1 files changed, 117 insertions, 61 deletions

diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index 24a62224b24b..c3a2bc8af675 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -231,7 +231,6 @@ int btrfs_alloc_stripe_hash_table(struct btrfs_fs_info *info)
 		cur = h + i;
 		INIT_LIST_HEAD(&cur->hash_list);
 		spin_lock_init(&cur->lock);
-		init_waitqueue_head(&cur->wait);
 	}
 
 	x = cmpxchg(&info->stripe_hash_table, NULL, table);
@@ -595,14 +594,31 @@ static int rbio_can_merge(struct btrfs_raid_bio *last,
 	 * bio list here, anyone else that wants to
 	 * change this stripe needs to do their own rmw.
 	 */
-	if (last->operation == BTRFS_RBIO_PARITY_SCRUB ||
-	    cur->operation == BTRFS_RBIO_PARITY_SCRUB)
+	if (last->operation == BTRFS_RBIO_PARITY_SCRUB)
 		return 0;
 
-	if (last->operation == BTRFS_RBIO_REBUILD_MISSING ||
-	    cur->operation == BTRFS_RBIO_REBUILD_MISSING)
+	if (last->operation == BTRFS_RBIO_REBUILD_MISSING)
 		return 0;
 
+	if (last->operation == BTRFS_RBIO_READ_REBUILD) {
+		int fa = last->faila;
+		int fb = last->failb;
+		int cur_fa = cur->faila;
+		int cur_fb = cur->failb;
+
+		if (last->faila >= last->failb) {
+			fa = last->failb;
+			fb = last->faila;
+		}
+
+		if (cur->faila >= cur->failb) {
+			cur_fa = cur->failb;
+			cur_fb = cur->faila;
+		}
+
+		if (fa != cur_fa || fb != cur_fb)
+			return 0;
+	}
 	return 1;
 }
 
@@ -670,7 +686,6 @@ static noinline int lock_stripe_add(struct btrfs_raid_bio *rbio)
 	struct btrfs_raid_bio *cur;
 	struct btrfs_raid_bio *pending;
 	unsigned long flags;
-	DEFINE_WAIT(wait);
 	struct btrfs_raid_bio *freeit = NULL;
 	struct btrfs_raid_bio *cache_drop = NULL;
 	int ret = 0;
@@ -816,15 +831,6 @@ static noinline void unlock_stripe(struct btrfs_raid_bio *rbio)
 			}
 
 			goto done_nolock;
-			/*
-			 * The barrier for this waitqueue_active is not needed,
-			 * we're protected by h->lock and can't miss a wakeup.
-			 */
-		} else if (waitqueue_active(&h->wait)) {
-			spin_unlock(&rbio->bio_list_lock);
-			spin_unlock_irqrestore(&h->lock, flags);
-			wake_up(&h->wait);
-			goto done_nolock;
 		}
 	}
 done:
@@ -858,10 +864,17 @@ static void __free_raid_bio(struct btrfs_raid_bio *rbio)
 	kfree(rbio);
 }
 
-static void free_raid_bio(struct btrfs_raid_bio *rbio)
+static void rbio_endio_bio_list(struct bio *cur, blk_status_t err)
 {
-	unlock_stripe(rbio);
-	__free_raid_bio(rbio);
+	struct bio *next;
+
+	while (cur) {
+		next = cur->bi_next;
+		cur->bi_next = NULL;
+		cur->bi_status = err;
+		bio_endio(cur);
+		cur = next;
+	}
 }
 
 /*
@@ -871,20 +884,26 @@ static void free_raid_bio(struct btrfs_raid_bio *rbio)
 static void rbio_orig_end_io(struct btrfs_raid_bio *rbio, blk_status_t err)
 {
 	struct bio *cur = bio_list_get(&rbio->bio_list);
-	struct bio *next;
+	struct bio *extra;
 
 	if (rbio->generic_bio_cnt)
 		btrfs_bio_counter_sub(rbio->fs_info, rbio->generic_bio_cnt);
 
-	free_raid_bio(rbio);
+	/*
+	 * At this moment, rbio->bio_list is empty, however since rbio does not
+	 * always have RBIO_RMW_LOCKED_BIT set and rbio is still linked on the
+	 * hash list, rbio may be merged with others so that rbio->bio_list
+	 * becomes non-empty.
+	 * Once unlock_stripe() is done, rbio->bio_list will not be updated any
+	 * more and we can call bio_endio() on all queued bios.
+	 */
+	unlock_stripe(rbio);
+	extra = bio_list_get(&rbio->bio_list);
+	__free_raid_bio(rbio);
 
-	while (cur) {
-		next = cur->bi_next;
-		cur->bi_next = NULL;
-		cur->bi_status = err;
-		bio_endio(cur);
-		cur = next;
-	}
+	rbio_endio_bio_list(cur, err);
+	if (extra)
+		rbio_endio_bio_list(extra, err);
 }
 
 /*
@@ -1326,6 +1345,9 @@ write_data:
 
 cleanup:
 	rbio_orig_end_io(rbio, BLK_STS_IOERR);
+
+	while ((bio = bio_list_pop(&bio_list)))
+		bio_put(bio);
 }
 
 /*
@@ -1348,6 +1370,7 @@ static int find_bio_stripe(struct btrfs_raid_bio *rbio,
 		stripe_start = stripe->physical;
 		if (physical >= stripe_start &&
 		    physical < stripe_start + rbio->stripe_len &&
+		    stripe->dev->bdev &&
 		    bio->bi_disk == stripe->dev->bdev->bd_disk &&
 		    bio->bi_partno == stripe->dev->bdev->bd_partno) {
 			return i;
@@ -1432,14 +1455,13 @@ static int fail_bio_stripe(struct btrfs_raid_bio *rbio,
  */
 static void set_bio_pages_uptodate(struct bio *bio)
 {
-	struct bio_vec bvec;
-	struct bvec_iter iter;
+	struct bio_vec *bvec;
+	int i;
 
-	if (bio_flagged(bio, BIO_CLONED))
-		bio->bi_iter = btrfs_io_bio(bio)->iter;
+	ASSERT(!bio_flagged(bio, BIO_CLONED));
 
-	bio_for_each_segment(bvec, bio, iter)
-		SetPageUptodate(bvec.bv_page);
+	bio_for_each_segment_all(bvec, bio, i)
+		SetPageUptodate(bvec->bv_page);
 }
 
 /*
@@ -1582,6 +1604,10 @@ static int raid56_rmw_stripe(struct btrfs_raid_bio *rbio)
 
 cleanup:
 	rbio_orig_end_io(rbio, BLK_STS_IOERR);
+
+	while ((bio = bio_list_pop(&bio_list)))
+		bio_put(bio);
+
 	return -EIO;
 
 finish:
@@ -1961,15 +1987,34 @@ cleanup:
 	kfree(pointers);
 
 cleanup_io:
-	if (rbio->operation == BTRFS_RBIO_READ_REBUILD) {
-		if (err == BLK_STS_OK)
+	/*
+	 * Similar to READ_REBUILD, REBUILD_MISSING at this point also has a
+	 * valid rbio which is consistent with ondisk content, thus such a
+	 * valid rbio can be cached to avoid further disk reads.
+	 */
+	if (rbio->operation == BTRFS_RBIO_READ_REBUILD ||
+	    rbio->operation == BTRFS_RBIO_REBUILD_MISSING) {
+		/*
+		 * - In case of two failures, where rbio->failb != -1:
+		 *
+		 *   Do not cache this rbio since the above read reconstruction
+		 *   (raid6_datap_recov() or raid6_2data_recov()) may have
+		 *   changed some content of stripes which are not identical to
+		 *   on-disk content any more, otherwise, a later write/recover
+		 *   may steal stripe_pages from this rbio and end up with
+		 *   corruptions or rebuild failures.
+		 *
+		 * - In case of single failure, where rbio->failb == -1:
+		 *
+		 *   Cache this rbio iff the above read reconstruction is
+		 *   excuted without problems.
+		 */
+		if (err == BLK_STS_OK && rbio->failb < 0)
 			cache_rbio_pages(rbio);
 		else
 			clear_bit(RBIO_CACHE_READY_BIT, &rbio->flags);
 
 		rbio_orig_end_io(rbio, err);
-	} else if (rbio->operation == BTRFS_RBIO_REBUILD_MISSING) {
-		rbio_orig_end_io(rbio, err);
 	} else if (err == BLK_STS_OK) {
 		rbio->faila = -1;
 		rbio->failb = -1;
@@ -2107,6 +2152,10 @@ cleanup:
 	if (rbio->operation == BTRFS_RBIO_READ_REBUILD ||
 	    rbio->operation == BTRFS_RBIO_REBUILD_MISSING)
 		rbio_orig_end_io(rbio, BLK_STS_IOERR);
+
+	while ((bio = bio_list_pop(&bio_list)))
+		bio_put(bio);
+
 	return -EIO;
 }
 
@@ -2159,11 +2208,21 @@ int raid56_parity_recover(struct btrfs_fs_info *fs_info, struct bio *bio,
 	}
 
 	/*
-	 * reconstruct from the q stripe if they are
-	 * asking for mirror 3
+	 * Loop retry:
+	 * for 'mirror == 2', reconstruct from all other stripes.
+	 * for 'mirror_num > 2', select a stripe to fail on every retry.
 	 */
-	if (mirror_num == 3)
-		rbio->failb = rbio->real_stripes - 2;
+	if (mirror_num > 2) {
+		/*
+		 * 'mirror == 3' is to fail the p stripe and
+		 * reconstruct from the q stripe.  'mirror > 3' is to
+		 * fail a data stripe and reconstruct from p+q stripe.
+		 */
+		rbio->failb = rbio->real_stripes - (mirror_num - 1);
+		ASSERT(rbio->failb > 0);
+		if (rbio->failb <= rbio->faila)
+			rbio->failb--;
+	}
 
 	ret = lock_stripe_add(rbio);
 
@@ -2231,12 +2290,18 @@ raid56_parity_alloc_scrub_rbio(struct btrfs_fs_info *fs_info, struct bio *bio,
 	ASSERT(!bio->bi_iter.bi_size);
 	rbio->operation = BTRFS_RBIO_PARITY_SCRUB;
 
-	for (i = 0; i < rbio->real_stripes; i++) {
+	/*
+	 * After mapping bbio with BTRFS_MAP_WRITE, parities have been sorted
+	 * to the end position, so this search can start from the first parity
+	 * stripe.
+	 */
+	for (i = rbio->nr_data; i < rbio->real_stripes; i++) {
 		if (bbio->stripes[i].dev == scrub_dev) {
 			rbio->scrubp = i;
 			break;
 		}
 	}
+	ASSERT(i < rbio->real_stripes);
 
 	/* Now we just support the sectorsize equals to page size */
 	ASSERT(fs_info->sectorsize == PAGE_SIZE);
@@ -2454,6 +2519,9 @@ submit_write:
 
 cleanup:
 	rbio_orig_end_io(rbio, BLK_STS_IOERR);
+
+	while ((bio = bio_list_pop(&bio_list)))
+		bio_put(bio);
 }
 
 static inline int is_data_stripe(struct btrfs_raid_bio *rbio, int stripe)
@@ -2563,12 +2631,12 @@ static void raid56_parity_scrub_stripe(struct btrfs_raid_bio *rbio)
 	int stripe;
 	struct bio *bio;
 
+	bio_list_init(&bio_list);
+
 	ret = alloc_rbio_essential_pages(rbio);
 	if (ret)
 		goto cleanup;
 
-	bio_list_init(&bio_list);
-
 	atomic_set(&rbio->error, 0);
 	/*
 	 * build a list of bios to read all the missing parts of this
@@ -2636,6 +2704,10 @@ static void raid56_parity_scrub_stripe(struct btrfs_raid_bio *rbio)
 
 cleanup:
 	rbio_orig_end_io(rbio, BLK_STS_IOERR);
+
+	while ((bio = bio_list_pop(&bio_list)))
+		bio_put(bio);
+
 	return;
 
 finish:
@@ -2700,24 +2772,8 @@ raid56_alloc_missing_rbio(struct btrfs_fs_info *fs_info, struct bio *bio,
 	return rbio;
 }
 
-static void missing_raid56_work(struct btrfs_work *work)
-{
-	struct btrfs_raid_bio *rbio;
-
-	rbio = container_of(work, struct btrfs_raid_bio, work);
-	__raid56_parity_recover(rbio);
-}
-
-static void async_missing_raid56(struct btrfs_raid_bio *rbio)
-{
-	btrfs_init_work(&rbio->work, btrfs_rmw_helper,
-			missing_raid56_work, NULL, NULL);
-
-	btrfs_queue_work(rbio->fs_info->rmw_workers, &rbio->work);
-}
-
 void raid56_submit_missing_rbio(struct btrfs_raid_bio *rbio)
 {
 	if (!lock_stripe_add(rbio))
-		async_missing_raid56(rbio);
+		async_read_rebuild(rbio);
 }