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;
Date:   Tue, 11 Feb 2020 00:21:49 +0300
From:   "Kirill A. Shutemov" <kirill@shutemov.name>
To:     Qian Cai <cai@lca.pw>
Cc:     Matthew Wilcox <willy@infradead.org>, akpm@linux-foundation.org,
        elver@google.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH -next] mm/filemap: fix a data race in filemap_fault()
Message-ID: <20200210212149.u6zlpi2jefi2vkfg@box>
References: <1581354029-20154-1-git-send-email-cai@lca.pw>
 <20200210172511.GL8731@bombadil.infradead.org>
 <20200210180546.vt7yhdjav5oinij7@box>
 <1581364697.7365.45.camel@lca.pw>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1581364697.7365.45.camel@lca.pw>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Feb 10, 2020 at 02:58:17PM -0500, Qian Cai wrote:
> On Mon, 2020-02-10 at 21:05 +0300, Kirill A. Shutemov wrote:
> > On Mon, Feb 10, 2020 at 09:25:11AM -0800, Matthew Wilcox wrote:
> > > On Mon, Feb 10, 2020 at 12:00:29PM -0500, Qian Cai wrote:
> > > > @@ -2622,7 +2622,7 @@ void filemap_map_pages(struct vm_fault *vmf,
> > > >  		if (page->index >= max_idx)
> > > >  			goto unlock;
> > > >  
> > > > -		if (file->f_ra.mmap_miss > 0)
> > > > +		if (data_race(file->f_ra.mmap_miss > 0))
> > > >  			file->f_ra.mmap_miss--;
> > > 
> > > How is this safe?  Two threads can each see 1, and then both decrement the
> > > in-memory copy, causing it to end up at -1.
> > 
> > Right, it is bogus.
> > 
> > Below is my completely untested attempt on fix this. It still allows
> > races, but they will only lead to missed accounting, but not underflow.
> 
> Looks good to me. Do you plan to send out an official patch?

Feel free to submit it. After testing.

-- 
 Kirill A. Shutemov