Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 13 Mar 2019 11:01:43 -0700
From:   Matthew Wilcox <willy@infradead.org>
To:     Laurent Dufour <ldufour@linux.ibm.com>
Cc:     lsf-pc@lists.linux-foundation.org, Linux-MM <linux-mm@kvack.org>,
        linux-kernel@vger.kernel.org,
        "Liam R. Howlett" <Liam.Howlett@Oracle.com>
Subject: Re: [LSF/MM TOPIC] Using XArray to manage the VMA
Message-ID: <20190313180142.GK19508@bombadil.infradead.org>
References: <7da20892-f92a-68d8-4804-c72c1cb0d090@linux.ibm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <7da20892-f92a-68d8-4804-c72c1cb0d090@linux.ibm.com>
User-Agent: Mutt/1.9.2 (2017-12-15)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Mar 13, 2019 at 04:10:14PM +0100, Laurent Dufour wrote:
> If this is not too late and if there is still place available, I would like
> to attend the MM track and propose a topic about using the XArray to replace
> the VMA's RB tree and list.

If there isn't room on the schedule, then Laurent and I are definitely
going to sneak off and talk about this ourselves at some point.  Having a
high-bandwidth conversation about this is going to be really important
for us, and I think having other people involved would be good.

If there're still spots, it'd be good to have Liam Howlett join us.
He's doing the actual writing-of-code for the Maple Tree at the moment
(I did some earlier on, but recent commits are all him).

> Using the XArray in place of the VMA's tree and list seems to be a first
> step to the long way of removing/replacing the mmap_sem.
> However, there are still corner cases to address like the VMA splitting and
> merging which may raise some issue. Using the XArray's specifying locking
> would not be enough to handle the memory management, and additional fine
> grain locking like a per VMA one could be studied, leading to further
> discussion about the merging of the VMA.
> 
> In addition, here are some topics I'm interested in:
> - Test cases to choose for demonstrating mm features or fixing mm bugs
> proposed by Balbir Singh
> - mm documentation proposed by Mike Rapoport
> 
> Laurent.
>