1 files changed, 113 insertions, 58 deletions

diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index aa5290cb7467..871ec35ea8e8 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -136,7 +136,7 @@ static void nilfs_dispose_list(struct the_nilfs *, struct list_head *, int);
 
 #define nilfs_cnt32_ge(a, b)   \
 	(typecheck(__u32, a) && typecheck(__u32, b) && \
-	 ((__s32)(a) - (__s32)(b) >= 0))
+	 ((__s32)((a) - (b)) >= 0))
 
 static int nilfs_prepare_segment_lock(struct super_block *sb,
 				      struct nilfs_transaction_info *ti)
@@ -1639,39 +1639,30 @@ static void nilfs_begin_folio_io(struct folio *folio)
 	folio_unlock(folio);
 }
 
-static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci)
+/**
+ * nilfs_prepare_write_logs - prepare to write logs
+ * @logs: logs to prepare for writing
+ * @seed: checksum seed value
+ *
+ * nilfs_prepare_write_logs() adds checksums and prepares the block
+ * buffers/folios for writing logs.  In order to stabilize folios of
+ * memory-mapped file blocks by putting them in writeback state before
+ * calculating the checksums, first prepare to write payload blocks other
+ * than segment summary and super root blocks in which the checksums will
+ * be embedded.
+ */
+static void nilfs_prepare_write_logs(struct list_head *logs, u32 seed)
 {
 	struct nilfs_segment_buffer *segbuf;
 	struct folio *bd_folio = NULL, *fs_folio = NULL;
+	struct buffer_head *bh;
 
-	list_for_each_entry(segbuf, &sci->sc_segbufs, sb_list) {
-		struct buffer_head *bh;
-
-		list_for_each_entry(bh, &segbuf->sb_segsum_buffers,
-				    b_assoc_buffers) {
-			if (bh->b_folio != bd_folio) {
-				if (bd_folio) {
-					folio_lock(bd_folio);
-					folio_clear_dirty_for_io(bd_folio);
-					folio_start_writeback(bd_folio);
-					folio_unlock(bd_folio);
-				}
-				bd_folio = bh->b_folio;
-			}
-		}
-
+	/* Prepare to write payload blocks */
+	list_for_each_entry(segbuf, logs, sb_list) {
 		list_for_each_entry(bh, &segbuf->sb_payload_buffers,
 				    b_assoc_buffers) {
-			if (bh == segbuf->sb_super_root) {
-				if (bh->b_folio != bd_folio) {
-					folio_lock(bd_folio);
-					folio_clear_dirty_for_io(bd_folio);
-					folio_start_writeback(bd_folio);
-					folio_unlock(bd_folio);
-					bd_folio = bh->b_folio;
-				}
+			if (bh == segbuf->sb_super_root)
 				break;
-			}
 			set_buffer_async_write(bh);
 			if (bh->b_folio != fs_folio) {
 				nilfs_begin_folio_io(fs_folio);
@@ -1679,13 +1670,49 @@ static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci)
 			}
 		}
 	}
+	nilfs_begin_folio_io(fs_folio);
+
+	nilfs_add_checksums_on_logs(logs, seed);
+
+	/* Prepare to write segment summary blocks */
+	list_for_each_entry(segbuf, logs, sb_list) {
+		list_for_each_entry(bh, &segbuf->sb_segsum_buffers,
+				    b_assoc_buffers) {
+			mark_buffer_dirty(bh);
+			if (bh->b_folio == bd_folio)
+				continue;
+			if (bd_folio) {
+				folio_lock(bd_folio);
+				folio_wait_writeback(bd_folio);
+				folio_clear_dirty_for_io(bd_folio);
+				folio_start_writeback(bd_folio);
+				folio_unlock(bd_folio);
+			}
+			bd_folio = bh->b_folio;
+		}
+	}
+
+	/* Prepare to write super root block */
+	bh = NILFS_LAST_SEGBUF(logs)->sb_super_root;
+	if (bh) {
+		mark_buffer_dirty(bh);
+		if (bh->b_folio != bd_folio) {
+			folio_lock(bd_folio);
+			folio_wait_writeback(bd_folio);
+			folio_clear_dirty_for_io(bd_folio);
+			folio_start_writeback(bd_folio);
+			folio_unlock(bd_folio);
+			bd_folio = bh->b_folio;
+		}
+	}
+
 	if (bd_folio) {
 		folio_lock(bd_folio);
+		folio_wait_writeback(bd_folio);
 		folio_clear_dirty_for_io(bd_folio);
 		folio_start_writeback(bd_folio);
 		folio_unlock(bd_folio);
 	}
-	nilfs_begin_folio_io(fs_folio);
 }
 
 static int nilfs_segctor_write(struct nilfs_sc_info *sci,
@@ -1725,14 +1752,8 @@ static void nilfs_end_folio_io(struct folio *folio, int err)
 		return;
 	}
 
-	if (!err) {
-		if (!nilfs_folio_buffers_clean(folio))
-			filemap_dirty_folio(folio->mapping, folio);
-		folio_clear_error(folio);
-	} else {
+	if (err || !nilfs_folio_buffers_clean(folio))
 		filemap_dirty_folio(folio->mapping, folio);
-		folio_set_error(folio);
-	}
 
 	folio_end_writeback(folio);
 }
@@ -1791,6 +1812,9 @@ static void nilfs_segctor_abort_construction(struct nilfs_sc_info *sci,
 	nilfs_abort_logs(&logs, ret ? : err);
 
 	list_splice_tail_init(&sci->sc_segbufs, &logs);
+	if (list_empty(&logs))
+		return; /* if the first segment buffer preparation failed */
+
 	nilfs_cancel_segusage(&logs, nilfs->ns_sufile);
 	nilfs_free_incomplete_logs(&logs, nilfs);
 
@@ -2035,7 +2059,7 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 
 		err = nilfs_segctor_begin_construction(sci, nilfs);
 		if (unlikely(err))
-			goto out;
+			goto failed;
 
 		/* Update time stamp */
 		sci->sc_seg_ctime = ktime_get_real_seconds();
@@ -2073,10 +2097,7 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 		nilfs_segctor_update_segusage(sci, nilfs->ns_sufile);
 
 		/* Write partial segments */
-		nilfs_segctor_prepare_write(sci);
-
-		nilfs_add_checksums_on_logs(&sci->sc_segbufs,
-					    nilfs->ns_crc_seed);
+		nilfs_prepare_write_logs(&sci->sc_segbufs, nilfs->ns_crc_seed);
 
 		err = nilfs_segctor_write(sci, nilfs);
 		if (unlikely(err))
@@ -2102,10 +2123,9 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 	return err;
 
  failed_to_write:
-	if (sci->sc_stage.flags & NILFS_CF_IFILE_STARTED)
-		nilfs_redirty_inodes(&sci->sc_dirty_files);
-
  failed:
+	if (mode == SC_LSEG_SR && nilfs_sc_cstage_get(sci) >= NILFS_ST_IFILE)
+		nilfs_redirty_inodes(&sci->sc_dirty_files);
 	if (nilfs_doing_gc())
 		nilfs_redirty_inodes(&sci->sc_gc_inodes);
 	nilfs_segctor_abort_construction(sci, nilfs, err);
@@ -2124,8 +2144,10 @@ static void nilfs_segctor_start_timer(struct nilfs_sc_info *sci)
 {
 	spin_lock(&sci->sc_state_lock);
 	if (!(sci->sc_state & NILFS_SEGCTOR_COMMIT)) {
-		sci->sc_timer.expires = jiffies + sci->sc_interval;
-		add_timer(&sci->sc_timer);
+		if (sci->sc_task) {
+			sci->sc_timer.expires = jiffies + sci->sc_interval;
+			add_timer(&sci->sc_timer);
+		}
 		sci->sc_state |= NILFS_SEGCTOR_COMMIT;
 	}
 	spin_unlock(&sci->sc_state_lock);
@@ -2172,19 +2194,36 @@ static int nilfs_segctor_sync(struct nilfs_sc_info *sci)
 	struct nilfs_segctor_wait_request wait_req;
 	int err = 0;
 
-	spin_lock(&sci->sc_state_lock);
 	init_wait(&wait_req.wq);
 	wait_req.err = 0;
 	atomic_set(&wait_req.done, 0);
+	init_waitqueue_entry(&wait_req.wq, current);
+
+	/*
+	 * To prevent a race issue where completion notifications from the
+	 * log writer thread are missed, increment the request sequence count
+	 * "sc_seq_request" and insert a wait queue entry using the current
+	 * sequence number into the "sc_wait_request" queue at the same time
+	 * within the lock section of "sc_state_lock".
+	 */
+	spin_lock(&sci->sc_state_lock);
 	wait_req.seq = ++sci->sc_seq_request;
+	add_wait_queue(&sci->sc_wait_request, &wait_req.wq);
 	spin_unlock(&sci->sc_state_lock);
 
-	init_waitqueue_entry(&wait_req.wq, current);
-	add_wait_queue(&sci->sc_wait_request, &wait_req.wq);
-	set_current_state(TASK_INTERRUPTIBLE);
 	wake_up(&sci->sc_wait_daemon);
 
 	for (;;) {
+		set_current_state(TASK_INTERRUPTIBLE);
+
+		/*
+		 * Synchronize only while the log writer thread is alive.
+		 * Leave flushing out after the log writer thread exits to
+		 * the cleanup work in nilfs_segctor_destroy().
+		 */
+		if (!sci->sc_task)
+			break;
+
 		if (atomic_read(&wait_req.done)) {
 			err = wait_req.err;
 			break;
@@ -2200,7 +2239,7 @@ static int nilfs_segctor_sync(struct nilfs_sc_info *sci)
 	return err;
 }
 
-static void nilfs_segctor_wakeup(struct nilfs_sc_info *sci, int err)
+static void nilfs_segctor_wakeup(struct nilfs_sc_info *sci, int err, bool force)
 {
 	struct nilfs_segctor_wait_request *wrq, *n;
 	unsigned long flags;
@@ -2208,7 +2247,7 @@ static void nilfs_segctor_wakeup(struct nilfs_sc_info *sci, int err)
 	spin_lock_irqsave(&sci->sc_wait_request.lock, flags);
 	list_for_each_entry_safe(wrq, n, &sci->sc_wait_request.head, wq.entry) {
 		if (!atomic_read(&wrq->done) &&
-		    nilfs_cnt32_ge(sci->sc_seq_done, wrq->seq)) {
+		    (force || nilfs_cnt32_ge(sci->sc_seq_done, wrq->seq))) {
 			wrq->err = err;
 			atomic_set(&wrq->done, 1);
 		}
@@ -2326,10 +2365,21 @@ int nilfs_construct_dsync_segment(struct super_block *sb, struct inode *inode,
  */
 static void nilfs_segctor_accept(struct nilfs_sc_info *sci)
 {
+	bool thread_is_alive;
+
 	spin_lock(&sci->sc_state_lock);
 	sci->sc_seq_accepted = sci->sc_seq_request;
+	thread_is_alive = (bool)sci->sc_task;
 	spin_unlock(&sci->sc_state_lock);
-	del_timer_sync(&sci->sc_timer);
+
+	/*
+	 * This function does not race with the log writer thread's
+	 * termination.  Therefore, deleting sc_timer, which should not be
+	 * done after the log writer thread exits, can be done safely outside
+	 * the area protected by sc_state_lock.
+	 */
+	if (thread_is_alive)
+		del_timer_sync(&sci->sc_timer);
 }
 
 /**
@@ -2346,7 +2396,7 @@ static void nilfs_segctor_notify(struct nilfs_sc_info *sci, int mode, int err)
 	if (mode == SC_LSEG_SR) {
 		sci->sc_state &= ~NILFS_SEGCTOR_COMMIT;
 		sci->sc_seq_done = sci->sc_seq_accepted;
-		nilfs_segctor_wakeup(sci, err);
+		nilfs_segctor_wakeup(sci, err, false);
 		sci->sc_flush_request = 0;
 	} else {
 		if (mode == SC_FLUSH_FILE)
@@ -2355,7 +2405,7 @@ static void nilfs_segctor_notify(struct nilfs_sc_info *sci, int mode, int err)
 			sci->sc_flush_request &= ~FLUSH_DAT_BIT;
 
 		/* re-enable timer if checkpoint creation was not done */
-		if ((sci->sc_state & NILFS_SEGCTOR_COMMIT) &&
+		if ((sci->sc_state & NILFS_SEGCTOR_COMMIT) && sci->sc_task &&
 		    time_before(jiffies, sci->sc_timer.expires))
 			add_timer(&sci->sc_timer);
 	}
@@ -2545,6 +2595,7 @@ static int nilfs_segctor_thread(void *arg)
 	int timeout = 0;
 
 	sci->sc_timer_task = current;
+	timer_setup(&sci->sc_timer, nilfs_construction_timeout, 0);
 
 	/* start sync. */
 	sci->sc_task = current;
@@ -2612,6 +2663,7 @@ static int nilfs_segctor_thread(void *arg)
  end_thread:
 	/* end sync. */
 	sci->sc_task = NULL;
+	timer_shutdown_sync(&sci->sc_timer);
 	wake_up(&sci->sc_wait_task); /* for nilfs_segctor_kill_thread() */
 	spin_unlock(&sci->sc_state_lock);
 	return 0;
@@ -2675,7 +2727,6 @@ static struct nilfs_sc_info *nilfs_segctor_new(struct super_block *sb,
 	INIT_LIST_HEAD(&sci->sc_gc_inodes);
 	INIT_LIST_HEAD(&sci->sc_iput_queue);
 	INIT_WORK(&sci->sc_iput_work, nilfs_iput_work_func);
-	timer_setup(&sci->sc_timer, nilfs_construction_timeout, 0);
 
 	sci->sc_interval = HZ * NILFS_SC_DEFAULT_TIMEOUT;
 	sci->sc_mjcp_freq = HZ * NILFS_SC_DEFAULT_SR_FREQ;
@@ -2729,6 +2780,13 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci)
 		|| sci->sc_seq_request != sci->sc_seq_done);
 	spin_unlock(&sci->sc_state_lock);
 
+	/*
+	 * Forcibly wake up tasks waiting in nilfs_segctor_sync(), which can
+	 * be called from delayed iput() via nilfs_evict_inode() and can race
+	 * with the above log writer thread termination.
+	 */
+	nilfs_segctor_wakeup(sci, 0, true);
+
 	if (flush_work(&sci->sc_iput_work))
 		flag = true;
 
@@ -2754,7 +2812,6 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci)
 
 	down_write(&nilfs->ns_segctor_sem);
 
-	timer_shutdown_sync(&sci->sc_timer);
 	kfree(sci);
 }
 
@@ -2790,8 +2847,6 @@ int nilfs_attach_log_writer(struct super_block *sb, struct nilfs_root *root)
 	if (!nilfs->ns_writer)
 		return -ENOMEM;
 
-	inode_attach_wb(nilfs->ns_bdev->bd_inode, NULL);
-
 	err = nilfs_segctor_start_thread(nilfs->ns_writer);
 	if (unlikely(err))
 		nilfs_detach_log_writer(sb);