You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@harmony.apache.org by Ivan Popov <iv...@gmail.com> on 2007/03/21 15:14:57 UTC
[drlvm][threading] DRLVM intermittently hangs on exit while detaching thread
Hi all,
I noticed several JPDA tests intermittently timed out on Linux and
there are problems in using Eclipse Java debugger. I submitted this in
HARMONY-3314.
The problem is that debuggee DRLVM process intermittently hangs on
exit while detaching thread. I've added stack trace to JIRA and I'd
appreciate if someone from DRLVM threads gurus can look at this
problem and help to identify the root cause.
I noticed tests failures only on a multi-threaded Linux box, but
Eclipse debugger also fails sometime on Windows (JVM process is not
terminated after debug session ended). I suspect this is a timing
issue.
Thanks.
Ivan
Re: [drlvm][threading] DRLVM intermittently hangs on exit while detaching thread
Posted by Weldon Washburn <we...@gmail.com>.
On 3/22/07, Ivan Popov <iv...@gmail.com> wrote:
>
> Weldon,
>
> Thanks for looking at this problem. Though this problem is
> intermittent I think it is major (but not critical) and worth to be
> fixed, because it often breaks Eclipse debugger work with Harmony. I'm
> not sure if this deadlock occur only in debug mode, it might be
> reproduced in usual mode and thus negatively affect overall DRLVM
> stability.
>
> The other problem is that CC run for jdktools often fail on Linux,
> because JPDA tests intermittently timed out and this makes monitoring
> regressions difficult. Putting all JPDA tests into exclude lists seems
> not a good idea.
Good point.
Incidentally I have been looking at threading code and see places where
there might be race conditions in thread join code. And the thread join
code might just be involved in the thread exit problems you are seeing. I
will keep this bug in mind as I work with the code.
Thanks.
> Ivan
>
> On 3/22/07, Weldon Washburn <we...@gmail.com> wrote:
> > Ivan,
> > I can't promise that I can get to this soon. I have zero idea if anyone
> > else will look at it. However, I am looking at other JIRA patches that
> > might also fix this problem as a side effect. I noticed it is currently
> > tagged as major priority. Is this still the case?
> >
> >
> > On 3/21/07, Ivan Popov <iv...@gmail.com> wrote:
> > >
> > > Hi all,
> > >
> > > I noticed several JPDA tests intermittently timed out on Linux and
> > > there are problems in using Eclipse Java debugger. I submitted this in
> > > HARMONY-3314.
> > >
> > > The problem is that debuggee DRLVM process intermittently hangs on
> > > exit while detaching thread. I've added stack trace to JIRA and I'd
> > > appreciate if someone from DRLVM threads gurus can look at this
> > > problem and help to identify the root cause.
> > >
> > > I noticed tests failures only on a multi-threaded Linux box, but
> > > Eclipse debugger also fails sometime on Windows (JVM process is not
> > > terminated after debug session ended). I suspect this is a timing
> > > issue.
> > >
> > > Thanks.
> > > Ivan
> > >
> >
> >
> >
> > --
> > Weldon Washburn
> > Intel Enterprise Solutions Software Division
> >
>
--
Weldon Washburn
Intel Enterprise Solutions Software Division
Re: [drlvm][threading] DRLVM intermittently hangs on exit while detaching thread
Posted by Ivan Popov <iv...@gmail.com>.
Weldon,
Thanks for looking at this problem. Though this problem is
intermittent I think it is major (but not critical) and worth to be
fixed, because it often breaks Eclipse debugger work with Harmony. I'm
not sure if this deadlock occur only in debug mode, it might be
reproduced in usual mode and thus negatively affect overall DRLVM
stability.
The other problem is that CC run for jdktools often fail on Linux,
because JPDA tests intermittently timed out and this makes monitoring
regressions difficult. Putting all JPDA tests into exclude lists seems
not a good idea.
Thanks.
Ivan
On 3/22/07, Weldon Washburn <we...@gmail.com> wrote:
> Ivan,
> I can't promise that I can get to this soon. I have zero idea if anyone
> else will look at it. However, I am looking at other JIRA patches that
> might also fix this problem as a side effect. I noticed it is currently
> tagged as major priority. Is this still the case?
>
>
> On 3/21/07, Ivan Popov <iv...@gmail.com> wrote:
> >
> > Hi all,
> >
> > I noticed several JPDA tests intermittently timed out on Linux and
> > there are problems in using Eclipse Java debugger. I submitted this in
> > HARMONY-3314.
> >
> > The problem is that debuggee DRLVM process intermittently hangs on
> > exit while detaching thread. I've added stack trace to JIRA and I'd
> > appreciate if someone from DRLVM threads gurus can look at this
> > problem and help to identify the root cause.
> >
> > I noticed tests failures only on a multi-threaded Linux box, but
> > Eclipse debugger also fails sometime on Windows (JVM process is not
> > terminated after debug session ended). I suspect this is a timing
> > issue.
> >
> > Thanks.
> > Ivan
> >
>
>
>
> --
> Weldon Washburn
> Intel Enterprise Solutions Software Division
>
Re: [drlvm][threading] DRLVM intermittently hangs on exit while detaching thread
Posted by Weldon Washburn <we...@gmail.com>.
Ivan,
I can't promise that I can get to this soon. I have zero idea if anyone
else will look at it. However, I am looking at other JIRA patches that
might also fix this problem as a side effect. I noticed it is currently
tagged as major priority. Is this still the case?
On 3/21/07, Ivan Popov <iv...@gmail.com> wrote:
>
> Hi all,
>
> I noticed several JPDA tests intermittently timed out on Linux and
> there are problems in using Eclipse Java debugger. I submitted this in
> HARMONY-3314.
>
> The problem is that debuggee DRLVM process intermittently hangs on
> exit while detaching thread. I've added stack trace to JIRA and I'd
> appreciate if someone from DRLVM threads gurus can look at this
> problem and help to identify the root cause.
>
> I noticed tests failures only on a multi-threaded Linux box, but
> Eclipse debugger also fails sometime on Windows (JVM process is not
> terminated after debug session ended). I suspect this is a timing
> issue.
>
> Thanks.
> Ivan
>
--
Weldon Washburn
Intel Enterprise Solutions Software Division