>
> Hi,
>
> > SIGEMT me thinks is a BSD thing, I have to provide it for SunOS
> > compatibility if that means anything... And whats even more strange
> > is that the NetBSD Sparc code sends this signal to a process if it
> > gets a tagged addition overflow trap ??!?!?! Possibly Ultrix or IRIX
> > or some other MIPS based Unix has this? As another point of reference
> > my Sparc gdb stub happens to use SIGEMT for fp disabled traps so it
> > seems as if SIGEMT is another one of the MAGIC Unix things ;)
>
> Hmpf... I think I'm gonna do some tests with Waldorf's RISC/os machines
> and take a look at the BSD 4.4 Lite sources.
>
No matter why, I believe that this is a GDB bug. It should take
care of include/signal.h, or needs a way to configure the signal
meanings.
But it's also the question why Linux assigns bus error to signal 7
while all others assign this to signal 10. Should be corrected, IMHO.
Cheers,
Andy
|