RFR(S) 8215113: Sampling interval not always correct

[Jean Christophe Beyler]

Hi Serguei,

Sorry it took so long to answer, I just got submit repo to work with it. It
passes there, I think that should be enough for now but we could do a mach5
testing if you prefer (and could you/someone do it please?)

I also opened https://bugs.openjdk.java.net/browse/JDK-8217267 to really
study the bias and accuracy of the sampling a bit more. So this might get
revisited in the near future but this particular webrev is "important" as
it fixes the tests to be doing what we want them to do.

Let me know what you think,
Jc

On Tue, Jan 22, 2019 at 1:14 PM <serguei.spitsyn at oracle.com> wrote:

>
>
> On 1/22/19 12:54 PM, serguei.spitsyn at oracle.com wrote:
>
> Hi Jc,
>
> It looks good to me too.
>
>
> Forgot to say that I hope it was tested well.
>
> Thanks,
> Serguei
>
>
> Thanks,
> Serguei
>
>
> On 1/11/19 5:07 PM, Hohensee, Paul wrote:
>
> Nits:
>
>
>
> memAllocator.cpp:
>
>
>
> Copyright line should read
>
>
>
> * Copyright (c) 2018, 2019, Oracle and/or its affiliates. All rights
> reserved.
>
>
>
> The comment
>
>
>
> // Tell tlab to forget the bytes_since if we passed it to the heap sampler.
>
>
>
> would be clearer if it read
>
>
>
> // Tell tlab to forget bytes_since_last if we passed it to the heap sampler.
>
> Same copyright line comment for HeapMonitor.java,
> HeapMonitorArrayAllSampledTest.java and HeapMonitorStatArrayCorrectnessTest.java
> as for memAllocator.cpp.
>
> Otherwise lgtm, no need for another webrev on my account.
>
>
> Thanks,
>
>
>
> Paul
>
>
>
> *From: *serviceability-dev <serviceability-dev-bounces at openjdk.java.net>
> <serviceability-dev-bounces at openjdk.java.net> on behalf of JC Beyler
> <jcbeyler at google.com> <jcbeyler at google.com>
> *Date: *Thursday, January 10, 2019 at 11:28 AM
> *To: *OpenJDK Serviceability <serviceability-dev at openjdk.java.net>
> <serviceability-dev at openjdk.java.net>
> *Subject: *RFR(S) 8215113: Sampling interval not always correct
>
>
>
> Hi all,
>
>
>
> Could I get a review for this:
>
> Webrev: http://cr.openjdk.java.net/~jcbeyler/8215113/webrev.00/
>
> Bug: https://bugs.openjdk.java.net/browse/JDK-8215113
>
>
>
> This fixes the code to do the right accounting in sampling and get the
> right sample counts and sampled objects. My error percentage calculation
> was wrong so the two tests I had added were false positives (my apologies).
>
>
>
> Thanks,
>
> Jc
>
>
>
>

-- 

Thanks,
Jc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/serviceability-dev/attachments/20190129/d9c76226/attachment-0001.html>