Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <4627b565-0a5c-080e-726a-01b773c985e8@landley.net>
 <20200603170357.GG1079@brightrain.aerifal.cx> <CAK8P3a2QYkZ9StDssxeZ3_zR5OH_P3v-D1iZp7=1E+2EaqoNcQ@mail.gmail.com>
In-Reply-To: <CAK8P3a2QYkZ9StDssxeZ3_zR5OH_P3v-D1iZp7=1E+2EaqoNcQ@mail.gmail.com>
From:   Masahiro Yamada <masahiroy@kernel.org>
Date:   Thu, 4 Jun 2020 10:07:29 +0900
Message-ID: <CAK7LNATDg=5U=GjwN9+gfW+kxJ3gN25wGrr7odiqsf7HBsb73w@mail.gmail.com>
Subject: Re: headers_install builds break on a lot of targets?
To:     Arnd Bergmann <arnd@arndb.de>
Cc:     Rich Felker <dalias@libc.org>, Rob Landley <rob@landley.net>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        David Woodhouse <dwmw2@infradead.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Jun 4, 2020 at 2:59 AM Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Wed, Jun 3, 2020 at 7:04 PM Rich Felker <dalias@libc.org> wrote:
> > On Wed, Jun 03, 2020 at 08:49:54AM -0500, Rob Landley wrote:
> > >     make ARCH=3D$i distclean defconfig headers_install \
> > >
> > > On the bright side, the resulting fruitbasket.tar.xz is 1.5 megabytes=
. The
> > > downside is I have no idea how broken the resulting header files are =
after this
> > > error-fest:
> > >
> > > alpha
> > > arc
> > > gcc: error: unrecognized command line option =E2=80=98-mmedium-calls=
=E2=80=99
> > > gcc: error: unrecognized command line option =E2=80=98-mno-sdata=E2=
=80=99; did you mean
> > > =E2=80=98-fno-stats=E2=80=99?
> > > gcc: error: unrecognized command line option =E2=80=98-mmedium-calls=
=E2=80=99
> > > gcc: error: unrecognized command line option =E2=80=98-mno-sdata=E2=
=80=99; did you mean
> > > =E2=80=98-fno-stats=E2=80=99?
> > > [...]
> >
> > Uhg. Surely there should be some fix for whatever mistaken dep is
> > behind this? Headers shouldn't actually depend on any config/compiler
> > output, should they??
>
> The first one of the two comes from "make defconfig", which definitely ne=
eds
> a working $TARGET compiler, but isn't actually needed before
> "make headers_install" as I just checked.
>
> > Or is that machinery somehow involved in
> > generating the syscall lists and similar?
>
> The syscall list for ARC is still not generated (that's on my todo list),=
 but
> something does call it even for "make headers_install".
>
> What it does is to set Makefile variables such as
>
> CC_VERSION_TEXT =3D $(shell $(CC) --version 2>/dev/null | head -n 1)
> and trying out all kinds of gcc options that may or may not be supported
> such as  "$(call cc-option,-fno-tree-loop-im)".
>
> Setting CC=3D: avoids this, like
>
> make -s CC=3D: ARCH=3D$i headers_install  INSTALL_HDR_PATH=3D"$PWD/fruitb=
asket/$i"
>
>       Arnd


You can do headers_install without the .config file.
'make defconfig' is unneeded.

headers_install should be independent of $(CC) too.


linux-next has some progress about this.

You can get clean headers_install without CROSS_COMPILE
for parisc, h8300, hexagon.


You still get warnings for ia64 and arc.
Our goal is to not require $(CC) for non-dot-config target
like 'make headers_install', 'make help', etc.


--
Best Regards
Masahiro Yamada