md: improve handling of bio with REQ_PREFLUSH in md_flush_request()

If pers->make_request fails in md_flush_request(), the bio is lost. To
fix this, pass back a bool to indicate if the original make_request call
should continue to handle the I/O and instead of assuming the flush logic
will push it to completion.

Convert md_flush_request to return a bool and no longer calls the raid
driver's make_request function.  If the return is true, then the md flush
logic has or will complete the bio and the md make_request call is done.
If false, then the md make_request function needs to keep processing like
it is a normal bio. Let the original call to md_handle_request handle any
need to retry sending the bio to the raid driver's make_request function
should it be needed.

Also mark md_flush_request and the make_request function pointer as
__must_check to issue warnings should these critical return values be
ignored.

Fixes: 2bc13b83 ("md: batch flush requests.")
Cc: stable@vger.kernel.org # # v4.19+
Cc: NeilBrown <neilb@suse.com>
Signed-off-by: David Jeffery <djeffery@redhat.com>
Reviewed-by: Xiao Ni <xni@redhat.com>
Signed-off-by: Song Liu <songliubraving@fb.com>

drivers/md/md-linear.c

@@ -244,10 +244,9 @@ static bool linear_make_request(struct mddev *mddev, struct bio *bio)
 	sector_t start_sector, end_sector, data_offset;
 	sector_t bio_sector = bio->bi_iter.bi_sector;
 
-	if (unlikely(bio->bi_opf & REQ_PREFLUSH)) {
-		md_flush_request(mddev, bio);
+	if (unlikely(bio->bi_opf & REQ_PREFLUSH)
+	    && md_flush_request(mddev, bio))
 		return true;
-	}
 
 	tmp_dev = which_dev(mddev, bio_sector);
 	start_sector = tmp_dev->end_sector - tmp_dev->rdev->sectors;

drivers/md/md-multipath.c

@@ -104,10 +104,9 @@ static bool multipath_make_request(struct mddev *mddev, struct bio * bio)
 	struct multipath_bh * mp_bh;
 	struct multipath_info *multipath;
 
-	if (unlikely(bio->bi_opf & REQ_PREFLUSH)) {
-		md_flush_request(mddev, bio);
+	if (unlikely(bio->bi_opf & REQ_PREFLUSH)
+	    && md_flush_request(mddev, bio))
 		return true;
-	}
 
 	mp_bh = mempool_alloc(&conf->pool, GFP_NOIO);
 

drivers/md/md.c

@@ -550,7 +550,13 @@ static void md_submit_flush_data(struct work_struct *ws)
 	}
 }
 
-void md_flush_request(struct mddev *mddev, struct bio *bio)
+/*
+ * Manages consolidation of flushes and submitting any flushes needed for
+ * a bio with REQ_PREFLUSH.  Returns true if the bio is finished or is
+ * being finished in another context.  Returns false if the flushing is
+ * complete but still needs the I/O portion of the bio to be processed.
+ */
+bool md_flush_request(struct mddev *mddev, struct bio *bio)
 {
 	ktime_t start = ktime_get_boottime();
 	spin_lock_irq(&mddev->lock);
@@ -575,9 +581,10 @@ void md_flush_request(struct mddev *mddev, struct bio *bio)
 			bio_endio(bio);
 		else {
 			bio->bi_opf &= ~REQ_PREFLUSH;
-			mddev->pers->make_request(mddev, bio);
+			return false;
 		}
 	}
+	return true;
 }
 EXPORT_SYMBOL(md_flush_request);
 

drivers/md/md.h

@@ -550,7 +550,7 @@ struct md_personality
 	int level;
 	struct list_head list;
 	struct module *owner;
-	bool (*make_request)(struct mddev *mddev, struct bio *bio);
+	bool __must_check (*make_request)(struct mddev *mddev, struct bio *bio);
 	/*
 	 * start up works that do NOT require md_thread. tasks that
 	 * requires md_thread should go into start()
@@ -703,7 +703,7 @@ extern void md_error(struct mddev *mddev, struct md_rdev *rdev);
 extern void md_finish_reshape(struct mddev *mddev);
 
 extern int mddev_congested(struct mddev *mddev, int bits);
-extern void md_flush_request(struct mddev *mddev, struct bio *bio);
+extern bool __must_check md_flush_request(struct mddev *mddev, struct bio *bio);
 extern void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
 			   sector_t sector, int size, struct page *page);
 extern int md_super_wait(struct mddev *mddev);

drivers/md/raid0.c

@@ -575,10 +575,9 @@ static bool raid0_make_request(struct mddev *mddev, struct bio *bio)
 	unsigned chunk_sects;
 	unsigned sectors;
 
-	if (unlikely(bio->bi_opf & REQ_PREFLUSH)) {
-		md_flush_request(mddev, bio);
+	if (unlikely(bio->bi_opf & REQ_PREFLUSH)
+	    && md_flush_request(mddev, bio))
 		return true;
-	}
 
 	if (unlikely((bio_op(bio) == REQ_OP_DISCARD))) {
 		raid0_handle_discard(mddev, bio);

drivers/md/raid1.c

@@ -1567,10 +1567,9 @@ static bool raid1_make_request(struct mddev *mddev, struct bio *bio)
 {
 	sector_t sectors;
 
-	if (unlikely(bio->bi_opf & REQ_PREFLUSH)) {
-		md_flush_request(mddev, bio);
+	if (unlikely(bio->bi_opf & REQ_PREFLUSH)
+	    && md_flush_request(mddev, bio))
 		return true;
-	}
 
 	/*
 	 * There is a limit to the maximum size, but

drivers/md/raid10.c

@@ -1525,10 +1525,9 @@ static bool raid10_make_request(struct mddev *mddev, struct bio *bio)
 	int chunk_sects = chunk_mask + 1;
 	int sectors = bio_sectors(bio);
 
-	if (unlikely(bio->bi_opf & REQ_PREFLUSH)) {
-		md_flush_request(mddev, bio);
+	if (unlikely(bio->bi_opf & REQ_PREFLUSH)
+	    && md_flush_request(mddev, bio))
 		return true;
-	}
 
 	if (!md_write_start(mddev, bio))
 		return false;

drivers/md/raid5.c

@@ -5592,8 +5592,8 @@ static bool raid5_make_request(struct mddev *mddev, struct bio * bi)
 		if (ret == 0)
 			return true;
 		if (ret == -ENODEV) {
-			md_flush_request(mddev, bi);
-			return true;
+			if (md_flush_request(mddev, bi))
+				return true;
 		}
 		/* ret == -EAGAIN, fallback */
 		/*