Re: Which POWER implementations need the HWCAP_DFP check?

[Florian Weimer <fweimer at redhat dot com>]

On 05/14/2018 10:29 PM, Tulio Magno Quites Machado Filho wrote:
> Florian Weimer <fweimer@redhat.com> writes:
>
> > On 05/14/2018 09:09 PM, Tulio Magno Quites Machado Filho wrote:
> > > > I would have suggested something like
> > > >
> > > > #define DL_HWCAP_HAS_DFP (GLRO(dl_hwcap) & PPC_FEATURE_HAS_DFP)
> > > >
> > > > in the general powerpc sysdeps file, and
> > > >
> > > > #define DL_HWCAP_HAS_DFP 1
> > > >
> > > > for power6 (which I expect will propagate upwards).  I expect that this
> > > > will result in somewhat nicer machine code, too.
> >
> > > It's safe to use the newer mtfsf in both cases.  The extra fields from the
> > > new mtfsf are ignored by older processors.
> >
> > But couldn't dropping the conditional have impact on other processors?
> > Then we should do that only for POWER6 and later.
>
> Agreed.

Good.  We may have to address this issue if we want to implement support 
for --enable-math-noprivate on ppc64le, along with some spurious IFUNC 
resolvers which always resolve to the same implementation on POWER8 and 
later.

<https://sourceware.org/ml/libc-alpha/2018-05/msg00501.html>
<https://sourceware.org/ml/libc-alpha/2018-05/msg00540.html>
<https://sourceware.org/ml/libc-alpha/2018-05/msg00541.html>

Thanks,
Florian