md: add a new recovery_flag MD_RECOVERY_LAZY_RECOVER

This flag is used by llbitmap in later patches to skip raid456 initial
recover and delay building initial xor data to first write.

https://lore.kernel.org/linux-raid/20250829080426.1441678-10-yukuai1@huaweicloud.com
Signed-off-by: Yu Kuai <yukuai3@huawei.com>

drivers/md/md.c

@@ -9199,6 +9199,39 @@ static sector_t md_sync_max_sectors(struct mddev *mddev,
 	}
 }
 
+/*
+ * If lazy recovery is requested and all rdevs are in sync, select the rdev with
+ * the higest index to perfore recovery to build initial xor data, this is the
+ * same as old bitmap.
+ */
+static bool mddev_select_lazy_recover_rdev(struct mddev *mddev)
+{
+	struct md_rdev *recover_rdev = NULL;
+	struct md_rdev *rdev;
+	bool ret = false;
+
+	rcu_read_lock();
+	rdev_for_each_rcu(rdev, mddev) {
+		if (rdev->raid_disk < 0)
+			continue;
+
+		if (test_bit(Faulty, &rdev->flags) ||
+		    !test_bit(In_sync, &rdev->flags))
+			break;
+
+		if (!recover_rdev || recover_rdev->raid_disk < rdev->raid_disk)
+			recover_rdev = rdev;
+	}
+
+	if (recover_rdev) {
+		clear_bit(In_sync, &recover_rdev->flags);
+		ret = true;
+	}
+
+	rcu_read_unlock();
+	return ret;
+}
+
 static sector_t md_sync_position(struct mddev *mddev, enum sync_action action)
 {
 	sector_t start = 0;
@@ -9230,6 +9263,14 @@ static sector_t md_sync_position(struct mddev *mddev, enum sync_action action)
 				start = rdev->recovery_offset;
 		rcu_read_unlock();
 
+		/*
+		 * If there are no spares, and raid456 lazy initial recover is
+		 * requested.
+		 */
+		if (test_bit(MD_RECOVERY_LAZY_RECOVER, &mddev->recovery) &&
+		    start == MaxSector && mddev_select_lazy_recover_rdev(mddev))
+			start = 0;
+
 		/* If there is a bitmap, we need to make sure all
 		 * writes that started before we added a spare
 		 * complete before we start doing a recovery.
@@ -9791,6 +9832,7 @@ static bool md_choose_sync_action(struct mddev *mddev, int *spares)
 
 		set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
 		clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
+		clear_bit(MD_RECOVERY_LAZY_RECOVER, &mddev->recovery);
 		return true;
 	}
 
@@ -9799,6 +9841,7 @@ static bool md_choose_sync_action(struct mddev *mddev, int *spares)
 		remove_spares(mddev, NULL);
 		set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
 		clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
+		clear_bit(MD_RECOVERY_LAZY_RECOVER, &mddev->recovery);
 		return true;
 	}
 
@@ -9808,7 +9851,7 @@ static bool md_choose_sync_action(struct mddev *mddev, int *spares)
 	 * re-add.
 	 */
 	*spares = remove_and_add_spares(mddev, NULL);
-	if (*spares) {
+	if (*spares || test_bit(MD_RECOVERY_LAZY_RECOVER, &mddev->recovery)) {
 		clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
 		clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
 		clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
@@ -10021,6 +10064,7 @@ void md_check_recovery(struct mddev *mddev)
 			}
 
 			clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
+			clear_bit(MD_RECOVERY_LAZY_RECOVER, &mddev->recovery);
 			clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 			clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags);
 
@@ -10131,6 +10175,7 @@ void md_reap_sync_thread(struct mddev *mddev)
 	clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
 	clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
 	clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
+	clear_bit(MD_RECOVERY_LAZY_RECOVER, &mddev->recovery);
 	/*
 	 * We call mddev->cluster_ops->update_size here because sync_size could
 	 * be changed by md_update_sb, and MD_RECOVERY_RESHAPE is cleared,

drivers/md/md.h

@@ -667,6 +667,8 @@ enum recovery_flags {
 	MD_RECOVERY_RESHAPE,
 	/* remote node is running resync thread */
 	MD_RESYNCING_REMOTE,
+	/* raid456 lazy initial recover */
+	MD_RECOVERY_LAZY_RECOVER,
 };
 
 enum md_ro_state {

drivers/md/raid5.c

@@ -4705,10 +4705,21 @@ static void analyse_stripe(struct stripe_head *sh, struct stripe_head_state *s)
 			}
 		} else if (test_bit(In_sync, &rdev->flags))
 			set_bit(R5_Insync, &dev->flags);
-		else if (sh->sector + RAID5_STRIPE_SECTORS(conf) <= rdev->recovery_offset)
-			/* in sync if before recovery_offset */
-			set_bit(R5_Insync, &dev->flags);
-		else if (test_bit(R5_UPTODATE, &dev->flags) &&
+		else if (sh->sector + RAID5_STRIPE_SECTORS(conf) <=
+			 rdev->recovery_offset) {
+			/*
+			 * in sync if:
+			 *  - normal IO, or
+			 *  - resync IO that is not lazy recovery
+			 *
+			 * For lazy recovery, we have to mark the rdev without
+			 * In_sync as failed, to build initial xor data.
+			 */
+			if (!test_bit(STRIPE_SYNCING, &sh->state) ||
+			    !test_bit(MD_RECOVERY_LAZY_RECOVER,
+				      &conf->mddev->recovery))
+				set_bit(R5_Insync, &dev->flags);
+		} else if (test_bit(R5_UPTODATE, &dev->flags) &&
 			 test_bit(R5_Expanded, &dev->flags))
 			/* If we've reshaped into here, we assume it is Insync.
 			 * We will shortly update recovery_offset to make