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: <20200827013636.149307-1-allen.lkml@gmail.com> <CAHk-=whiEUUrtnbgUH2xsD0+jNyoXudYJ4hGCA55MCjryaHGjw@mail.gmail.com>
 <1598553133.4237.8.camel@HansenPartnership.com> <CAHk-=wi8o+FvfQkUiH_2MUs3J19FzfMzumOViAJ2aboGg9qY7Q@mail.gmail.com>
In-Reply-To: <CAHk-=wi8o+FvfQkUiH_2MUs3J19FzfMzumOViAJ2aboGg9qY7Q@mail.gmail.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Thu, 27 Aug 2020 11:48:19 -0700
Message-ID: <CAHk-=wingJWToQfoc+m2am7Q=7r8XD+6p0FXasCRAzOdcRyngw@mail.gmail.com>
Subject: Re: [PATCH] linux/kernel.h: add container_from()
To:     James Bottomley <James.Bottomley@hansenpartnership.com>
Cc:     Allen Pais <allen.lkml@gmail.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Kees Cook <keescook@chromium.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jens Axboe <axboe@kernel.dk>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Aug 27, 2020 at 11:40 AM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> On Thu, Aug 27, 2020 at 11:32 AM James Bottomley
> <James.Bottomley@hansenpartnership.com> wrote:
> >
> >
> > The tasklet rework people don't want to use container_of, which was our
> > first suggestion, because it produces lines which are "too long".
>
> WTF?

Side note: I'm entirely serious. If somebody has problems with "too
long lines", they are doing things wrong. Just ignore them.

Changing standard kernel interfaces is the wrong solution. What's
next? Using 2-character indentation like some broken projects do just
to make lines shorter and encourage people to do deep nesting?

No. The solution is to not write crap code with overly complex expressions.

"container_of()" has been a _very_ successful model, and the only
reason to ever wrap it is if you already *know* the type, and you wrap
it with an inline function that actually checks it.

For a lot of (usually) good examples of this, just do a

     git grep return.container_of

and find things like

    static inline struct kvm_pit *pit_state_to_pit(struct kvm_kpit_state *ps)
    {
        return container_of(ps, struct kvm_pit, pit_state);
    }

which now results in a type-checked *true* simplification of container-of.

It basically creates a specialized version which documents what it
does in the name, and does proper type checking, and doesn't try to be
another name for the existing generic container_of().

Seriously. It sounds to me like the tasklet rework people are people
we want to avoid. They're doing completely the wrong thing.

So just throw that garbage out.

                      Linus