[11u] RFC: Stack dump improvements

[Lindenmaier, Goetz]

Hi,

I think this are useful improvements to the thread dumps
which are also on our side heavily used.
The changes are of limited scope, I would not expect 
any problems with this.

Best regards,
  Goetz.


> -----Original Message-----
> From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.java.net> On
> Behalf Of Aleksey Shipilev
> Sent: Wednesday, January 26, 2022 12:00 PM
> To: jdk-updates-dev at openjdk.java.net
> Subject: [11u] RFC: Stack dump improvements
> 
> Hi,
> 
> I was looking at customer support case in 11u recently, which looks like yet
> another a class
> initialization deadlock. JDK 12+ has two very good improvements to stack
> dumps, would these
> backports be acceptable?
> 
> I am asking because the dependency tree between these issues is a bit
> complicated, and it would
> probably raise questions if I suggest 6 dependent backports at once. So I
> though I'd give an
> overview here.
> 
> The dependency tree looks like this:
> 
>   - JDK-8150689: Thread dump report "waiting to re-lock in wait()" shows
> incorrectly
>     - JDK-8214499: SA should follow 8150689
>     - JDK-8214148: [TESTBUG]
> serviceability/tmtools/jstack/WaitNotifyThreadTest.java...
>        - JDK-8215050: [TESTBUG]
> serviceability/tmtools/jstack/WaitNotifyThreadTest.java...
>     - JDK-8213397: Stack dump should show more clearly when a thread is
> blocked...
>        - JDK-8217337: [TESTBUG] TestThreadDumpClassInitMonitor.java fails
> with Unexpected
> 
> The key issues are JDK-8150689 and JDK-8213397. JDK-8150689 is apparently a
> regression from JDK 9.
> Doing JDK-8150689 before JDK-8213397 makes all backports clean. The rest
> are SA/testbug fixups.
> 
> The overall change is not large:
> 
> $ git diff master --stat
>   src/hotspot/share/oops/instanceKlass.cpp                                    |  23 +++++----
>   src/hotspot/share/runtime/thread.cpp                                        |   3 +-
>   src/hotspot/share/runtime/thread.hpp                                        |  10 ++++
>   src/hotspot/share/runtime/thread.inline.hpp                                 |  12 +++++
>   src/hotspot/share/runtime/vframe.cpp                                        |  33 +++++------
> -
> 
> src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/oops/OopUtilities.jav
> a  |  22 ++++----
> 
> src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/runtime/JavaVFrame
> .java |  14 ++----
>   test/hotspot/jtreg/runtime/Thread/TestThreadDumpClassInitMonitor.java
> | 202
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> ++++++++++++++++
>   test/hotspot/jtreg/serviceability/tmtools/jstack/WaitNotifyThreadTest.java
> |  53 ++++++++++----------
>   9 files changed, 293 insertions(+), 79 deletions(-)
> 
> I ran the composite patch through tier{1,2,3} and jdk_svc on Linux x86_64
> {fastdebug,release}
> without problems.
> 
> Thoughts?
> 
> --
> Thanks,
> -Aleksey