Steve French [off-list ref] wrote:

On Thu, Oct 21, 2021 at 5:21 PM Omar Sandoval [off-list ref] wrote:

quoted

On Mon, Oct 18, 2021 at 03:50:15PM +0100, David Howells wrote:
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.

Have changes been made to O_TMPFILE?  It is problematic for network
filesystems because it is not an atomic operation, and would be great if it
were possible to create a tmpfile and open it atomically (at the file system
level).

In this case, it's nothing to do with the network filesystem that's using the
cache per se.  Cachefiles is using tmpfiles on the backing filesystem, so as
long as that's, say, ext4, xfs or btrfs, it should work fine.  The backing
filesystem also needs to support SEEK_HOLE and SEEK_DATA.

I'm not sure I'd recommend putting your cache on a network filesystem.

David