Hotfix double entity removal making entity scheduler retire call

The assumption that the setRemoved method will never be called more than once is flawed, considering even vanilla code seems to indicate it might happen. Especially with plugins, throwing an exception is not something reasonably maintainable across all the places it *could* happen.
If it is called a second time after already having been removed due to changing dimensions, that's definitely bad, so no extra check for that
Fixes #9420

patches/server/Folia-scheduler-and-owned-region-API.patch

@@ -1194,12 +1194,20 @@ index 0000000000000000000000000000000000000000..00000000000000000000000000000000
          this.remove(Entity.RemovalReason.DISCARDED);
      }
  
+@@ -0,0 +0,0 @@ public abstract class Entity implements Nameable, EntityAccess, CommandSource {
+             return;
+         }
+         // Paper end - rewrite chunk system
++        final boolean alreadyRemoved = this.removalReason != null;
+         if (this.removalReason == null) {
+             this.removalReason = reason;
+         }
 @@ -0,0 +0,0 @@ public abstract class Entity implements Nameable, EntityAccess, CommandSource {
  
          if (reason != RemovalReason.UNLOADED_TO_CHUNK) this.getPassengers().forEach(Entity::stopRiding); // Paper - chunk system - don't adjust passenger state when unloading, it's just not safe (and messes with our logic in entity chunk unload)
          this.levelCallback.onRemove(reason);
 +        // Paper start - Folia schedulers
-+        if (!(this instanceof ServerPlayer) && reason != RemovalReason.CHANGED_DIMENSION) {
++        if (!(this instanceof ServerPlayer) && reason != RemovalReason.CHANGED_DIMENSION && !alreadyRemoved) {
 +            // Players need to be special cased, because they are regularly removed from the world
 +            this.retireScheduler();
 +        }