On Wed, Jan 03, 2024 at 11:32:17PM -0800, Darrick J. Wong wrote: > > Any reason this doesn't actually remove the page from shmem? > > I think I skipped the shmem_truncate_range call because the next btree > block allocation will re-use the page immediately. Maybe add a comment explaining that? Note that shmemfs pages once dirties will keep space allocated for them in memory/swap until explicitly punched out.
