Date: Tue, 31 Jan 2017 19:09:01 -0800 From: Mark Millard <markmi@dsl-only.net> To: Justin Hibbits <chmeeedalf@gmail.com> Cc: FreeBSD PowerPC ML <freebsd-ppc@freebsd.org> Subject: Re: svn commit: r313005 - in head/sys: conf modules/zfs [The related 215947: -mlongcall use instead of -mminimal-toc] Message-ID: <E248B6D5-3B85-49E2-BCCE-F8A808D40F5B@dsl-only.net> In-Reply-To: <2B31BF44-CB4F-4B07-AFFE-0D2083855BB2@gmail.com> References: <676ECF89-F9AE-436A-942A-003B76C6BFE2@dsl-only.net> <2B31BF44-CB4F-4B07-AFFE-0D2083855BB2@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2017-Jan-31, at 6:44 PM, Justin Hibbits <chmeeedalf at gmail.com> =
wrote:
> On Jan 31, 2017, at 4:26 AM, Mark Millard wrote:
>>> Modified: head/sys/modules/zfs/Makefile
>>> =3D
>>> =3D
>>> =3D
>>> =3D
>>> =3D
>>> =3D
>>> =3D
>>> =3D
>>> =3D
>>> =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>> --- head/sys/modules/zfs/Makefile Tue Jan 31 01:48:55 2017 =
(r313004)
>>> +++ head/sys/modules/zfs/Makefile Tue Jan 31 01:55:29 2017 =
(r313005)
>>> @@ -94,7 +94,7 @@ CFLAGS+=3D-I${SUNW}/common
>>> CFLAGS+=3D-DBUILDING_ZFS
>>>=20
>>> .if ${MACHINE_ARCH} =3D=3D "powerpc64"
>>> -CFLAGS+=3D-mminimal-toc
>>> +CFLAGS.gcc+=3D-mminimal-toc
>>> .endif
>>>=20
>>> .ifdef ZFS_DEBUG
>>=20
>> You may want to process my bugzilla 215107 with =
head/sys/modules/zfs/Makefile
>> as it also deals with that -mminimal-toc: in other words to declare =
215107
>> fixed by this update.
>>=20
>> =3D=3D=3D
>> Mark Millard
>> markmi at dsl-only.net
>>=20
>=20
> Ah, thanks for reminding me, I'll take that.
>=20
> - Justin
Thanks.
215947 is sort of related in that it eliminated -mlongcall
usage because (A) clang 3.8.0 did not have it and (B)
the usage was not needed under the medium model. It also
limited the forced use of gcc in:
/usr/src/lib/csu/powerpc64/Makefile
But I've had the patch in use for a long time and -mlongcall
now exists if one requires a recent enough clang. (I only
recently made a bugzilla for the issue so bugzilla does
not reflect the patch's age.)
Even if you leave in the -mlongcall usage something
like part of the patch should be applied to avoid forcing
gcc use.
And the comment about why -mlongcall is needed is now wrong
even if the -mlongcall 's are left in place.
=3D=3D=3D
Mark Millard
markmi at dsl-only.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E248B6D5-3B85-49E2-BCCE-F8A808D40F5B>