Libthr stable enough for testing

[Glenn Johnson]

On Thu, May 29, 2003 at 05:31:32PM -0700, Julian Elischer wrote:


> On Thu, 29 May 2003, Glenn Johnson wrote:
>
> > On Thu, May 29, 2003 at 03:54:22PM -0700, Julian Elischer wrote:
> >
> >
> > > On Thu, 29 May 2003, Glenn Johnson wrote:
> > >
> > > > On Thu, May 29, 2003 at 02:20:02AM -0400, Mike Makonnen wrote:
> > > >
> > > > > On Thu, 29 May 2003 00:35:32 -0500 Dan Nelson
> > > > > <dnelson at allantgroup.com> wrote:
> > > > >
> > > > >
> > > > > > The real problem is in the kernel, though.  A userland
> > > > > > non-root process should not be able to hard lock the system.
> > > > > > One of the threads people will probably have to get an SMP
> > > > > > machine to be able to debug it.
> > > > >
> > > > >
> > > > > Upon first reading it I had assumed he meant gnome locked
> > > > > up. Can you confirm this Glenn? Is the machine itself locked
> > > > > solid (no ping, no ssh, not vtys, etc)?  However, even if that
> > > > > is the case a bug in the kernel does not preclude a bug in
> > > > > libthr.
> > > >
> > > > The machine locks up solid.  I can not do anything with it      
> > > > except hit the reset button.  I mentioned gnome because gnome   
> > > > will trigger the lock up.  I would imagine I would see the same 
> > > > thing with kde however.                                         
> > >
> > > what about kernel debugger?
> >
> > I do not have the debugger enabled on the SMP box.  I can set that
> > up but probably not until next week.  Will the debugger work if the
> > machine locks up?
>
> that depends.. when it locked up, did it still respond to pings?

No, the networking was dead as well.

-- 
Glenn Johnson
glennpj at charter.net