From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Date: Sat, 7 Oct 2017 03:31:52 +0100
From: Al Viro <viro@ZenIV.linux.org.uk>
To: Jia-Ju Bai <baijiaju1990@163.com>
Cc: bcrl@kvack.org, linux-aio@kvack.org, linux-fsdevel@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [BUG] fs/aio: A possible sleep-in-atomic bug in aio_migratepage
Message-ID: <20171007023152.GQ21978@ZenIV.linux.org.uk>
References: <dee03130-37ef-3958-455b-b1fca415b2c1@163.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <dee03130-37ef-3958-455b-b1fca415b2c1@163.com>
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Sat, Oct 07, 2017 at 09:36:48AM +0800, Jia-Ju Bai wrote:
> According to fs/aio.c, cond_resched is called under a spinlock,
> and the function call path is:
> aio_migratepage (acquire the spinlock)
>   migrate_page_copy
>     copy_huge_page
>       __copy_gigantic_page
>         cond_resched
>           might_sleep
> 
> This bug is found by my static analysis tool and my code review.
> A possible fix is to remove cond_resched in __copy_gigantic_page.

How the hell would a huge page end up in page cache of inode created
by aio_private_file()???