Hi Yunda,
Thanks for fixing this. I actually have the exact same fix for this
problem, along with fixes for some other problems in SA in my local
workspace, but haven't come around to sending it out for public review yet.
I think this is just a missed change in JDK-7054512. I have created a
JIRA issue for you: JDK-8008796: SA: Oop.iterateFields() should support
CompressedKlassPointers again. I'll prepare a webrev and send it out for
official review later.
Thanks,
Kris
On 2013/2/25 2:58, 云达(Yunda) wrote:
Hi all,
When I used CLHSDB just now I met this error:
hsdb> inspect 0x00000000ee255080
instance of "java/io/InputStream" @ 0x00000000ee255080 @
0x00000000ee255080 (size = 24)
Exception in thread "main" java.lang.InternalError: unimplemented
at sun.jvm.hotspot.oops.Oop.iterateFields(Oop.java:151)
at sun.jvm.hotspot.oops.Instance.iterateFields(Instance.java:66)
at sun.jvm.hotspot.oops.Oop.iterate(Oop.java:143)
at
sun.jvm.hotspot.ui.tree.OopTreeNodeAdapter.getChildCount(OopTreeNodeAdapter.java:65)
at
sun.jvm.hotspot.CommandProcessor$Command.printNode(CommandProcessor.java:231)
at
sun.jvm.hotspot.CommandProcessor$24.doit(CommandProcessor.java:1008)
at
sun.jvm.hotspot.CommandProcessor.executeCommand(CommandProcessor.java:1897)
at
sun.jvm.hotspot.CommandProcessor.executeCommand(CommandProcessor.java:1867)
at
sun.jvm.hotspot.CommandProcessor.run(CommandProcessor.java:1747)
at sun.jvm.hotspot.CLHSDB.run(CLHSDB.java:91)
at sun.jvm.hotspot.CLHSDB.main(CLHSDB.java:35)
I found it’s caused by the code of
sun.jvm.hotspot.oops.Oop.iterateFields():
void iterateFields(OopVisitor visitor, boolean doVMFields) {
if (doVMFields) {
visitor.doCInt(mark, true);
if (VM.getVM().isCompressedKlassPointersEnabled()) {
*throw new InternalError("unimplemented");*
} else {
visitor.doMetadata(klass, true);
}
}
}
When compressed oops( which is by default) are used an InternalError
of “unimplemented” will be throwed. But actually it can be implemented
easily by just one line of code:
visitor.doMetadata(compressedKlass, true);
I checked the hotspot-rt repo and I found it was implemented this way
before changeset 3601, the main implementation of NPG. Since 3601
changed a whole lot of stuff, the ‘compressedKlass’ field couldn’t get
the right value, as in line 51:
// compressedKlass = new
CIntField(type.getCIntegerField("_metadata._compressed_klass"), 0);
So the code in iterateFields() was changed to throwing an
InternalError accordingly.
But the ‘compressedKlass’ field can get the right value now. So I
think it’s time to change the code back and here’s the diff against
the latest hotspot-rt:
diff -r 2f881161d085 agent/src/share/classes/sun/jvm/hotspot/oops/Oop.java
--- a/agent/src/share/classes/sun/jvm/hotspot/oops/Oop.java Mon Feb 25
18:25:24 2013 +0800
+++ b/agent/src/share/classes/sun/jvm/hotspot/oops/Oop.java Mon Feb 25
18:47:08 2013 +0800
@@ -148,7 +148,7 @@
if (doVMFields) {
visitor.doCInt(mark, true);
if (VM.getVM().isCompressedKlassPointersEnabled()) {
- throw new InternalError("unimplemented");
+ visitor.doMetadata(compressedKlass, true);
} else {
visitor.doMetadata(klass, true);
}
Regards,
Yunda
------------------------------------------------------------------------
This email (including any attachments) is confidential and may be
legally privileged. If you received this email in error, please delete
it immediately and do not copy it or use it for any purpose or
disclose its contents to any other person. Thank you.
本电邮(包括任何附件)可能含有机密资料并受法律保护。如您不是正确的收件
人，请您立即删除本邮件。请不要将本电邮进行复制并用作任何其 他用途、或
透露本邮件之内容。谢谢。
