On Tue, May 05, 2020 at 11:12:27AM -0600, Alex Williamson wrote: > > As noted in the comment, the fault handler can simply do: > > mutex_lock(&vdev->vma_lock); > down_read(&vdev->memory_lock); > > This should be deadlock free now, so we can drop the retry handling That does look like the right direction, because the memory_lock can be done at the very end it means it doesn't need to be nested inside mmap_sem This is much cleaner! Jason
