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;
Subject: Re: [PATCH 2/2] fuse: Replace page without copying in
 fuse_writepage_in_flight()
To:     Miklos Szeredi <miklos@szeredi.hu>
Cc:     linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
References: <154322517208.18737.3297786654135648324.stgit@localhost.localdomain>
 <154322557765.18737.14337090699283695815.stgit@localhost.localdomain>
 <CAJfpegvKxJ06hfXR5yxWkdz0GzTw7vR2x8E5bjdO1df1J1C54w@mail.gmail.com>
 <30c6d750-01b1-fb7e-eb80-0f60f242974a@virtuozzo.com>
 <CAJfpeguXy_Eebok9z_FCGvjGM8_qHHyTqsTMFeaJbmrW12kqYA@mail.gmail.com>
From:   Kirill Tkhai <ktkhai@virtuozzo.com>
Message-ID: <1cf4d11f-224f-ca0c-55cb-01754d429fe1@virtuozzo.com>
Date:   Tue, 15 Jan 2019 19:38:35 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.4.0
MIME-Version: 1.0
In-Reply-To: <CAJfpeguXy_Eebok9z_FCGvjGM8_qHHyTqsTMFeaJbmrW12kqYA@mail.gmail.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 15.01.2019 19:36, Miklos Szeredi wrote:
> On Tue, Jan 15, 2019 at 5:14 PM Kirill Tkhai <ktkhai@virtuozzo.com> wrote:
>>
>> On 15.01.2019 18:39, Miklos Szeredi wrote:
>>> On Mon, Nov 26, 2018 at 10:46 AM Kirill Tkhai <ktkhai@virtuozzo.com> wrote:
>>>>
>>>> It looks like we can optimize old_req page replacement
>>>> and avoid copying by simple updating the request's page.
>>>>
>>>> Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
>>>> ---
>>>>  fs/fuse/file.c |    2 +-
>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/fs/fuse/file.c b/fs/fuse/file.c
>>>> index c6650c68b31a..83b54b082c86 100644
>>>> --- a/fs/fuse/file.c
>>>> +++ b/fs/fuse/file.c
>>>> @@ -1778,7 +1778,7 @@ static bool fuse_writepage_in_flight(struct fuse_req *new_req,
>>>>         if (old_req->num_pages == 1 && old_req != first_req) {
>>>>                 struct backing_dev_info *bdi = inode_to_bdi(page->mapping->host);
>>>>
>>>> -               copy_highpage(old_req->pages[0], page);
>>>> +               swap(old_req->pages[0], page);
>>>
>>> This would mess up refcounting for all pages involved.   need to swap
>>> with the temp page in new_req.    Fixed version in #for-next.
>>
>> You are sure, page is just a simple pointer, not struct **page.
>> Then we would have had to change fuse_writepage_in_flight() to use ** pointer.
> 
> Using a struct page** would still have been broken, not because of
> refcounting, but because of putting the wrong page into the request
> (we do the temporary copy to avoid some issues with adding the page
> cache page directly into the request)

Ok, thanks for the explanation.

Kirill