In some email I received from Greg 'groggy' Lehey, sie wrote:

Checking application/pgp-signature: FAILURE
-- Start of PGP signed section.
> On Sunday,  4 May 2003 at  1:04:07 +1000, Darren Reed wrote:
> >
> > There would appear to be a number of panic messages in the kernel that
> > are not unique, giving panic/crash diagnosis an extra ounce of difficulty.
> > Whilst I'm well aware of other mechanisms such as that used by Solaris
> > present a unique identifier for each message, I'd like to recommend an
> > interim step of doing something simple to make each message different.
> > This might be as simple as numbering them or as complex as changing them
> > to better reflect the condition.
> >
> > What do others feel about this sort of change ?
> 
> Wouldn't it make more sense to print a stack back trace with function
> names?  That would give much more information.

No.  That shows the path that led to the problem, but it doesn't pinpoint
where the problem occurs.

In most cases, the similar panic messages occur in the same function,
and stack trace is of little help.  The problem that stirred me up to
write this email was a panic in m_copydata.  There are three different
panic's within that single function, all have the same panic message
and of which the stack trace might help me decide between 1 or the
other 2.  This is less than ideal.  I'm sure there are other, similar,
problems elsewhere.

Darren