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;
MIME-Version: 1.0
References: <20201006230930.3908-1-rcampbell@nvidia.com> <20201007082517.GC6984@quack2.suse.cz>
In-Reply-To: <20201007082517.GC6984@quack2.suse.cz>
From:   Dan Williams <dan.j.williams@intel.com>
Date:   Wed, 7 Oct 2020 14:59:59 -0700
Message-ID: <CAPcyv4j=q4n1PDW5vwOFpkeDEoDv2_xXBL+xoGsRwPn_ej=hnA@mail.gmail.com>
Subject: Re: [PATCH] ext4/xfs: add page refcount helper
To:     Jan Kara <jack@suse.cz>
Cc:     Ralph Campbell <rcampbell@nvidia.com>,
        Linux MM <linux-mm@kvack.org>,
        linux-xfs <linux-xfs@vger.kernel.org>,
        linux-fsdevel <linux-fsdevel@vger.kernel.org>,
        linux-nvdimm <linux-nvdimm@lists.01.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-ext4 <linux-ext4@vger.kernel.org>,
        Matthew Wilcox <willy@infradead.org>,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        "Theodore Ts'o" <tytso@mit.edu>, Christoph Hellwig <hch@lst.de>,
        Andreas Dilger <adilger.kernel@dilger.ca>,
        "Darrick J. Wong" <darrick.wong@oracle.com>,
        Andrew Morton <akpm@linux-foundation.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Oct 7, 2020 at 1:25 AM Jan Kara <jack@suse.cz> wrote:
>
> On Tue 06-10-20 16:09:30, Ralph Campbell wrote:
> > There are several places where ZONE_DEVICE struct pages assume a reference
> > count == 1 means the page is idle and free. Instead of open coding this,
> > add a helper function to hide this detail.
> >
> > Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
> > Reviewed-by: Christoph Hellwig <hch@lst.de>
>
> Looks as sane direction but if we are going to abstract checks when
> ZONE_DEVICE page is idle, we should also update e.g.
> mm/swap.c:put_devmap_managed_page() or
> mm/gup.c:__unpin_devmap_managed_user_page() (there may be more places like
> this but I found at least these two...). Maybe Dan has more thoughts about
> this.

Yes, but I think that cleanup comes once the idle page count is
unified to be 0 across typical and ZONE_DEVICE pages. Then
free_devmap_managed_page() can be moved internal to __put_page(). For
this patch it's just hiding the "idle == 1" assumption from
dax-filesystems.