On Sat, 18 Dec 2004, Florian Weimer wrote: > * James Yonan: > > >> >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> > >> Program terminated with signal 10, Bus error. > >> #0 0x120023904 in mroute_extract_addr_from_packet > >> (src=0x1ffffb220, dest=0x1ffffb208, buf=0x1, tunnel_type=-20292) at > >> mroute.c:113 > >> 113 memcpy (src->addr, &ip->saddr, 4); > >> >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> > >> > >> I think my system choke on the 'ip->saddr' address in the memcpy statement. > > > > That's strange that it would hit an alignment fault on a memcpy. memcpy > > is supposed to operate at byte granularity. > > I suspect this statement is the problem: > > const struct openvpn_iphdr *ip = (const struct openvpn_iphdr *) BPTR (buf); > > The cast results in undefined behavior if BPTR (buf) has the wrong > alignment. Recent versions of GCC know this and therefore conclude > that BPTR must be properly aligned. As a result, it is possible to > replace memcpy() with something that copies word-wise. > > The char * casts in your other message won't fix this, they come too > late. It's a common misconception that char pointers behave like > machine addresses. They don't. > > Just to be clear, this is not a bug in GCC. Your code is not valid C.
That's a great observation that probably explains the alpha alignment issues -- thanks! James
