Thanks for responding. On 2022/07/15 23:25, Mike Snitzer wrote: > I'm not applying such a fundamental change to the DM subsystem when > the header for the change leads with: "Note: This patch was made > blindly and completely untested." Of course, I don't think this patch will be applied as-is. > > I've been busy... when is the deadline for this conversion? I read > commit c4f135d643823a86 but it doesn't say. There is no deadline. I estimate we need about 6 months for getting rid of all in-tree flush_scheduled_work() users. > > I don't think I have the time to make and test such change in time for > 5.20 as I'm working to resolve other issues. That's no problem. This patch was proposed for heads-up purpose. Please clarify which work items does flush_scheduled_work() from local_exit() needs to wait. Depending on the clarification result, we might be able to simply remove flush_scheduled_work() from local_exit() and/or replace with flush_work() instead of introducing dedicated workqueues. -- dm-devel mailing list dm-devel@xxxxxxxxxx https://listman.redhat.com/mailman/listinfo/dm-devel
