Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <CAK8P3a2Zg2QDS1_Ysn8-Zqqd+K7bbTFS7JV7gPabp6nvPiKaog@mail.gmail.com>
 <91E67F46-A3C7-4159-9E0C-C6C6306F3669@inria.fr> <CAK8P3a2iAsemAQdbTZ_E7GGGCXAOeWbjSjLgXEsd5sg_buZWhw@mail.gmail.com>
 <CAHk-=wgO0V9OdY+DFm-f0qZYMyFSm0ptReO+_qgSTEpBLtFV7Q@mail.gmail.com>
 <d971a684-ccd9-3839-1e30-c166fd55cf49@inria.fr> <CAHk-=wiViikY0szsJGipSxFmMwdsvxjm7SwDfwNfMHYvQ64kAA@mail.gmail.com>
 <CAK8P3a1m80u+eVnoSJ-APihjNQ1se9=FG+E6tKBb-hRJx5FAVg@mail.gmail.com> <74bed19a-713f-1a25-8142-cf32984beada@I-love.SAKURA.ne.jp>
In-Reply-To: <74bed19a-713f-1a25-8142-cf32984beada@I-love.SAKURA.ne.jp>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Thu, 2 Jun 2022 14:11:13 +0200
Message-ID: <CAK8P3a02fpOPkXSEEd2eDoryVN2zZcRvzJ2qRceAQmCqB8ghag@mail.gmail.com>
Subject: Re: mainline build failure due to f1e4c916f97f ("drm/edid: add EDID
 block count and size helpers")
To:     Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc:     Arnd Bergmann <arnd@arndb.de>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Keisuke Nishimura <keisuke.nishimura@inria.fr>,
        Kentaro Takeda <takedakn@nttdata.co.jp>,
        Ayush Sawal <ayush.sawal@chelsio.com>,
        Vinay Kumar Yadav <vinay.yadav@chelsio.com>,
        Rohit Maheshwari <rohitm@chelsio.com>,
        Julia Lawall <Julia.Lawall@inria.fr>,
        Jani Nikula <jani.nikula@intel.com>,
        Sudip Mukherjee <sudipm.mukherjee@gmail.com>,
        Russell King <linux@armlinux.org.uk>,
        Viresh Kumar <vireshk@kernel.org>,
        Shiraz Hashim <shiraz.linux.kernel@gmail.com>,
        =?UTF-8?B?VmlsbGUgU3lyasOkbMOk?= <ville.syrjala@linux.intel.com>,
        Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
        Maxime Ripard <mripard@kernel.org>,
        Thomas Zimmermann <tzimmermann@suse.de>,
        David Airlie <airlied@linux.ie>,
        Daniel Vetter <daniel@ffwll.ch>,
        dri-devel <dri-devel@lists.freedesktop.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        SoC Team <soc@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Jun 2, 2022 at 1:21 PM Tetsuo Handa
<penguin-kernel@i-love.sakura.ne.jp> wrote:
> On 2022/06/02 16:38, Arnd Bergmann wrote:
> >> But let's cc the tomoyo and chelsio people.
> >
> > I think both of them work because the structures are always
> > embedded inside of larger structures that have at least word
> > alignment. This is the thing I was looking for, and the
> > __packed attribute was added in error, most likely copied
> > from somewhere else.
>
> The __packed in "struct tomoyo_shared_acl_head" is to embed next
> naturally-aligned member of a larger struct into the bytes that
> would have been wasted if __packed is not specified. For example,
>
> struct tomoyo_shared_acl_head {
>         struct list_head list;
>         atomic_t users;
> } __packed;
>
> struct tomoyo_condition {
>         struct tomoyo_shared_acl_head head;
>         u32 size; /* Memory size allocated for this entry. */
>         (...snipped...)
> };
>
> saves 4 bytes on 64 bits build.
>
> If the next naturally-aligned member of a larger struct is larger than
> the bytes that was saved by __packed, the saved bytes will be unused.

Ok, got it. I think as gcc should still be able to always figure out the
alignment when accessing the atomic, without ever falling back
to byte access on an atomic_get() or atomic_set().

To be on the safe side, I would still either move the __packed attribute
to the 'list' member, or make the structure '__aligned(4)'.

       Arnd