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;
Organization: Red Hat UK Ltd. Registered Address: Red Hat UK Ltd, Amberley
        Place, 107-111 Peascod Street, Windsor, Berkshire, SI4 1TE, United
        Kingdom.
        Registered in England and Wales under Company Registration No. 3798903
From:   David Howells <dhowells@redhat.com>
In-Reply-To: <67f55d920f40bf6c49643af08fe8a5cfc97a9542.camel@kernel.org>
References: <67f55d920f40bf6c49643af08fe8a5cfc97a9542.camel@kernel.org> <163456861570.2614702.14754548462706508617.stgit@warthog.procyon.org.uk> <163456871794.2614702.15398637170877934146.stgit@warthog.procyon.org.uk>
To:     Jeff Layton <jlayton@kernel.org>
Cc:     dhowells@redhat.com, linux-cachefs@redhat.com,
        Dave Wysochanski <dwysocha@redhat.com>,
        Trond Myklebust <trond.myklebust@hammerspace.com>,
        Anna Schumaker <anna.schumaker@netapp.com>,
        linux-nfs@vger.kernel.org, Ilya Dryomov <idryomov@gmail.com>,
        ceph-devel@vger.kernel.org, Steve French <sfrench@samba.org>,
        linux-cifs@vger.kernel.org, Eric Van Hensbergen <ericvh@gmail.com>,
        Latchesar Ionkov <lucho@ionkov.net>,
        Dominique Martinet <asmadeus@codewreck.org>,
        v9fs-developer@lists.sourceforge.net,
        Trond Myklebust <trondmy@hammerspace.com>,
        Matthew Wilcox <willy@infradead.org>,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Omar Sandoval <osandov@osandov.com>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        linux-afs@lists.infradead.org, linux-fsdevel@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 06/67] nfs, cifs, ceph, 9p: Disable use of fscache prior to its rewrite
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <3164228.1634727463.1@warthog.procyon.org.uk>
Date:   Wed, 20 Oct 2021 11:57:43 +0100
Message-ID: <3164229.1634727463@warthog.procyon.org.uk>
Precedence: bulk

Jeff Layton <jlayton@kernel.org> wrote:

> The typical way to do this would be to rebrand the existing FSCACHE
> Kconfig symbols into FSCACHE_OLD and then build the new fscache
> structure such that it exists in parallel with the old.

That, there, is nub of the problem.

You can't have parallel cachefiles drivers: There's a single userspace
interface (/dev/cachefiles) and only one driver can register it.  You would
need to decide at compile time whether you want the converted or the
unconverted network filesystems to be cached.

> You'd then just drop the old infrastructure once all of the fs's are
> converted to the new. You could even make them conflict with one another in
> Kconfig too, so that only one could be built in during the transition period
> if supporting both at runtime is too difficult.
> 
> This approach of disabling everything is much more of an all-or-nothing
> affair. It may mean less "churn" overall, but it seems less "nice"
> because you have an interval of commits where fscache is non-functional.
> 
> I'm not necessarily opposed to this approach, but I'd like to better
> understand why doing it this way was preferred.

I'm trying to avoid adding two parallel drivers, but change in place so that I
can test parts of it as I go along.

David