md: Runtime support for multiple ppls

Increase PPL area to 1MB and use it as circular buffer to store PPL. The
entry with highest generation number is the latest one. If PPL to be
written is larger then space left in a buffer, rewind the buffer to the
start (don't wrap it).

Signed-off-by: Pawel Baldysiak <pawel.baldysiak@intel.com>
Signed-off-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Signed-off-by: Shaohua Li <shli@fb.com>

drivers/md/md.c

@@ -1536,7 +1536,8 @@ static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_
 	} else if (sb->bblog_offset != 0)
 		rdev->badblocks.shift = 0;
 
-	if (le32_to_cpu(sb->feature_map) & MD_FEATURE_PPL) {
+	if ((le32_to_cpu(sb->feature_map) &
+	    (MD_FEATURE_PPL | MD_FEATURE_MULTIPLE_PPLS))) {
 		rdev->ppl.offset = (__s16)le16_to_cpu(sb->ppl.offset);
 		rdev->ppl.size = le16_to_cpu(sb->ppl.size);
 		rdev->ppl.sector = rdev->sb_start + rdev->ppl.offset;
@@ -1655,10 +1656,15 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev)
 		if (le32_to_cpu(sb->feature_map) & MD_FEATURE_JOURNAL)
 			set_bit(MD_HAS_JOURNAL, &mddev->flags);
 
-		if (le32_to_cpu(sb->feature_map) & MD_FEATURE_PPL) {
+		if (le32_to_cpu(sb->feature_map) &
+		    (MD_FEATURE_PPL | MD_FEATURE_MULTIPLE_PPLS)) {
 			if (le32_to_cpu(sb->feature_map) &
 			    (MD_FEATURE_BITMAP_OFFSET | MD_FEATURE_JOURNAL))
 				return -EINVAL;
+			if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_PPL) &&
+			    (le32_to_cpu(sb->feature_map) &
+					    MD_FEATURE_MULTIPLE_PPLS))
+				return -EINVAL;
 			set_bit(MD_HAS_PPL, &mddev->flags);
 		}
 	} else if (mddev->pers == NULL) {
@@ -1875,7 +1881,11 @@ retry:
 		sb->feature_map |= cpu_to_le32(MD_FEATURE_JOURNAL);
 
 	if (test_bit(MD_HAS_PPL, &mddev->flags)) {
-		sb->feature_map |= cpu_to_le32(MD_FEATURE_PPL);
+		if (test_bit(MD_HAS_MULTIPLE_PPLS, &mddev->flags))
+			sb->feature_map |=
+			    cpu_to_le32(MD_FEATURE_MULTIPLE_PPLS);
+		else
+			sb->feature_map |= cpu_to_le32(MD_FEATURE_PPL);
 		sb->ppl.offset = cpu_to_le16(rdev->ppl.offset);
 		sb->ppl.size = cpu_to_le16(rdev->ppl.size);
 	}

drivers/md/md.h

@@ -236,6 +236,7 @@ enum mddev_flags {
 				 * never cause the array to become failed.
 				 */
 	MD_HAS_PPL,		/* The raid array has PPL feature set */
+	MD_HAS_MULTIPLE_PPLS,	/* The raid array has multiple PPLs feature set */
 };
 
 enum mddev_sb_flags {

drivers/md/raid0.c

@@ -30,7 +30,8 @@
 	((1L << MD_HAS_JOURNAL) |	\
 	 (1L << MD_JOURNAL_CLEAN) |	\
 	 (1L << MD_FAILFAST_SUPPORTED) |\
-	 (1L << MD_HAS_PPL))
+	 (1L << MD_HAS_PPL) |		\
+	 (1L << MD_HAS_MULTIPLE_PPLS))
 
 static int raid0_congested(struct mddev *mddev, int bits)
 {

drivers/md/raid1.c

@@ -48,7 +48,8 @@
 #define UNSUPPORTED_MDDEV_FLAGS		\
 	((1L << MD_HAS_JOURNAL) |	\
 	 (1L << MD_JOURNAL_CLEAN) |	\
-	 (1L << MD_HAS_PPL))
+	 (1L << MD_HAS_PPL) |		\
+	 (1L << MD_HAS_MULTIPLE_PPLS))
 
 /*
  * Number of guaranteed r1bios in case of extreme VM load:

drivers/md/raid5-ppl.c

@@ -87,6 +87,8 @@
  * The current io_unit accepting new stripes is always at the end of the list.
  */
 
+#define PPL_SPACE_SIZE (128 * 1024)
+
 struct ppl_conf {
 	struct mddev *mddev;
 
@@ -122,6 +124,10 @@ struct ppl_log {
 					 * always at the end of io_list */
 	spinlock_t io_list_lock;
 	struct list_head io_list;	/* all io_units of this log */
+
+	sector_t next_io_sector;
+	unsigned int entry_space;
+	bool use_multippl;
 };
 
 #define PPL_IO_INLINE_BVECS 32
@@ -264,13 +270,12 @@ static int ppl_log_stripe(struct ppl_log *log, struct stripe_head *sh)
 	int i;
 	sector_t data_sector = 0;
 	int data_disks = 0;
-	unsigned int entry_space = (log->rdev->ppl.size << 9) - PPL_HEADER_SIZE;
 	struct r5conf *conf = sh->raid_conf;
 
 	pr_debug("%s: stripe: %llu\n", __func__, (unsigned long long)sh->sector);
 
 	/* check if current io_unit is full */
-	if (io && (io->pp_size == entry_space ||
+	if (io && (io->pp_size == log->entry_space ||
 		   io->entries_count == PPL_HDR_MAX_ENTRIES)) {
 		pr_debug("%s: add io_unit blocked by seq: %llu\n",
 			 __func__, io->seq);
@@ -451,12 +456,25 @@ static void ppl_submit_iounit(struct ppl_io_unit *io)
 	pplhdr->entries_count = cpu_to_le32(io->entries_count);
 	pplhdr->checksum = cpu_to_le32(~crc32c_le(~0, pplhdr, PPL_HEADER_SIZE));
 
+	/* Rewind the buffer if current PPL is larger then remaining space */
+	if (log->use_multippl &&
+	    log->rdev->ppl.sector + log->rdev->ppl.size - log->next_io_sector <
+	    (PPL_HEADER_SIZE + io->pp_size) >> 9)
+		log->next_io_sector = log->rdev->ppl.sector;
+
+
 	bio->bi_end_io = ppl_log_endio;
 	bio->bi_opf = REQ_OP_WRITE | REQ_FUA;
 	bio->bi_bdev = log->rdev->bdev;
-	bio->bi_iter.bi_sector = log->rdev->ppl.sector;
+	bio->bi_iter.bi_sector = log->next_io_sector;
 	bio_add_page(bio, io->header_page, PAGE_SIZE, 0);
 
+	pr_debug("%s: log->current_io_sector: %llu\n", __func__,
+	    (unsigned long long)log->next_io_sector);
+
+	if (log->use_multippl)
+		log->next_io_sector += (PPL_HEADER_SIZE + io->pp_size) >> 9;
+
 	list_for_each_entry(sh, &io->stripe_list, log_list) {
 		/* entries for full stripe writes have no partial parity */
 		if (test_bit(STRIPE_FULL_WRITE, &sh->state))
@@ -1031,6 +1049,7 @@ static int ppl_load(struct ppl_conf *ppl_conf)
 static void __ppl_exit_log(struct ppl_conf *ppl_conf)
 {
 	clear_bit(MD_HAS_PPL, &ppl_conf->mddev->flags);
+	clear_bit(MD_HAS_MULTIPLE_PPLS, &ppl_conf->mddev->flags);
 
 	kfree(ppl_conf->child_logs);
 
@@ -1099,6 +1118,22 @@ static int ppl_validate_rdev(struct md_rdev *rdev)
 	return 0;
 }
 
+static void ppl_init_child_log(struct ppl_log *log, struct md_rdev *rdev)
+{
+	if ((rdev->ppl.size << 9) >= (PPL_SPACE_SIZE +
+				      PPL_HEADER_SIZE) * 2) {
+		log->use_multippl = true;
+		set_bit(MD_HAS_MULTIPLE_PPLS,
+			&log->ppl_conf->mddev->flags);
+		log->entry_space = PPL_SPACE_SIZE;
+	} else {
+		log->use_multippl = false;
+		log->entry_space = (log->rdev->ppl.size << 9) -
+				   PPL_HEADER_SIZE;
+	}
+	log->next_io_sector = rdev->ppl.sector;
+}
+
 int ppl_init_log(struct r5conf *conf)
 {
 	struct ppl_conf *ppl_conf;
@@ -1196,6 +1231,7 @@ int ppl_init_log(struct r5conf *conf)
 			q = bdev_get_queue(rdev->bdev);
 			if (test_bit(QUEUE_FLAG_WC, &q->queue_flags))
 				need_cache_flush = true;
+			ppl_init_child_log(log, rdev);
 		}
 	}
 
@@ -1261,6 +1297,7 @@ int ppl_modify_log(struct r5conf *conf, struct md_rdev *rdev, bool add)
 		if (!ret) {
 			log->rdev = rdev;
 			ret = ppl_write_empty_header(log);
+			ppl_init_child_log(log, rdev);
 		}
 	} else {
 		log->rdev = NULL;

drivers/md/raid5.c

@@ -7236,6 +7236,7 @@ static int raid5_run(struct mddev *mddev)
 		pr_warn("md/raid:%s: using journal device and PPL not allowed - disabling PPL\n",
 			mdname(mddev));
 		clear_bit(MD_HAS_PPL, &mddev->flags);
+		clear_bit(MD_HAS_MULTIPLE_PPLS, &mddev->flags);
 	}
 
 	if (mddev->private == NULL)

include/uapi/linux/raid/md_p.h

@@ -324,9 +324,10 @@ struct mdp_superblock_1 {
 #define	MD_FEATURE_RECOVERY_BITMAP	128 /* recovery that is happening
 					     * is guided by bitmap.
 					     */
-#define MD_FEATURE_CLUSTERED		256 /* clustered MD */
+#define	MD_FEATURE_CLUSTERED		256 /* clustered MD */
 #define	MD_FEATURE_JOURNAL		512 /* support write cache */
 #define	MD_FEATURE_PPL			1024 /* support PPL */
+#define	MD_FEATURE_MULTIPLE_PPLS	2048 /* support for multiple PPLs */
 #define	MD_FEATURE_ALL			(MD_FEATURE_BITMAP_OFFSET	\
 					|MD_FEATURE_RECOVERY_OFFSET	\
 					|MD_FEATURE_RESHAPE_ACTIVE	\
@@ -338,6 +339,7 @@ struct mdp_superblock_1 {
 					|MD_FEATURE_CLUSTERED		\
 					|MD_FEATURE_JOURNAL		\
 					|MD_FEATURE_PPL			\
+					|MD_FEATURE_MULTIPLE_PPLS	\
 					)
 
 struct r5l_payload_header {