PaperMC/patches/server/0350-Set-cap-on-JDK-per-thread-native-byte-buffer-cache.patch
Bjarne Koll c5a10665b8
Remove wall-time / unused skip tick protection (#11412)
Spigot still maintains some partial implementation of "tick skipping", a
practice in which the MinecraftServer.currentTick field is updated not
by an increment of one per actual tick, but instead set to
System.currentTimeMillis() / 50. This behaviour means that the tracked
tick may "skip" a tick value in case a previous tick took more than the
expected 50ms.

To compensate for this in important paths, spigot/craftbukkit
implements "wall-time". Instead of incrementing/decrementing ticks on
block entities/entities by one for each call to their tick() method,
they instead increment/decrement important values, like
an ItemEntity's age or pickupDelay, by the difference of
`currentTick - lastTick`, where `lastTick` is the value of
`currentTick` during the last tick() call.

These "fixes" however do not play nicely with minecraft's simulation
distance as entities/block entities implementing the above behaviour
would "catch up" their values when moving from a non-ticking chunk to a
ticking one as their `lastTick` value remains stuck on the last tick in
a ticking chunk and hence lead to a large "catch up" once ticked again.

Paper completely removes the "tick skipping" behaviour (See patch
"Further-improve-server-tick-loop"), making the above precautions
completely unnecessary, which also rids paper of the previous described
incompatibility with non-ticking chunks.
2024-09-19 16:36:07 +02:00

30 lines
1.5 KiB
Diff

From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Aikar <aikar@aikar.co>
Date: Mon, 4 May 2020 01:08:56 -0400
Subject: [PATCH] Set cap on JDK per-thread native byte buffer cache
See: https://www.evanjones.ca/java-bytebuffer-leak.html
This is potentially a source of lots of native memory usage.
We are clearly seeing native usage upwards to 1-4GB which doesn't make sense.
Region File usage fixed in previous patch should of tecnically only been somewhat
temporary until GC finally gets it some time later, but between all the various
plugins doing IO on various threads, this hidden detail of the JDK could be
keeping long lived large direct buffers in cache.
Set system properly at server startup if not set already to help protect from this.
diff --git a/src/main/java/org/bukkit/craftbukkit/Main.java b/src/main/java/org/bukkit/craftbukkit/Main.java
index ed167d0d399924d54d9ff99c10ab8ee093efc149..168cbb239ac5d632908f2b0aca82cbcfdc35651f 100644
--- a/src/main/java/org/bukkit/craftbukkit/Main.java
+++ b/src/main/java/org/bukkit/craftbukkit/Main.java
@@ -27,6 +27,7 @@ public class Main {
}
// Paper end
// Todo: Installation script
+ if (System.getProperty("jdk.nio.maxCachedBufferSize") == null) System.setProperty("jdk.nio.maxCachedBufferSize", "262144"); // Paper - cap per-thread NIO cache size; https://www.evanjones.ca/java-bytebuffer-leak.html
OptionParser parser = new OptionParser() {
{
this.acceptsAll(Main.asList("?", "help"), "Show the help");