On Sun, May 3, 2009 at 2:25 PM, John David Anglin <dave@xxxxxxxxxxxxxxxxxx> wrote: > I'm not able to build gcc with 2.6.30-rc4. Builds don't usually get > beyond stage1. A variety of applications generate SEGVs, the most common > one being /bin/sh (bash). I looked at one that seemed reasonably easy > to understand. Quite a few of the SEGVs occur running _dl_fixup. That is odd, but certainly a place where I would expect memory corruption to bite you, since if you resolve the wrong address you will likely crash. > So, the values in $r20 and $ret0 are not consistent with the values in > memory pointed to by $r22. I looked at the assembler code for do_lookup_x > and as far as I can tell, there is no way to avoid the path through > do_lookup_x+364. Further, $rp is consistent with do_lookup_x being > called at its normal entry position. So, either memory or $r22 is > corrupt in the core dump. Could you run the entire build with LD_BIND_NOW=1 set? Does it crash anywhere else? > The values like 0x000ca... probably are from bash. Yes, the application would be at a low address. Cheers, Carlos. -- To unsubscribe from this list: send the line "unsubscribe linux-parisc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
