Faster than what I've expected, I've found what caused the segfault: Pcap.xs is allowing unsafe signals during the callback. I don't know why upstream did this, but this has to be discussed with them and filled as a bug in CPAN.
With the attached patch, my original testcase no longer segfaults. On Sat, Feb 14, 2009 at 21:32, Martín Ferrari <tin...@debian.org> wrote: > On Sat, Feb 14, 2009 at 21:08, Martín Ferrari <tin...@debian.org> wrote: >> Meanwhile this is solved, see the attached patch that's useful as a >> workaround for fwknopd > > My bad, that patch didn't work. It goes again, this time correctly tested :) > > > -- > Martín Ferrari > -- Martín Ferrari
Index: Pcap.xs =================================================================== --- Pcap.xs (revision 30710) +++ Pcap.xs (working copy) @@ -410,10 +410,7 @@ callback_fn = newSVsv(callback); user = newSVsv(user); - SAVE_signals = PL_signals; /* Allow the call to be interrupted by signals */ - PL_signals |= PERL_SIGNALS_UNSAFE_FLAG; RETVAL = pcap_loop(p, cnt, callback_wrapper, (u_char *)user); - PL_signals = SAVE_signals; SvREFCNT_dec(user); SvREFCNT_dec(callback_fn);
