Hi Daniel,
On 1/02/2018 2:45 AM, stewartd.qdt wrote:
Hi Jini, David,
Please have a look at the revised webrev:
http://cr.openjdk.java.net/~dstewart/8196361/webrev.01/
In this webrev I have changed the approach to finding the addresses. This was
necessary because in the case of matching for the locks the addresses are
before what is matched and in the case of Method the address is after it. The
existing code only looked for the addresses after the matched string. I've also
tried to align what tokens are being looked for in the lock case. I've taken
an approach of breaking the jstack output into lines and then searching each
line for it containing what we want. Once found, the line is broken into pieces
to find the actual address we want.
Please let me know if this is an unacceptable approach or any changes you would
like to see.
I'm not clear on the overall approach as I'm unclear exactly how inspect
operates or exactly what the test is trying to verify. One comment on
breaking things into lines though:
73 String newline = System.getProperty("line.separator");
74 String[] lines = jstackOutput.split(newline);
As split() takes a regex, I suggest using \R to cover all potential
line-breaks, rather than the platform specific line-seperator. We've
been recently bitten by the distinction between output that comes from
reading a process's stdout/stderr (and for which a newline \n is
translated into the platform line-seperator), and output that comes
across a socket connection (for which \n is not translated). This could
result in failing to parse things correctly on Windows. It's
safer/simpler to expect any kind of line-seperator.
Thanks,
David
Thanks,
Daniel
-----Original Message-----
From: Jini George [mailto:jini.geo...@oracle.com]
Sent: Tuesday, January 30, 2018 6:58 AM
To: David Holmes <david.hol...@oracle.com>; stewartd.qdt
<stewartd....@qualcommdatacenter.com>
Cc: serviceability-dev <serviceability-dev@openjdk.java.net>;
hotspot-...@openjdk.java.net
Subject: Re: RFR: 8196361: JTReg failure in serviceability/sa/ClhsdbInspect.java
Hi Daniel, David,
Thanks, Daniel, for bringing this up. The intent of the test is to get the oop
address corresponding to a java.lang.ref.ReferenceQueue$Lock,
which can typically be obtained from the stack traces of the Common-Cleaner or
the Finalizer threads. The stack traces which I had been noticing were
typically of the form:
"Common-Cleaner" #8 daemon prio=8 tid=0x00007f09c82ac000 nid=0xf6e in
Object.wait() [0x00007f09a18d2000]
java.lang.Thread.State: TIMED_WAITING (on object monitor)
JavaThread state: _thread_blocked
- java.lang.Object.wait(long) @bci=0, pc=0x00007f09b7d6480b,
Method*=0x00007f09acc43d60 (Interpreted frame)
- waiting on <0x000000072e61f6e0> (a
java.lang.ref.ReferenceQueue$Lock)
- java.lang.ref.ReferenceQueue.remove(long) @bci=59, line=151,
pc=0x00007f09b7d55243, Method*=0x00007f09acdab9b0 (Interpreted frame)
- waiting to re-lock in wait() <0x000000072e61f6e0> (a
java.lang.ref.ReferenceQueue$Lock)
...
I chose 'waiting to re-lock in wait' since that was what I had been observing
next to the oop address of java.lang.ref.ReferenceQueue$Lock.
But I see how with a timing difference, one could get 'waiting to lock'
as in your case. So, a good way to fix might be to check for the line
containing '(a java.lang.ref.ReferenceQueue$Lock)', getting the oop address
from that line (should be the address appearing immediately before '(a
java.lang.ref.ReferenceQueue$Lock)') and passing that to the 'inspect' command.
Thanks much,
Jini.
On 1/30/2018 3:35 AM, David Holmes wrote:
Hi Daniel,
Serviceability issues should go to serviceability-dev@openjdk.java.net
- now cc'd.
On 30/01/2018 7:53 AM, stewartd.qdt wrote:
Please review this webrev [1] which attempts to fix a test error in
serviceability/sa/ClhsdbInspect.java when it is run under an AArch64
system (not necessarily exclusive to this system, but it was the
system under test). The bug report [2] provides further details.
Essentially the line "waiting to re-lock in wait" never actually
occurs. Instead I have the line "waiting to lock" which occurs for
the referenced item of /java/lang/ref/ReferenceQueue$Lock.
Unfortunately the test is written such that only the first "waiting to lock"
occurrence is seen (for java/lang/Class), which is already accounted
for in the test.
I can't tell exactly what the test expects, or why, but it would be
extremely hard to arrange for "waiting to re-lock in wait" to be seen
for the ReferenceQueue lock! That requires acquiring the lock
yourself, issuing a notify() to unblock the wait(), and then issuing
the jstack command while still holding the lock!
David
-----
I'm not overly happy with this approach as it actually removes a test
line. However, the test line does not actually appear in the output
(at least on my system) and the test is not currently written to look
for the second occurrence of the line "waiting to lock". Perhaps the
original author could chime in and provide further guidance as to the
intention of the test.
I am happy to modify the patch as necessary.
Regards,
Daniel Stewart
[1] - http://cr.openjdk.java.net/~dstewart/8196361/webrev.00/
[2] - https://bugs.openjdk.java.net/browse/JDK-8196361
