On 2018 May 26 (Sat) at 21:53:51 +0200 (+0200), Markus Hennecke wrote:
:On Sat, 26 May 2018, Alexander Bluhm wrote:
:
:> On Sat, May 26, 2018 at 10:23:25PM +1000, Jonathan Gray wrote:
:> > > After seeing some mails regarding NEON support on armv7 another approach
:> > > would be to pass --cc-abi-flags="-mfloat-abi=softfp" to configure which
:> > > fixes the build on armv7 as well. Going that way the port has to be
:> > > updated after the switch to hard fp abi. Running with --disable-neon
:> > > results in a little slower regression test, 1105 seconds vs 1135 seconds.
:> >
:> > softfp ABI is the default since a few weeks ago
:> >
:> > https://marc.info/?l=openbsd-cvs&m=152554719006083&w=2
:>
:> I am involved in two threads discussing different solutions without
:> real progress. I have no idea what is needed on arm and I cannot
:> test it.
:>
:> Jonathan, could you just commit whatever is suitable?
:>
:> bluhm
:
:Just in case it helps, the armv7 patch is no longer needed as Jonathan
:already pointed out. I just retested on armv7. What is left is to fix the
:build on aarch64.
:
:Index: Makefile
:===================================================================
:RCS file: /cvs/ports/security/botan2/Makefile,v
:retrieving revision 1.16
:diff -u -p -r1.16 Makefile
:--- Makefile 12 Apr 2018 09:53:54 -0000 1.16
:+++ Makefile 26 May 2018 16:53:19 -0000
:@@ -45,6 +45,10 @@ CONFIGURE_ARGS = --with-build-dir=${WRKB
: --with-zlib \
: --without-sphinx
:
:+.if ${MACHINE_ARCH} == "aarch64"
:+CONFIGURE_ARGS += --disable-armv8crypto
:+.endif
:+
: .if ${MACHINE_ARCH} == "i386" || ${MACHINE_ARCH} == "amd64"
: PKG_ARGS = -Dx86=1
: .else
:
ok and committed
--
IBM had a PL/I,
Its syntax worse than JOSS;
And everywhere this language went,
It was a total loss.
