Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 9 Jul 2021 15:33:16 -0700
From:   Andrew Morton <akpm@linux-foundation.org>
To:     Yang Huan <link@vivo.com>
Cc:     linux-kernel@vger.kernel.org, linux-mm@kvack.org, kernel@vivo.com,
        syzbot+b07d8440edb5f8988eea@syzkaller.appspotmail.com,
        Wang Qing <wangqing@vivo.com>
Subject: Re: [PATCH] mm/page_alloc: fix alloc_pages_bulk/set_page_owner
 panic on irq disabled
Message-Id: <20210709153316.a81214a453920dbcc2714ae6@linux-foundation.org>
In-Reply-To: <20210709090600.9804-1-link@vivo.com>
References: <20210709090600.9804-1-link@vivo.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Fri,  9 Jul 2021 17:05:59 +0800 Yang Huan <link@vivo.com> wrote:

> The problem is caused by set_page_owner alloc memory to save stack with
> GFP_KERNEL in local_riq disabled.
> So, we just can't assume that alloc flags should be same with new page,
> let's split it. But in most situation, same is ok, in alloc_pages_bulk,
> input GFP_ATOMIC when prep_new_pages

Please more fully describe the bug which is being fixed.  A link to the
sysbot report would be helpful.  I assume there was a stack backtrace
available?  Seeing the will help others to understand the bug.

And if you are able to identify the patch which introduced the bug, a
Fixes: tag would be helpful as well.

Thanks.