linux-kernel - Re: [PATCH] media: fsl-viu: Use proper check for presence of {out,in}

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAMuHMdUKePx7WWAAB3LCFWd29yis1zC2oCPrgmdH2n8fOhKyMQ@mail.gmail.com>
Date:   Fri, 18 Jan 2019 11:45:16 +0100
From:   Geert Uytterhoeven <geert@...ux-m68k.org>
To:     Hans Verkuil <hverkuil@...all.nl>
Cc:     Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        Michael Ellerman <mpe@...erman.id.au>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Arnd Bergmann <arnd@...db.de>,
        linuxppc-dev <linuxppc-dev@...ts.ozlabs.org>,
        Linux Media Mailing List <linux-media@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] media: fsl-viu: Use proper check for presence of {out,in}_be32()

Hi Hans,

On Fri, Jan 18, 2019 at 11:08 AM Hans Verkuil <hverkuil@...all.nl> wrote:
> This patch is still in my patchwork todo list, and I wonder who will pick this up,
> especially the change to arch/powerpc/include/asm/io.h.

I think the powerpc tree makes most sense.

> Wouldn't it be easier to just fix this in fsl-viu.c only by doing this:
>
> #ifndef CONFIG_PPC
> #ifndef out_be32
> #define out_be32(v, a)  iowrite32be(a, (void __iomem *)v)
> #endif
> #ifndef in_be32
> #define in_be32(a)      ioread32be((void __iomem *)a)
> #endif
> #endif
>
> Basically just your patch, but without removing #ifndef CONFIG_PPC.
>
> That way there is no need to touch arch/powerpc/include/asm/io.h.

While I agree that avoids touching the powerpc tree, it doesn't solve
the problem
in the long run.
All drivers using e.g. out_be32() would need an extra check for CONFIG_PPC
to enable compile-testing.  The same is true when handling this in asm-generic.

> On 6/15/18 10:10 AM, Hans Verkuil wrote:
> > On 08/06/18 11:48, Geert Uytterhoeven wrote:
> >> When compile-testing on m68k or microblaze:
> >>
> >>     drivers/media/platform/fsl-viu.c:41:1: warning: "out_be32" redefined
> >>     drivers/media/platform/fsl-viu.c:42:1: warning: "in_be32" redefined
> >>
> >> Fix this by replacing the check for PowerPC by checks for the presence
> >> of {out,in}_be32().
> >>
> >> As PowerPC implements the be32 accessors using inline functions instead
> >> of macros, identity definions are added for all accessors to make the
> >> above checks work.
> >>
> >> Fixes: 29d750686331a1a9 ("media: fsl-viu: allow building it with COMPILE_TEST")
> >> Signed-off-by: Geert Uytterhoeven <geert@...ux-m68k.org>
> >
> > Reviewed-by: Hans Verkuil <hans.verkuil@...co.com>
> >
> > Should this go through the media tree or powerpc tree? Either way works for me.
> >
> > Regards,
> >
> >       Hans
> >
> >> ---
> >> Compile-tested on m68k, microblaze, and powerpc.
> >> Assembler output before/after compared for powerpc.
> >> ---
> >>  arch/powerpc/include/asm/io.h    | 14 ++++++++++++++
> >>  drivers/media/platform/fsl-viu.c |  4 +++-
> >>  2 files changed, 17 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/arch/powerpc/include/asm/io.h b/arch/powerpc/include/asm/io.h
> >> index e0331e7545685c5f..3741183ae09349f1 100644
> >> --- a/arch/powerpc/include/asm/io.h
> >> +++ b/arch/powerpc/include/asm/io.h
> >> @@ -222,6 +222,20 @@ static inline void out_be64(volatile u64 __iomem *addr, u64 val)
> >>  #endif
> >>  #endif /* __powerpc64__ */
> >>
> >> +#define in_be16 in_be16
> >> +#define in_be32 in_be32
> >> +#define in_be64 in_be64
> >> +#define in_le16 in_le16
> >> +#define in_le32 in_le32
> >> +#define in_le64 in_le64
> >> +
> >> +#define out_be16 out_be16
> >> +#define out_be32 out_be32
> >> +#define out_be64 out_be64
> >> +#define out_le16 out_le16
> >> +#define out_le32 out_le32
> >> +#define out_le64 out_le64
> >> +
> >>  /*
> >>   * Low level IO stream instructions are defined out of line for now
> >>   */
> >> diff --git a/drivers/media/platform/fsl-viu.c b/drivers/media/platform/fsl-viu.c
> >> index e41510ce69a40815..5d5e030c9c980647 100644
> >> --- a/drivers/media/platform/fsl-viu.c
> >> +++ b/drivers/media/platform/fsl-viu.c
> >> @@ -37,8 +37,10 @@
> >>  #define VIU_VERSION         "0.5.1"
> >>
> >>  /* Allow building this driver with COMPILE_TEST */
> >> -#ifndef CONFIG_PPC
> >> +#ifndef out_be32
> >>  #define out_be32(v, a)      iowrite32be(a, (void __iomem *)v)
> >> +#endif
> >> +#ifndef in_be32
> >>  #define in_be32(a)  ioread32be((void __iomem *)a)
> >>  #endif

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds