On 20/11/20 09:05, Andrew Jones wrote:
So I finally looked closely enough at the dirty-ring stuff to see that patch 2 was always a dumb idea. dirty_ring_create_vm_done() has a comment that says "Switch to dirty ring mode after VM creation but before any of the vcpu creation". I'd argue that that comment would be better served at the log_mode_create_vm_done() call, but that doesn't excuse my sloppiness here. Maybe someday we can add a patch that adds that comment and also tries to use common code for the number of pages calculation for the VM, but not today. Regarding this series, if the other three patches look good, then we can just drop 2/4. 3/4 and 4/4 should still apply cleanly and work.
Yes, the rest is good. Paolo
