Wojtek Mitus <[email protected]> writes:
> Hello,
> I have a problem with ltrace (pulled from git and built today). Every
> binary which i try to ltrace ends up with SIGSEGV in the same place -
> SYS_ioctl(1,,,) (example attached below). I'll be grateful for any
> hints on what can cause such behavior...
> Thanks,
> WM
It doesn't seem like it should make a difference, but would you please
recheck with the released tarball 0.7.2? That was reported to work on
MIPS.
Then you should re-try with a minimal binary, something like
int main() { return puts (""); }
Presumably it will fail as well. Re-run with -D77, and send us the
resulting log, that might give us a clue as to whether breakpoints are
where expected.
> [0x2aaa9c9c] SYS_mmap(0, , , )
> = 0x2aaae000
> [0x2aaaa9c8] SYS_open(0x7f8ed1b8, , , )
> = 3
BTW, all those missing arguments mean that argument-fetching is broken.
It seems like only the first argument is ever decoded properly. I don't
really know how that is supposed to look on MIPS, perhaps it's expected.
Thank you,
PM
_______________________________________________
Ltrace-devel mailing list
[email protected]
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/ltrace-devel
