SIGSEGV on PhaseIdealLoop::split_up?

[Uwe Schindler]

OK, will try tomorrow.

The big problem is that it triggers only from time to time. So maybe I can start some loop to run until it fails.

Uwe

-----
Uwe Schindler
uschindler at apache.org 
ASF Member, Apache Lucene PMC / Committer
Bremen, Germany
http://lucene.apache.org/

> -----Original Message-----
> From: Aleksey Shipilev <shade at redhat.com>
> Sent: Wednesday, January 30, 2019 5:10 PM
> To: Uwe Schindler <uschindler at apache.org>; 'Dawid Weiss'
> <dawid.weiss at gmail.com>; 'hotspot-dev' <hotspot-dev at openjdk.java.net>
> Subject: Re: SIGSEGV on PhaseIdealLoop::split_up?
> 
> On 1/30/19 3:51 PM, Uwe Schindler wrote:
> > as many of you know: I have no access to the OpenJDK bugtracker. I am
> currently on travel, but
> > later today, I'd take the current problem and I may open a new bug in the
> OpenJDK issue tracker
> > about this issue. The stack traces all look very similar, so there should be
> enough information
> > to figure out where exactly the SIGSEGV is happening. To my knowledge of
> development, this should
> > help to get the condition, why this is happening. Together with the Lucene
> Source code, you may
> > figure out the conditions. Because this bug is quite new, there must be
> some change in Lucene
> > that triggers this bug. With the information about where the SIGSEGV is
> happening in OpenJDK code
> > and the commit logs in Lucene, we may figure out how the Lucene code
> changed so it causes that
> > issue.
> >
> > I will take the time and try to collect all information an OpenJDK developer
> needs to (hopefully)
> > reproduce the issue.
> >
> > Does this sound like a plan? Uwe
> 
> Yes. And try to run with fastdebug build, it might meaningfully assert, which
> may point to existing
> bugreports or even fixes. You can build it yourself, or pull one from here:
> https://builds.shipilev.net/
> 
> -Aleksey
>