On Sun, Feb 2, 2020 at 10:18 PM Bernd Schubert <bernd.schubert@xxxxxxxxxxx> wrote: > For network file systems it also common to globally enforce fuse > direct-io to reduce/avoid cache coherency issues - the application > typically doesn't ask for that on its own. And that is where this lock > is badly hurting. Hmm, maybe we should differentiate between > fuse-internal direct-io and application direct-io requests here? I'm not against optionally removing the locking for direct write. Theoretically that should be doable. But we need to be very careful about not breaking any assumptions in the kernel and libfuse code. Obviously this would need to be enabled with a flag (e.g. FOPEN_PARALLEL_WRITES). Thanks, Miklos