Bug#869122: Bug#869373: Bug#869122: perl: 5.26 FTBFS on hppa: broken miniperl?

[Niko Tyni]

On Mon, Jul 24, 2017 at 02:01:16PM -0400, John David Anglin wrote:
> On 2017-07-24, at 6:45 AM, John Paul Adrian Glaubitz wrote:
> 
> > On Sun, Jul 23, 2017 at 09:32:04PM -0400, John David Anglin wrote:
> >> Maybe the patch will also fix the sh build.
> > 
> > Yes, it does. I just tested the patch and I can confirm with the patch
> > applied, I can build src:perl just fine on sh4 without "noopt".
> 
> I created a gcc bug report:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81538

Thanks for your great work on this!

I feel somewhat uneasy about applying a workaround patch that affects
all the architectures for the sake of non-release architectures.

I can see that it's just a code reorganization that presumably avoids
the buggy gcc optimization (at least for the time being), but it's a bit
hard to feed upstream as-is and I wouldn't like to carry it for the major
architectures without upstreaming it.

Do you know why the GCC optimization issue only seems to affect hppa+sh4?

I wonder if lowering the optimization of the affected
code would be an easier workaround. On that note, looking
at the build logs makes me think the cflags fiddling in
debian/patches/debian/hppa_op_optimize_workaround.diff is still working,
but the problem here is that op.c is linked to opmini.c for the miniperl
build and that still gets compiled at -O2.

Does it work for you without the Perl_custom_op_get_field() code
reorganization if you patch cflags.SH further like this?

-    op) : work around http://bugs.debian.org/838613
+    op|opmini) : work around http://bugs.debian.org/838613

This should make it build both op.c and opmini.c at -O0.

Or does this hit the hppa "-O0 -fPIC" problem again?

(For sh4, the archname check on the next lines obviously needs adjustment
too.)

Thanks again, especially for filing the gcc upstream bug so we can
eventually get rid of these workarounds.
-- 
Niko