It becomes possible to do discards after a journal flush, which naturally the journal code is reponsible for.
A prior refactoring seems to have broken this - which went unnoticed because the foreground allocator path can also trigger discards. Signed-off-by: Kent Overstreet <[email protected]> --- fs/bcachefs/journal_io.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c index dd3f3434c1b0..9e028dbcc3d0 100644 --- a/fs/bcachefs/journal_io.c +++ b/fs/bcachefs/journal_io.c @@ -1767,6 +1767,7 @@ static CLOSURE_CALLBACK(journal_write_done) closure_wake_up(&c->freelist_wait); bch2_reset_alloc_cursors(c); + do_discards = true; } j->seq_ondisk = seq; -- 2.50.0
