On Nov 27, 2012, at 11:21 AM, Robert Williams wrote:
> Any advice or pointers appreciated!
So, when you see these alignment errors or tracebacks, they are always a
software defect. Typically this is something doing bogus pointer math, but the
event was non-fatal. (As compared to an ALIGN-1-FATAL message).
Take this entry:
Address Count Traceback
8 226 0x40D42728 0x40D425B4 0x40D42618 0x418D8C74
0x40306030 0x40306970 0x4030A2FC 0x4313B13C
Someone (with that stack trace) was trying to look at the memory at 0x8, and
did 226 times for that stack trace. This is likely code that does something
like this:
struct somestruct *foo = NULL;
printf("%s\n", foo->bar);
but since foo is null, and bar is located 8 bytes in a normal struct of type
'somestruct' you get that address.
With a better 'show version' Cisco should be able to identify if it is a known
defect or a new one.
You should open a case and IMHO cisco should at least triage it even if you
don't have support so the defect can be fixed for other customers.
The output decoder/interpreter may also be able to diagnose this as it will
decode the stack trace and match against known public bugs.
Hope this helps.
- Jared
_______________________________________________
cisco-nsp mailing list cisco-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
