Upstream/Paper
1
0
Fork 0
mirror of https://github.com/PaperMC/Paper.git synced 2024-08-20 19:33:32 +02:00
Code Issues Projects Releases Wiki Activity
1ccff6fabb
Paper/Spigot-Server-Patches/0501-Set-cap-on-JDK-per-thread-native-byte-buffer-cache.patch

34 lines
1.4 KiB
Diff
Raw Normal View History

Optimize Network Manager to not need synchronization Removes synchronization from sending packets Makes normal packet sends no longer need to be wrapped and queued like it use to work. Adds more packet queue immunities on top of keep alive to let the following scenarios go out without delay: - Keep Alive - Chat - Kick - All of the packets during the Player Joined World event Hoping that latter one helps join timeout issues more too for slow connections. Removes processing packet queue off of main thread - for the few cases where it is allowed, order is not necessary nor should it even be happening concurrently in first place (handshaking/login/status) Ensures packets sent asynchronously are dispatched on main thread This helps ensure safety for ProtocolLib as packet listeners are commonly accessing world state. This will allow you to schedule a packet to be sent async, but itll be dispatched sync for packet listeners to process. This should solve some deadlock risks This may provide a decent performance improvement because thread synchronization incurs a cache reset so by avoiding ever entering a synchronized block, we get to avoid that, and packet sending is a really hot activity.
2020-05-06 11:22:03 +02:00
From 1a054b86819314f6e6aae65af695b3f048122466 Mon Sep 17 00:00:00 2001
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.
2020-05-04 07:14:10 +02:00
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
Optimize Network Manager to not need synchronization Removes synchronization from sending packets Makes normal packet sends no longer need to be wrapped and queued like it use to work. Adds more packet queue immunities on top of keep alive to let the following scenarios go out without delay: - Keep Alive - Chat - Kick - All of the packets during the Player Joined World event Hoping that latter one helps join timeout issues more too for slow connections. Removes processing packet queue off of main thread - for the few cases where it is allowed, order is not necessary nor should it even be happening concurrently in first place (handshaking/login/status) Ensures packets sent asynchronously are dispatched on main thread This helps ensure safety for ProtocolLib as packet listeners are commonly accessing world state. This will allow you to schedule a packet to be sent async, but itll be dispatched sync for packet listeners to process. This should solve some deadlock risks This may provide a decent performance improvement because thread synchronization incurs a cache reset so by avoiding ever entering a synchronized block, we get to avoid that, and packet sending is a really hot activity.
2020-05-06 11:22:03 +02:00
index 093dbeae278..93340e9470b 100644
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.
2020-05-04 07:14:10 +02:00
--- a/src/main/java/org/bukkit/craftbukkit/Main.java
+++ b/src/main/java/org/bukkit/craftbukkit/Main.java
@@ -21,6 +21,7 @@ public class Main {
public static void main(String[] args) {
// Todo: Installation script
+ if (System.getProperty("jdk.nio.maxCachedBufferSize") == null) System.setProperty("jdk.nio.maxCachedBufferSize", "262144"); // Paper - cap per-thread NIO cache size
OptionParser parser = new OptionParser() {
{
acceptsAll(asList("?", "help"), "Show the help");
--
Optimize Network Manager to not need synchronization Removes synchronization from sending packets Makes normal packet sends no longer need to be wrapped and queued like it use to work. Adds more packet queue immunities on top of keep alive to let the following scenarios go out without delay: - Keep Alive - Chat - Kick - All of the packets during the Player Joined World event Hoping that latter one helps join timeout issues more too for slow connections. Removes processing packet queue off of main thread - for the few cases where it is allowed, order is not necessary nor should it even be happening concurrently in first place (handshaking/login/status) Ensures packets sent asynchronously are dispatched on main thread This helps ensure safety for ProtocolLib as packet listeners are commonly accessing world state. This will allow you to schedule a packet to be sent async, but itll be dispatched sync for packet listeners to process. This should solve some deadlock risks This may provide a decent performance improvement because thread synchronization incurs a cache reset so by avoiding ever entering a synchronized block, we get to avoid that, and packet sending is a really hot activity.
2020-05-06 11:22:03 +02:00
2.26.2
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.
2020-05-04 07:14:10 +02:00
Reference in New Issue Copy Permalink