[GitHub] srowen commented on a change in pull request #23424: [SPARK-24421][CORE][FOLLOWUP] Use normal direct ByteBuffer allocation if Cleaner can't be set

2019-01-02 Thread GitBox
srowen commented on a change in pull request #23424: 
[SPARK-24421][CORE][FOLLOWUP] Use normal direct ByteBuffer allocation if 
Cleaner can't be set
URL: https://github.com/apache/spark/pull/23424#discussion_r244918136
 
 

 ##
 File path: common/unsafe/src/main/java/org/apache/spark/unsafe/Platform.java
 ##
 @@ -209,21 +209,25 @@ public static long reallocateMemory(long address, long 
oldSize, long newSize) {
   }
 
   /**
-   * Uses internal JDK APIs to allocate a DirectByteBuffer while ignoring the 
JVM's
-   * MaxDirectMemorySize limit (the default limit is too low and we do not 
want to require users
-   * to increase it).
+   * Allocate a DirectByteBuffer, potentially bypassing the JVM's 
MaxDirectMemorySize limit.
*/
   public static ByteBuffer allocateDirectBuffer(int size) {
 try {
+  if (CLEANER_CREATE_METHOD == null) {
+// Can't set a Cleaner (see comments on field), so need to allocate 
via normal Java APIs
+return ByteBuffer.allocateDirect(size);
 
 Review comment:
   OK, it does look like it will throw OutOfMemoryError here if off-heap 
allocation fails. That's a decent hint, yes.


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] srowen commented on a change in pull request #23424: [SPARK-24421][CORE][FOLLOWUP] Use normal direct ByteBuffer allocation if Cleaner can't be set

2019-01-01 Thread GitBox
srowen commented on a change in pull request #23424: 
[SPARK-24421][CORE][FOLLOWUP] Use normal direct ByteBuffer allocation if 
Cleaner can't be set
URL: https://github.com/apache/spark/pull/23424#discussion_r244661628
 
 

 ##
 File path: common/unsafe/src/main/java/org/apache/spark/unsafe/Platform.java
 ##
 @@ -209,21 +209,25 @@ public static long reallocateMemory(long address, long 
oldSize, long newSize) {
   }
 
   /**
-   * Uses internal JDK APIs to allocate a DirectByteBuffer while ignoring the 
JVM's
-   * MaxDirectMemorySize limit (the default limit is too low and we do not 
want to require users
-   * to increase it).
+   * Allocate a DirectByteBuffer, potentially bypassing the JVM's 
MaxDirectMemorySize limit.
*/
   public static ByteBuffer allocateDirectBuffer(int size) {
 try {
+  if (CLEANER_CREATE_METHOD == null) {
+// Can't set a Cleaner (see comments on field), so need to allocate 
via normal Java APIs
+return ByteBuffer.allocateDirect(size);
+  }
+  // Otherwise, use internal JDK APIs to allocate a DirectByteBuffer while 
ignoring the JVM's
+  // MaxDirectMemorySize limit (the default limit is too low and we do not 
want to
+  // require users to increase it).
   long memory = allocateMemory(size);
   ByteBuffer buffer = (ByteBuffer) DBB_CONSTRUCTOR.newInstance(memory, 
size);
-  if (CLEANER_CREATE_METHOD != null) {
-try {
-  DBB_CLEANER_FIELD.set(buffer,
-  CLEANER_CREATE_METHOD.invoke(null, buffer, (Runnable) () -> 
freeMemory(memory)));
-} catch (IllegalAccessException | InvocationTargetException e) {
-  throw new IllegalStateException(e);
-}
+  try {
+DBB_CLEANER_FIELD.set(buffer,
+CLEANER_CREATE_METHOD.invoke(null, buffer, (Runnable) () -> 
freeMemory(memory)));
+  } catch (IllegalAccessException | InvocationTargetException e) {
+freeMemory(memory);
 
 Review comment:
   Just to be totally safe, free the memory that was allocated but can't be 
used now in this case.


This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org