW;
@ -85,8 +86,16 @@ public class PlayerData implements IData, ICanHandleTimeRunningBackwards {
/** Monitor player task load across all players (nanoseconds per server tick). */
private static ActionFrequency taskLoad = new ActionFrequency(6, 7);
/** Some measure for heavy load, for skipping some of the (lazy) updating. */
private static float heavyLoad = 1000000f; // 1 ms = 2 % of a tick.
private static final int ticksMonitored = taskLoad.numberOfBuckets() * (int) taskLoad.bucketDuration();
private static final long msMonitored = ticksMonitored * 50;
/**
* Some measure for heavy load, for skipping some of the (lazy) updating.
* Assume 1% of a tick in average as heavy - just for this task. This still
* is problematic, if System.nanoTime() just has milliseconds resolution,
* because we'll be adding up zeros most of the time here. Perhaps just
* relaying to TickTask.getLag is enough.
*/
private static float heavyLoad = 500000f / (float) ticksMonitored;
// Default tags.
public static final String TAG_NOTIFY_OFF = "notify_off";
@ -263,23 +272,24 @@ public class PlayerData implements IData, ICanHandleTimeRunningBackwards {
}
long nanos = System.nanoTime();
taskLoad.update(tick);
final boolean isHeavyLoad = taskLoad.score(1f) > heavyLoad;
updatePermissionsLazy.mergePrimaryThread();
final Iterator<RegisteredPermission> it = updatePermissionsLazy.iteratorPrimaryThread();
// TODO: Load balancing with other tasks ?
while (it.hasNext()) {
hasPermission(it.next(), player);
it.remove();
if (isHeavyLoad) {
break;
}
}
boolean hasWrk = it.hasNext();
nanos = System.nanoTime() - nanos;
if (nanos > 0L) {
taskLoad.add(tick, nanos);
}
return !hasWrk;
final boolean isHeavyLoad = taskLoad.score(1f) > heavyLoad
|| TickTask.getLag(msMonitored, true) > 1.1f;
updatePermissionsLazy.mergePrimaryThread();
final Iterator<RegisteredPermission> it = updatePermissionsLazy.iteratorPrimaryThread();
// TODO: Load balancing with other tasks ?
while (it.hasNext()) {
hasPermission(it.next(), player);
it.remove();
if (isHeavyLoad) {
break;
}
}
boolean hasWrk = it.hasNext();
nanos = System.nanoTime() - nanos;
if (nanos > 0L) {
taskLoad.add(tick, nanos);
}
return !hasWrk;
}
private void registerFrequentPlayerTask() {