On Fri, May 14, 2021 at 05:27:42PM +0200, David Sterba wrote:

On Fri, May 14, 2021 at 11:03:08AM +0900, Naohiro Aota wrote:

quoted

As btrfs_set_free_space_cache_v1_active() is introduced, this patch uses
it to disable space cache v1 properly.

Can you please describe what problem is it fixing, of if it's a problem
at all? The two functions do have quite different effects, resetting
generation is simple but the new function starts transaction and
iterates over all space inodes. That's beyond obvious so this needs an
explanation. Thanks.

Sure. I'll rework the commit message to tell we want "cache_generation

0 iff the file system is using space_cache v1."

Thanks,