[jdk17u-dev] RFR: 8285835: SIGSEGV in PhaseIdealLoop::build_loop_late_post_work

Martin Doerr mdoerr at openjdk.org
Wed Jan 4 10:42:53 UTC 2023


On Tue, 3 Jan 2023 11:24:59 GMT, Goetz Lindenmaier <goetz at openjdk.org> wrote:

> I backport this for parity with 17.0.7-oracle.
> 
> 
> I had to resolve one hunk in escape.cpp because head has more tracing code at that
> place. Old code calls set_scalar_replaceable(false), new code
> calls set_not_scalar_replaceable(...) which traces and calls the 
> other method.
> Instead of editing the new code, I keep the tracing call and add 
> that method in escape.hpp.  But I stripped the output of the 
> tracing, as it would be incomplete as long as I don't fix all
> occurances.
> The tracing was added by "8281548: Add escape analysis tracing flag"

LGTM. Please fix comment indentation before integrating.

src/hotspot/share/opto/escape.cpp line 1862:

> 1860:         PointsToNode* base = i.get();
> 1861:         // 2. An object is not scalar replaceable if the field into which it is
> 1862:        // stored has multiple bases one of which is null.

Odd indentation.

-------------

Marked as reviewed by mdoerr (Reviewer).

PR: https://git.openjdk.org/jdk17u-dev/pull/1023


More information about the jdk-updates-dev mailing list