Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: Re: [PATCH] mm: memory-failure: fix potential page refcnt leak in
 memory_failure()
To:     Naoya Horiguchi <naoya.horiguchi@linux.dev>
CC:     <akpm@linux-foundation.org>, <naoya.horiguchi@nec.com>,
        <linux-mm@kvack.org>, <linux-kernel@vger.kernel.org>
References: <20230701072837.1994253-1-linmiaohe@huawei.com>
 <20230703063713.GA3012709@ik1-406-35019.vs.sakura.ne.jp>
From:   Miaohe Lin <linmiaohe@huawei.com>
Message-ID: <1e8ff9bf-c81b-9c39-9f44-2f82efcd36ec@huawei.com>
Date:   Mon, 3 Jul 2023 19:27:25 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101
 Thunderbird/78.6.0
MIME-Version: 1.0
In-Reply-To: <20230703063713.GA3012709@ik1-406-35019.vs.sakura.ne.jp>
Content-Type: text/plain; charset="utf-8"
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2023/7/3 14:37, Naoya Horiguchi wrote:
> On Sat, Jul 01, 2023 at 03:28:37PM +0800, Miaohe Lin wrote:
>> put_ref_page() is not called to drop extra refcnt when comes from madvise
>> in the case pfn is valid but pgmap is NULL leading to page refcnt leak.
> 
> Is this test scenario realistic one?  I don't think that we can call madvise() for
> such a device memory page.  If this is the case, this issue can be thought as
> potentioal one (so no need to send to stable). 

This is a theoretical problem. If pagemap can ever be NULL after page refcnt is increased, this
would occur. I think it's not needed to send to stable too.

> 
>>
>> Fixes: 1e8aaedb182d ("mm,memory_failure: always pin the page in madvise_inject_error")
>> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
> 
> Anyway, the patch looks good to me. Thanks you.
> 
> Acked-by: Naoya Horiguchi <naoya.horiguchi@nec.com>

Thanks for your review and comment.