Re: [OpenBIOS] [RFC 3/3] SPARC32/64: Mimic Sun's OBP behaviour if a divide by zero occurs.

[Blue Swirl]

On Sat, Jan 12, 2013 at 2:18 PM, Segher Boessenkool
 wrote:
> > Actually this got me thinking, since you're right in that integer division
> > should require a manual trap. This took me into OpenBIOS's bundled version
> > of libgcc and __udivmoddi4.c:
> >
> > __uint128_t __udivmodti4(__uint128_t num, __uint128_t den, __uint128_t
> > *rem_p)
> > {
> >   __uint128_t quot = 0, qbit = 1;
> >
> > if ( den == 0 ) {
> >     __divide_error();
> >     return 0;                   /* If trap returns... */
> >   }
>
> [It seems you showed __udivmodti4 instead -- but the code is much the same].
>
> That's not libgcc code: it is x86 Linux kernel code or klibc or something
> like that.  Libgcc forces a single-precision divide by zero when the double-
> precision routine divides by zero, so there can be no problem there.

Some software archeology follows.

The function  __udivmoddi4() was added by me in r4 as
arch/sparc32/libgcc/__udivmoddi4.c in 2006. Then I moved the file with
r61 to its current location in ./libgcc, with R62 copied this to
__udivmodti4.c and changed 64 bit types to 128 bits.

This and files __divdi3.c, __udivdi3.c and __umoddi3.c seem to
originate from klibc indeed. With
2f6cd4633e623c29bbcb793eff76c8c08c2c0f9c H. Peter Anvin added the
files to klibc/arch/i386/libgcc/ in 2002, they have been moved since
to usr/klibc/libgcc/ and reformatted. Before reformatting to kernel
style, the files were identical to our version except for a header
change.

Some klibc files like these are also used by Syslinux, iPXE and gPXE,
but they may actually be newer than OpenBIOS r4. So I think I used
klibc directly.

As specified by usr/klibc/LICENSE, klibc is a mix of GPLv2only, BSD
and MIT licensed files, the default being MIT unless files originate
from BSD or Linux. I think these files are therefore MIT licensed.
GPLv2only and BSD would be OK for OpenBIOS too.

Files ashldi3.c and ashrdi3.c are from GCC 2.95.2 libgcc which was
GPLv2+ at that time (way before GPLv3 switch). GPLv2+ is also fine for
OpenBIOS.

The license situation could be clarified by copying the LICENSE from
klibc to OpenBIOS.
> > A quick grep of the source shows that __divide_error() is manually set to
> > the unexpected exception trap handler in start.S:
> >
> > _GLOBAL(__divide_error):
> > trap_error:
> >         mflr    r3
> >         LOAD_REG_FUNC(r4, unexpected_excep)
> >         mtctr r4
> >         bctr
> >
> > Heh. So actually all we need to do is provide an empty function for
> > __divide_error() and we magically get the behaviour we want. Revised patch
> > to follow shortly.
>
> Or not have a __divide_error at all, which makes it clearer the intent
> is to have the double precision divide have the same behaviour as single
> precision on whatever hardware it runs on.  Your patch should work though
> :-)
>
> Segher