m68k/cpu32 TRAP instruction?
Jonathan Wakely
jwakely.gcc@gmail.com
Mon Oct 8 23:11:00 GMT 2018
On Mon, 8 Oct 2018 at 21:00, Josef Wolf wrote:
>
> On Sat, Oct 06, 2018 at 11:51:33AM +0800, Xi Ruoyao wrote:
> > > Simply crashing leaves you without any clue. You have no idea where to look
> > > for the bug. The crash could even be by a power outage. That's not exactly
> > > improving the situation.
> >
> > You have GDB.
>
> I'd happily do this. But unfortunately, this is an embedded system, and it
> doesn't have a debug interface.
>
> So gdb won't help very much here.
>
> > > Issuing a diagnostic "Oops, NULL-dereference at adsf.c:123" would improve
> > > the situation: you know where to look for the fix.
> >
> > You have -fsanitize=undefined.
>
> Ah! This seems to be exactly what I was looking for!
>
> Unfortunatly, the linker complains that it can't find -lubsan.
You can use -fsanitize-undefined-trap-on-error to avoid the need for
libubsan by inserting traps instead of calls to libubsan, but then
you're back where you started.
> Do I need to do something special to enable the build/install of the sanitizer
> runtime library on an embedded target? AFAICS, the sanitizer should be built
> by default?
I don't think libubsan is built for all targets.
More information about the Gcc-help
mailing list