Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: automatic stacktrace on panic, was: Re: Boot fail as kvm guest



On Sat, Feb 18, 2012 at 09:06:31PM -0500, Thor Lancelot Simon wrote:
> On Sat, Feb 18, 2012 at 04:05:22PM -0600, David Young wrote:
> > On Sat, Feb 18, 2012 at 08:05:23PM +0100, Manuel Bouyer wrote:
> > > On Sat, Feb 18, 2012 at 12:09:02PM -0600, David Young wrote:
> > > > > that's not enough: you still miss the panic message.
> > > > 
> > > > That's a good point, however, can't we print the panic message again, or
> > > > else postpone printing it?
> > > 
> > > if things go wrong when printing the stack trace, you also loose
> > > the panic message. that's the worse possible senario.
> > 
> > Why should anything go wrong when printing the stack trace?  I realize
> > that, in practice, things *do* go wrong, but aren't these defects in our
> > stack trace code?
> 
> No -- what if the stack was damaged so you fault while trying to unwind it?

I figure that now we will probably get a ddb loop/recursion or some
other ugliness.  I *think* that it is feasible for the stack trace
code to dereference the stack and pointers in the stack, after
range-checking, with a suitable fault handler in place, but there are
probably hazards I have not thought of.

Dave

-- 
David Young
dyoung%pobox.com@localhost    Urbana, IL    (217) 721-9981


Home | Main Index | Thread Index | Old Index