[11u] RFR 8214975: No hs-err file if fatal error is raised during dynamic initialization

Thomas Stüfe thomas.stuefe at gmail.com
Tue Sep 10 14:07:05 UTC 2019


Hi,

I would like to bring this patch to 11u.

Original Bug: https://bugs.openjdk.java.net/browse/JDK-8214975
Original Change: http://hg.openjdk.java.net/jdk/jdk/rev/0de1f006d3c3

The patch fixes an rare but annoying issue which causes the VM to behave
erratically if an assert or crash happens during dynamic C++ initialization.

The patch is rather simple, exchanging the global-scope-C++ object wrapping
the error log file handle - which needs dynamic C++ initialization - with
simple int, which does not and does work always. The patch comes with a new
test case.

Unfortunately, the patch did not apply cleanly, since 11u is missing
"8209856: Obsolete error reporter". But I found that one too big to bring
to 11u. The adjustments are trivial anyway:

http://cr.openjdk.java.net/~stuefe/webrevs/backports/8214975-No-hs-err-file-if-fatal-error-is-raised-during-dynamic-initialization-reject-fixes.patch

Thank you,

Thomas


More information about the hotspot-runtime-dev mailing list