Received-SPF: pass (google.com: domain of linux-kernel+bounces-78294-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240223143309.5ec298ec@canb.auug.org.au> <CACMJSevZxCiqa8uz+XU36psCa5T_rQGi5tVkMejpZj22Bj1k3Q@mail.gmail.com>
In-Reply-To: <CACMJSevZxCiqa8uz+XU36psCa5T_rQGi5tVkMejpZj22Bj1k3Q@mail.gmail.com>
From: Andy Shevchenko <andy.shevchenko@gmail.com>
Date: Fri, 23 Feb 2024 14:29:40 +0200
Message-ID: <CAHp75Vf4=huDQ6211zY4aQem_BX3px4YwhfsA+YXcbGCJSV6Tg@mail.gmail.com>
Subject: Re: linux-next: build failure after merge of the gpio-brgl tree
To: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>, Bartosz Golaszewski <brgl@bgdev.pl>, 
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>, 
	Linux Next Mailing List <linux-next@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Fri, Feb 23, 2024 at 9:51=E2=80=AFAM Bartosz Golaszewski
<bartosz.golaszewski@linaro.org> wrote:
> On Fri, 23 Feb 2024 at 04:33, Stephen Rothwell <sfr@canb.auug.org.au> wro=
te:

> > After merging the gpio-brgl tree, today's linux-next build (x86_64
> > allmodconfig) failed like this:
> >
> > x86_64-linux-gnu-ld: vmlinux.o: in function `bgpio_write64':
> > gpio-mmio.c:(.text+0x1489427): undefined reference to `iowrite64'
> > x86_64-linux-gnu-ld: vmlinux.o: in function `bgpio_read64':
> > gpio-mmio.c:(.text+0x14894a0): undefined reference to `ioread64'
> > x86_64-linux-gnu-ld: vmlinux.o: in function `bgpio_write64be':
> > gpio-mmio.c:(.text+0x1489527): undefined reference to `iowrite64be'
> > x86_64-linux-gnu-ld: vmlinux.o: in function `bgpio_read64be':
> > gpio-mmio.c:(.text+0x14895a0): undefined reference to `ioread64be'
> >
> > Caused by commit
> >
> >   36e44186e0ba ("gpio: mmio: Support 64-bit BE access")
> >
> > I have used the gpio-brgl from next-20240222 for today.

> I'm not seeing this issue with allmodconfig on x86_64, could you give
> me some more information on the build environment?
>
> Andy: Could it be that the ifdefs you added should depend on
> CONFIG_64BIT and not only on the BITS_PER_LONG value?

If so, it will be the old issue as I just used what is currently done
for LE 64-bit accessors in the same file.
I think the problem here is different, i.e. io*64*() are not available
on all (64-bit) architectures in comparison to readq()/writeq().
I'll look at this later today.


--=20
With Best Regards,
Andy Shevchenko