In article <200502271716.j1RHGkUq028320@ector.cs.purdue.edu>,
J Chapman Flack  <flack@cs.purdue.edu> wrote:
>Ok, the dodgy MP ProLiant has been replaced with an identical but working
>one, and I completely deinstalled mozilla-bin, sun-jdk15, sun-jre15,
>and suse_* (7.3), and reinstalled starting with the suse91 libraries
>as Thomas Bieg and Matthew Mondor suggested.  Still the same behavior, so
>that would seem to rule out both broken hardware and suse7.3 as the culprits.
>
>To replicate: (on ProLiant 800 with two PIII/600s, NetBSD 2.0)
>
>  cd /usr/pkgsrc/emulators/suse91_x11     &&  make bin-install
>  cd /usr/pkgsrc/emulators/suse91_compat  &&  make bin-install
>  cd /usr/pkgsrc/lang/sun-jre15  &&  make bin-install
>  cd /usr/pkgsrc/lang/sun-jdk15  &&  make bin-install
>
>  sun15-appletviewer /usr/pkg/java/sun-1.5/demo/applets/Animator/example1.html
>
>  Observe varying misbehavior including hangs (blocked or spinning),
>  jvm panic report, or occasional ( < 20% ) good runs.
>
>  Rerun after booting with only one processor enabled, observe flawless
>  operation.
>
>Should I PR this, or is there still something simple I might be missing?
>
>One interesting thing: when hotspot actually reports a panic (which only
>happens sometimes), the "SYSTEM" part of the report includes this line:
>
>  CPU:total 1 family 6, cmov, cx8, fxsr, mmx, sse
>
>I haven't seen the hotspot internals to know just what this line is
>intended to mean, but does it seem that hotspot thinks it's running on
>a uniprocessor?  that could be a problem ... I wonder if there's a suse or
>emulator call it's using to get that information, and it's getting the
>wrong story?
>
>-Chap

It is getting it from procfs, and I will commit a fix for it shortly.

christos