Re: [PATCH 2/2] fs/dcache: Make negative dentries easier to be reclaimed

From: Waiman Long <longman@redhat.com>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matthew Wilcox <willy@infradead.org>,
	Al Viro <viro@zeniv.linux.org.uk>,
	Jonathan Corbet <corbet@lwn.net>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>,
	linux-mm <linux-mm@kvack.org>,
	"open list:DOCUMENTATION" <linux-doc@vger.kernel.org>,
	"Luis R. Rodriguez" <mcgrof@kernel.org>,
	Kees Cook <keescook@chromium.org>, Jan Kara <jack@suse.cz>,
	Paul McKenney <paulmck@linux.vnet.ibm.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Ingo Molnar <mingo@kernel.org>,
	Miklos Szeredi <mszeredi@redhat.com>,
	Larry Woodman <lwoodman@redhat.com>,
	James Bottomley <James.Bottomley@hansenpartnership.com>,
	"Wangkai (Kevin,C)" <wangkai86@huawei.com>,
	Michal Hocko <mhocko@kernel.org>
Subject: Re: [PATCH 2/2] fs/dcache: Make negative dentries easier to be reclaimed
Date: Tue, 28 Aug 2018 21:18:18 -0400	[thread overview]
Message-ID: <644e34e7-48b0-3bf4-49c9-a04024d3ca1b@redhat.com> (raw)
In-Reply-To: <CA+55aFxy1vH2CamZ_pdFohKgSJgi1i2MkeaY1qX8NdFK8Xu8Ww@mail.gmail.com>

On 08/28/2018 07:10 PM, Linus Torvalds wrote:
> On Tue, Aug 28, 2018 at 3:29 PM Waiman Long <longman@redhat.com> wrote:
>> Yes, I can rewrite it. What is the problem with the abbreviated form?
> Either gcc rewrites it for you, or you end up _actually_ using a
> function pointer and calling through it.

Yes, function pointer will be really bad.
>
> The latter would be absolutely horribly bad for something like
> "list_add()", which should expand to just a couple of instructions.
>
> And the former would be ok, except for the "you wrote code the garbage
> way, and then depended on the compiler fixing it up". Which we
> generally try to avoid in the kernel.
>
> (Don't get me wrong - we definitely depend on the compiler doing a
> good job at CSE and dead code elimination etc, but generally we try to
> avoid the whole "compiler has to rewrite code to be good" model).
>
>                  Linus

I see your point here. I will rewrite to use the regular if-then-else.

Thanks,
Longman