Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
Date:   Fri, 6 May 2022 17:59:45 +0900
From:   Naoya Horiguchi <naoya.horiguchi@linux.dev>
To:     zhenwei pi <pizhenwei@bytedance.com>
Cc:     akpm@linux-foundation.org, naoya.horiguchi@nec.com,
        linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        Wu Fengguang <fengguang.wu@intel.com>
Subject: Re: [PATCH 4/4] mm/memofy-failure.c: add hwpoison_filter for soft
 offline
Message-ID: <20220506085945.GD1356094@u2004>
References: <20220429142206.294714-1-pizhenwei@bytedance.com>
 <20220429142206.294714-5-pizhenwei@bytedance.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20220429142206.294714-5-pizhenwei@bytedance.com>
Precedence: bulk

On Fri, Apr 29, 2022 at 10:22:06PM +0800, zhenwei pi wrote:
> hwpoison_filter is missing in the soft offline path, this leads an
> issue: after enabling the corrupt filter, the user process still has
> a chance to inject hwpoison fault by
> madvise(addr, len, MADV_SOFT_OFFLINE) at PFN which is expected to
> reject.

The motivation is fine to me. Thank you for finding this.

> 
> Cc: Wu Fengguang <fengguang.wu@intel.com>
> Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
> ---
>  mm/memory-failure.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index a6a27c8b800f..6564f5a34658 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -2313,7 +2313,9 @@ static void put_ref_page(struct page *page)
>   * @pfn: pfn to soft-offline
>   * @flags: flags. Same as memory_failure().
>   *
> - * Returns 0 on success, otherwise negated errno.
> + * Returns 0 on success
> + *         -EOPNOTSUPP for memory_filter() filtered the error event

Using word hwpoison_filter() rather than memory_filter() seems better to me.

> + *         < 0 otherwise negated errno.
>   *
>   * Soft offline a page, by migration or invalidation,
>   * without killing anything. This is for the case when
> @@ -2350,6 +2352,11 @@ int soft_offline_page(unsigned long pfn, int flags)
>  		return -EIO;
>  	}
>  
> +	if (hwpoison_filter(page)) {
> +		put_ref_page(ref_page);
> +		return -EOPNOTSUPP;
> +	}
> +

Based on the assumption behind hwpoison_filter(), calling it after
get_hwpoison_page() would be better?

Thanks,
Naoya Horiguchi