IBM jdk needs kernel patch to fix floating stacks problem? LD_ASSUME_KERNEL

[nigel kersten]

The problem appears to be this:

(it's not just a YDL specfic problem):

Known Issues / Trouble Spots:
-----------------------------

    o Some Java JVMs (both from Sun and IBM) don't work with the new
      floating stack feature of the included version of glibc.  The failures
      are due to programming assumptions in the JVMs that are now invalid.
      JVM vendors are working on making the necessary corrections.  Until
      corrected JVM packages are available, you may force glibc to use the
      deprecated stack model by setting the following environment variable:

                          LD_ASSUME_KERNEL=2.2.5


but from what I can work out, this doesn't work, so I guess the patch mustn't 
have been applied to YDL 2.2, and 2.1 must be using an older version of 
glibc?

Do Blackdown normally not have a JIT included with their java distribution? 
Seems kind of a strange thing to do....


nigel



On Fri, 17 May 2002 07:09, Christian Gross wrote:
> I gave up and went back to YDL 2.1.  Everything worked on that platform.
>
> I am waiting until the Blackdown sources have upgraded to 1.4.
>
> Christian Gross
>
> At 10:07 16/05/2002 +1000, you wrote:
> >Hi ppc'ers
> >        From discussion on the ibm.software.java.linux newsgroup
> > ---snipped-- below, it's clear
> >we ppc users need some kernel patch to get the IBM jdk's (1.3.0, 1.3.1) to
> >work.
> >Anybody got any  ideas if this is possible? Where to get the patch? Is
> >there a patch?
> >
> >Any views welcome,
> >                                      James.
> >
> >-----snipped from ibm.software.java.linux -----
> >
> >>Hi all,
> >>
> >>When trying to use the pSeries JDK with Tomcat v4.0.3, it bombs out on
> >>startup like so:
> >>
> >>**Out of memory, aborting**
> >>
> >>*** panic: JVMST058: Cannot allocate memory in initGcHelpers(3)
> >>
> >>It then leaves a 60MB or so core file, and exits.
> >>
> >>I have tried all sorts of combinations of -X options to do with memory,
> >>none of them make any difference.
> >>
> >>The target platform is LinuxPPC v2.4.19-pre7 (v2.4.18 has same problem)
> >>under YDL v2.2, with glibc 2.2.5-0.19a.
> >>
> >>Anyone shed any clue? Is anyone from IBM interested in the core dump so
> >>it can be debugged?
> >>
> >>Regards,
> >>Graham
> >
> >As I have just posted on another thread:
> >
> >
> >It looks very much like the floating stack thing.  If YellowDog don't
> >give you the LD_ASSUME_KERNEL patch, then you can compile
> >glibc with floating stacks disabled.  This course of action is only open
> >to the brave and the foolish.
> >
> >Neil
> >
> >-----snipped from ibm.software.java.linux -----
> >
> >
> >_______________________________________________
> >yellowdog-general mailing list
> >yellowdog-general@lists.terrasoftsolutions.com
> >http://lists.terrasoftsolutions.com/mailman/listinfo/yellowdog-general
>
> _______________________________________________
> yellowdog-general mailing list
> yellowdog-general@lists.terrasoftsolutions.com
> http://lists.terrasoftsolutions.com/mailman/listinfo/yellowdog-general