From: Valerie Aurora <val@xxxxxxxxxxxxxxxxx>
[kamal@xxxxxxxxxxxxx: patch restructure]
[jack@xxxxxxx: updated comments to reflect new situation]
Signed-off-by: Valerie Aurora <val@xxxxxxxxxxxxxxxxx>
Tested-by: Peter M. Petrakis <peter.petrakis@xxxxxxxxxxxxx>
Signed-off-by: Kamal Mostafa <kamal@xxxxxxxxxxxxx>
Signed-off-by: Jan Kara <jack@xxxxxxx>
fs/super.c | 18 ++++++++++++++++++
1 files changed, 18 insertions(+), 0 deletions(-)
diff --git a/fs/super.c b/fs/super.c
index 4aaad7e..557a4ef 100644
@@ -1312,6 +1312,24 @@ static void sb_freeze_to_level(struct super_block *sb,
* Syncs the super to make sure the filesystem is consistent and calls the fs's
* freeze_fs. Subsequent calls to this without first thawing the fs will
+ * During this function, sb->s_frozen goes through these values:
+ * SB_UNFROZEN: File system is normal, all writes progress as usual.
+ * SB_FREEZE_WRITE: The file system is in the process of being frozen.
+ * New writes should be blocked, we wait for in-progress writes (using
+ * s_writers[SB_FREEZE_WRITE-1]) and then sync the file system.
+ * SB_FREEZE_TRANS: The file system is frozen. We wait for all metadata writes
+ * to finish using s_writers[SB_FREEZE_TRANS-1] and after that call ->freeze_fs
+ * to finish filesystem freezing. After ->freeze_fs is done nothing should be
+ * dirty and all modifications are blocked until the file system is thawed.
+ * sb->s_frozen is protected by sb->s_umount. Additionally,
+ * SB_FREEZE_WRITE is only temporarily set during freeze/thaw while
+ * holding sb->s_umount for writing, so any other callers holding
+ * sb->s_umount will never see this state.
int freeze_super(struct super_block *sb)