On 21 Aug 2018, at 9:46, David Howells wrote:
Should changes to the compression options on a btrfs mount be atomic,
the
problem being that they're split across three variables?
Further to that, how much of an issue is it if the configuration is
split out
into its own struct that is accessed from struct btrfs_fs_info using
RCU?
Hi David,
They shouldn't need to be atomic, we're making compression decisions on
a per-extent basis and sometimes bailing on the compression if it
doesn't make things smaller.
-chris