Re: [PATCH] kernel.h: Add for_each_if()

From: Randy Dunlap <rdunlap@infradead.org>
To: NeilBrown <neilb@suse.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Daniel Vetter <daniel@ffwll.ch>
Cc: LKML <linux-kernel@vger.kernel.org>,
	DRI Development <dri-devel@lists.freedesktop.org>,
	Intel Graphics Development <intel-gfx@lists.freedesktop.org>,
	Gustavo Padovan <gustavo@padovan.org>,
	Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
	Sean Paul <seanpaul@chromium.org>,
	David Airlie <airlied@linux.ie>,
	Kees Cook <keescook@chromium.org>, Ingo Molnar <mingo@kernel.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Wei Wang <wvw@google.com>, Stefan Agner <stefan@agner.ch>,
	Andrei Vagin <avagin@openvz.org>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	Yisheng Xie <ysxie@foxmail.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Daniel Vetter <daniel.vetter@intel.com>
Subject: Re: [PATCH] kernel.h: Add for_each_if()
Date: Fri, 13 Jul 2018 16:42:59 -0700	[thread overview]
Message-ID: <79133322-b04b-f005-f1f6-25c28c5058e4@infradead.org> (raw)
In-Reply-To: <87bmbavhai.fsf@notabene.neil.brown.name>

On 07/13/2018 04:37 PM, NeilBrown wrote:
> On Wed, Jul 11 2018, Andrew Morton wrote:
> 
>> On Wed, 11 Jul 2018 13:51:08 +0200 Daniel Vetter <daniel@ffwll.ch> wrote:
>>
>>> But I still have the situation that a bunch of maintainers acked this
>>> and Andrew Morton defacto nacked it, which I guess means I'll keep the
>>> macro in drm? The common way to go about this seems to be to just push
>>> the patch series with the ack in some pull request to Linus and ignore
>>> the people who raised questions, but not really my thing.
>>
>> Heh.
>>
>> But, am I wrong?  Code which uses regular kernel style doesn't have
>> these issues.  We shouldn't be enabling irregular style - we should be
>> making such sites more regular.  The fact that the compiler generates a
>> nice warning in some cases simply helps us with that.
> 
> I think you are wrong .... or at least, not completely correct.
> 
> I think it is perfectly acceptable in Linux to have code like:
> 
>   for (....)
>   	if (x)
>         	something();
>         else
>         	something_else();
> 
> Would you agree?  If not, then I'm the one who is wrong.  Otherwise....

coding-style.rst says:
Also, use braces when a loop contains more than a single simple statement:

> The problem is that for certain poorly written for_each_foo() macros,
> such as blkg_for_each_descendant_pre() (and several others identified in
> this patch series), writing
> 
>    blkg_for_each_descendant_pre(...)
>      	if (x)
>         	something();
>         else
>         	something_else();
> 
> will trigger a compiler warning.  This is inconsistent with the
> behaviour of a simple "for".
> So I do think that the macros should be fixed, and I don't think that
> sprinkling extra braces is an appropriate response.
> 
> I'm not personally convinced that writing
>    if_no_else(cond)
> is easier than just writing
>    if (!(cond)); else

agreed.

> in these macros, but I do think that the macros should be fixed and
> maybe this is the path-of-least-resistance to getting it done.

I'm not opposed to fixing some macros, but some of these macros are just
ease-of-less-typing shortcuts.  They don't improve readability at all;
they harm it.  (of course, that is just one opinion :)

-- 
~Randy