Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
To:     rientjes@google.com, mhocko@kernel.org
Cc:     akpm@linux-foundation.org, aarcange@redhat.com, guro@fb.com,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [patch v2] mm, oom: fix concurrent munlock and oom reaperunmap
From:   Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
References: <20180419063556.GK17484@dhcp22.suse.cz>
        <alpine.DEB.2.21.1804191214130.157851@chino.kir.corp.google.com>
        <20180420082349.GW17484@dhcp22.suse.cz>
        <20180420124044.GA17484@dhcp22.suse.cz>
        <alpine.DEB.2.21.1804212019400.84222@chino.kir.corp.google.com>
In-Reply-To: <alpine.DEB.2.21.1804212019400.84222@chino.kir.corp.google.com>
Message-Id: <201804221248.CHE35432.FtOMOLSHOFJFVQ@I-love.SAKURA.ne.jp>
Date:   Sun, 22 Apr 2018 12:48:13 +0900
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

David Rientjes wrote:
> How have you tested this?
> 
> I'm wondering why you do not see oom killing of many processes if the 
> victim is a very large process that takes a long time to free memory in 
> exit_mmap() as I do because the oom reaper gives up trying to acquire 
> mm->mmap_sem and just sets MMF_OOM_SKIP itself.
> 

We can call __oom_reap_task_mm() from exit_mmap() (or __mmput()) before
exit_mmap() holds mmap_sem for write. Then, at least memory which could
have been reclaimed if exit_mmap() did not hold mmap_sem for write will
be guaranteed to be reclaimed before MMF_OOM_SKIP is set.