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;
Date:   Fri, 24 Jul 2020 16:45:25 +0200
From:   Oleg Nesterov <oleg@redhat.com>
To:     Linus Torvalds <torvalds@linux-foundation.org>
Cc:     Hugh Dickins <hughd@google.com>, Michal Hocko <mhocko@kernel.org>,
        Linux-MM <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        Tim Chen <tim.c.chen@linux.intel.com>,
        Michal Hocko <mhocko@suse.com>
Subject: Re: [RFC PATCH] mm: silence soft lockups from unlock_page
Message-ID: <20200724144525.GB17209@redhat.com>
References: <CAHk-=whewL14RgwLZTXcNAnrDPt0H+sRJS6iDq0oGb6zwaBMxg@mail.gmail.com>
 <CAHk-=whb0=rjc1WR+F_r_syw5Ld4=ebuNJmmpaPEzfjZRD5Y-w@mail.gmail.com>
 <alpine.LSU.2.11.2007221359450.1017@eggly.anvils>
 <CAHk-=wi=vuc6sdu0m9nYd3gb8x5Xgnc6=TH=DTOy7qU96rZ9nw@mail.gmail.com>
 <CAHk-=whEjnsANEhTA3aqpNLZ3vv7huP7QAmcAEd-GUxm2YMo-Q@mail.gmail.com>
 <20200723124749.GA7428@redhat.com>
 <CAHk-=wgyc7en4=HddEYiz_RKJXfqe1JYv3BzHc=+_wYq9ti+LQ@mail.gmail.com>
 <20200723180100.GA21755@redhat.com>
 <CAHk-=wj9KWfs799xU5eW0J_hkee52C5kvFFmBV-A+vN7qNWnjA@mail.gmail.com>
 <CAHk-=wg=zqJ69JH3S9Ux0XKwZzjKqBP+ruLrmjW6NbNrpMuBNw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAHk-=wg=zqJ69JH3S9Ux0XKwZzjKqBP+ruLrmjW6NbNrpMuBNw@mail.gmail.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 07/23, Linus Torvalds wrote:
>
> IOW, I think we should do something like this (this is on top of my
> patch, since it has that wake_page_function() change in it, but notice
> how we have the exact same issue in our traditional
> autoremove_wake_function() usage).

...

> +static inline void list_del_init_careful(struct list_head *entry)
> +{
> +	__list_del_entry(entry);
> +	entry->prev = entry;
> +	smp_store_release(&entry->next, entry);
> +}
> +
...
>  static inline int list_empty_careful(const struct list_head *head)
>  {
> -	struct list_head *next = head->next;
> +	struct list_head *next = smp_load_acquire(&head->next);
>  	return (next == head) && (next == head->prev);
>  }

This (and your previous email) answers my concerns about memory barriers.

IIUC, finish_wait() could even use this version of list_empty_careful(),

	struct list_head *next = smp_load_acquire(&head->next);
	return (next == head) && !WARN_ON(next != head->prev);

iow, it doesn't really need to check next == head->prev as long as only
list_del_init_careful() can remove it from list.

Thanks!

Oleg.