[8u] RFR: 8217362: Emergency dump does not work when disk=false is set

[Hohensee, Paul]

Lgtm.

Paul

On 10/19/20, 7:08 AM, "jdk8u-dev on behalf of Ekaterina Vergizova" <jdk8u-dev-retn at openjdk.java.net on behalf of katya at azul.com> wrote:

    Hello,
    I would like to backport JFR issue 8217362 to 8u.

    JBS: https://bugs.openjdk.java.net/browse/JDK-8217362
    Original patch: https://hg.openjdk.java.net/jdk/jdk/rev/3cabb47758c9
    Webrev for 8u
      hotspot: https://cr.openjdk.java.net/~apetushkov/jfr_backports_katya/8217362/hotspot/webrev.00/
      jdk: https://cr.openjdk.java.net/~apetushkov/jfr_backports_katya/8217362/jdk/webrev.00/

    The patch doesn't apply cleanly due to the context differences:
    - different logging (tty->print_cr in 8u instead of log_info/log_error) and includes list in jfrRepository.cpp and jfrEmergencyDump.cpp
    - different parameters of ProcessTools.createJavaProcessBuilder() in TestDumpOnCrash.java

    I reapplied the changes manually:
    - moved emergency dump methods from jfrRepository.cpp to jfrEmergencyDump.cpp
    - incorporated the fix for emergency dump when disk=false and other small changes in emergency dump methods
    - made sure the resulting patch differs from the original patch only by the context differences described above

    Tested with tier1 and jdk.jfr tests on Linux and Windows.

    Thanks,
    Ekaterina