[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YXHntB2O0ACr0pbz@relinquished.localdomain>
Date: Thu, 21 Oct 2021 15:20:36 -0700
From: Omar Sandoval <osandov@...ndov.com>
To: David Howells <dhowells@...hat.com>
Cc: linux-cachefs@...hat.com, ceph-devel@...r.kernel.org,
linux-afs@...ts.infradead.org,
Anna Schumaker <anna.schumaker@...app.com>,
linux-nfs@...r.kernel.org,
Kent Overstreet <kent.overstreet@...il.com>,
linux-mm@...ck.org, Matthew Wilcox <willy@...radead.org>,
linux-fsdevel@...r.kernel.org,
Dave Wysochanski <dwysocha@...hat.com>,
Marc Dionne <marc.dionne@...istor.com>,
Trond Myklebust <trond.myklebust@...merspace.com>,
Shyam Prasad N <nspmangalore@...il.com>,
Eric Van Hensbergen <ericvh@...il.com>,
v9fs-developer@...ts.sourceforge.net, linux-cifs@...r.kernel.org,
Latchesar Ionkov <lucho@...kov.net>,
Jeff Layton <jlayton@...nel.org>,
Steve French <sfrench@...ba.org>,
Al Viro <viro@...iv.linux.org.uk>,
Dominique Martinet <asmadeus@...ewreck.org>,
Ilya Dryomov <idryomov@...il.com>,
Trond Myklebust <trondmy@...merspace.com>,
Jeff Layton <jlayton@...hat.com>,
Linus Torvalds <torvalds@...ux-foundation.org>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 00/67] fscache: Rewrite index API and management system
On Mon, Oct 18, 2021 at 03:50:15PM +0100, David Howells wrote:
>
> Here's a set of patches that rewrites and simplifies the fscache index API
> to remove the complex operation scheduling and object state machine in
> favour of something much smaller and simpler. It is built on top of the
> set of patches that removes the old API[1].
>
> The operation scheduling API was intended to handle sequencing of cache
> operations, which were all required (where possible) to run asynchronously
> in parallel with the operations being done by the network filesystem, while
> allowing the cache to be brought online and offline and interrupt service
> with invalidation.
>
> However, with the advent of the tmpfile capacity in the VFS, an opportunity
> arises to do invalidation much more easily, without having to wait for I/O
> that's actually in progress: Cachefiles can simply cut over its file
> pointer for the backing object attached to a cookie and abandon the
> in-progress I/O, dismissing it upon completion.
>
> Future work there would involve using Omar Sandoval's vfs_link() with
> AT_LINK_REPLACE[2] to allow an extant file to be displaced by a new hard
> link from a tmpfile as currently I have to unlink the old file first.
I had forgotten about that. It'd be great to finish that someday, but
given the dead-end of the last discussion [1], we might need to hash it
out the next time we can convene in person.
1:https://lore.kernel.org/linux-fsdevel/364531.1579265357@warthog.procyon.org.uk/
Powered by blists - more mailing lists