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;
Date:   Fri, 11 Nov 2022 11:27:32 -0800
From:   Andrew Morton <akpm@linux-foundation.org>
To:     Zhongkun He <hezhongkun.hzk@bytedance.com>
Cc:     corbet@lwn.net, mhocko@suse.com, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org, linux-api@vger.kernel.org,
        linux-doc@vger.kernel.org
Subject: Re: [PATCH v2] mm: add new syscall pidfd_set_mempolicy().
Message-Id: <20221111112732.30e1696bcd0d5b711c188a9a@linux-foundation.org>
In-Reply-To: <20221111084051.2121029-1-hezhongkun.hzk@bytedance.com>
References: <20221111084051.2121029-1-hezhongkun.hzk@bytedance.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Fri, 11 Nov 2022 16:40:51 +0800 Zhongkun He <hezhongkun.hzk@bytedance.com> wrote:

> Page allocation usage of task or vma policy occurs in the fault
> path where we hold the mmap_lock for read. because replacing the
> task or vma policy requires that the mmap_lock be held for write,
> the policy can't be freed out from under us while we're using
> it for page allocation. But there are some corner cases(e.g.
> alloc_pages()) which not acquire any lock for read during the
> page allocation. For this reason, task_work is used in
> mpol_put_async() to free mempolicy in  pidfd_set_mempolicy().
> Thuse, it avoids into race conditions.

This sounds a bit suspicious.  Please share much more detail about
these races.  If we proced with this design then mpol_put_async()
shouild have comments which fully describe the need for the async free.

How do we *know* that these races are fully prevented with this
approach?  How do we know that mpol_put_async() won't free the data
until the race window has fully passed?

Also, in some situations mpol_put_async() will free the data
synchronously anyway, so aren't these races still present?


Secondly, why was the `flags' argument added?  We might use it one day?
For what purpose?  I mean, every syscall could have a does-nothing
`flags' arg, but we don't do that.  What's the plan here?