[PATCH 2/2] f2fs: fix to wake up all sleeping flusher

[Chao Yu]

In scenario of remount_ro vs flush, after flush_thread exits in
->remount_fs, flusher will only clean up golbal issue_list, but
without waking up flushers waiting on that list, result in hang
related user threads.

In order to fix this issue, this patch enables the flusher to
take charge of issue_flush thread: executes merged flush command,
and wake up all sleeping flushers.

Fixes: 5eba8c5d1fb3 ("f2fs: fix to access nullified flush_cmd_control pointer")
Signed-off-by: Chao Yu <yuchao0@xxxxxxxxxx>
---
 fs/f2fs/segment.c | 23 +++++++++++++++++++++--
 1 file changed, 21 insertions(+), 2 deletions(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 1215ca1bd4e2..265c3bc44f2d 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -558,8 +558,27 @@ int f2fs_issue_flush(struct f2fs_sb_info *sbi)
 		wait_for_completion(&cmd.wait);
 		atomic_dec(&fcc->issing_flush);
 	} else {
-		llist_del_all(&fcc->issue_list);
-		atomic_set(&fcc->issing_flush, 0);
+		struct llist_node *list;
+
+		list = llist_del_all(&fcc->issue_list);
+		if (!list) {
+			wait_for_completion(&cmd.wait);
+			atomic_dec(&fcc->issing_flush);
+		} else {
+			struct flush_cmd *tmp, *next;
+
+			ret = submit_flush_wait(sbi);
+
+			llist_for_each_entry_safe(tmp, next, list, llnode) {
+				if (tmp == &cmd) {
+					cmd.ret = ret;
+					atomic_dec(&fcc->issing_flush);
+					continue;
+				}
+				tmp->ret = ret;
+				complete(&tmp->wait);
+			}
+		}
 	}
 
 	return cmd.ret;
-- 
2.13.1.388.g69e6b9b4f4a9