[c-nsp] 2811 crash...

[Tim Franklin]

On Thu, January 24, 2008 1:29 pm, Brad Beck wrote:

> I've got a 2811 that crashed with a bus error within a few hours of
> enabling Netflow export.  I haven't opened a TAC case yet, but my
> curiousity was aroused by the address at which the error occured....
> "0xDEADBEF3".  This could almost be pronounced "Dead Beef"....

Yes - traditionally it's 0xDEADBEEF, I guess Cisco felt the need to
distinguish between different sides of beef and so dropped an 'E' to add
an index number :)

Seriously, it's an old coder utility / humour combo - if there's memory
that's you shouldn't be using, rather than leaving it full of random
values, you fill it with DEADBEEF.  If the program crashes or gives wrong
answers, and you see DEADBEEF show up, you know it's a problem with
looking at the wrong area of memory rather than some other kind of logic
error.

Why DEADBEEF and not some other 'magic' number is up there with why you
find so many workstations of long-time coders with the MAC of their
Ethernet card re-programmed to be nn:nn:nn:C0:FF:FE...

Regards,
Tim.