Re: [v3 PATCH] mm: move_pages: return valid node id in status if the page is already on the target node

On Dec 5, 2019, at 7:04 PM, John Hubbard jhubbard@nvidia.com wrote:

Felix's code is not random test code. It's code he wrote and he expected it to work.

Sure, but could he show a bit detail if the kernel will start to behavior as expected like what was written in the manpage to return ENOENT in this case, why is it not acceptable? i.e., why is it important to depend on the broken behavior?