Re: broken stack trace (Was: carp kernel trap)

From: Yar Tikhiy <yar_at_comp.chem.msu.su> Date: Tue, 27 Jun 2006 17:34:46 +0400 · This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:57 UTC

On Tue, Jun 27, 2006 at 05:17:15PM +0400, Maxim Konovalov wrote:
> Btw, Yar, while working on this panic I noticed the trace is not
> correct.  gdb shows sin=0x0 but oviously it is not.  Any ideas why?
> 
> > #12 0xc054733a in carp_set_addr (sc=0xc115bc00, sin=0x0) at ../../../netinet/ip_carp.c:1439
> >         ifp = (struct ifnet *) 0xc1291000
> >         cif = (struct carp_if *) 0x6b5
> >         ia = (struct in_ifaddr *) 0xc12a4500
> >         ia_if = (struct in_ifaddr *) 0xc1291000
> >         imo = (struct ip_moptions *) 0xc115bc0c
> >         addr = {s_addr = 301990112}
> >         iaddr = 3239427084
> >         own = 0
> >         error = -1066976812

It must be fallout from optimization by gcc.  Alas, absolutely
reliable debugging of function arguments and automatic variables
is possible only after -O0 was specified explicitly to gcc, according
to my experience with gcc and gdb.

-- 
Yar