Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <YJvwVq3Gl35RQrIe@casper.infradead.org>
In-Reply-To: <YJvwVq3Gl35RQrIe@casper.infradead.org>
From:   Steve French <smfrench@gmail.com>
Date:   Wed, 12 May 2021 14:28:41 -0500
Message-ID: <CAH2r5msOQsdeknBdTsfMXYzrb5=NuKEBPc4WD1CkYp10t19Guw@mail.gmail.com>
Subject: Re: Removing readpages aop
To:     Matthew Wilcox <willy@infradead.org>
Cc:     linux-nfs <linux-nfs@vger.kernel.org>,
        CIFS <linux-cifs@vger.kernel.org>,
        v9fs-developer@lists.sourceforge.net,
        linux-fsdevel <linux-fsdevel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

I don't have any objections as long as:
- we see at least mild performance benefit (or at least we are
confident that no performance loss)
- it passes regression tests (the usual xfstest bucket)
- it doesn't complicate the code too much (sounds like it actually
might simplify it, but needs a little more work)
- make sure that the usual tuning parms still work (e.g. "rsize" and
"rasize" mount options) or we can figure out a sane way to autotune
readhead so those wouldn't be needed for any workload

But currently since we get the most benefit from multichannel (as that
allows even better parallelization of i/o) ... I have been focused on
various multichannel issues (low credit situations, reconnect, fall
back to different channels when weird errors, adjusting channels
dynamically when server adds or removes adapters on the fly) for the
short term

On Wed, May 12, 2021 at 10:31 AM Matthew Wilcox <willy@infradead.org> wrote:
>
> In Linus' current tree, there are just three filesystems left using the
> readpages address_space_operation:
>
> $ git grep '\.readpages'
> fs/9p/vfs_addr.c:       .readpages = v9fs_vfs_readpages,
> fs/cifs/file.c: .readpages = cifs_readpages,
> fs/nfs/file.c:  .readpages = nfs_readpages,
>
> I'd love to finish getting rid of ->readpages as it would simplify
> the VFS.  AFS and Ceph were both converted since 5.12 to use
> netfs_readahead().  Is there any chance we might get the remaining three
> filesystems converted in the next merge window?
>


-- 
Thanks,

Steve