lightnvm: pblk: support packed metadata

pblk performs recovery of open lines by storing the LBA in the per LBA
metadata field. Recovery therefore only works for drives that has this
field.

This patch adds support for packed metadata, which store l2p mapping
for open lines in last sector of every write unit and enables drives
without per IO metadata to recover open lines.

After this patch, drives with OOB size <16B will use packed metadata
and metadata size larger than16B will continue to use the device per
IO metadata.
Reviewed-by: Javier González <javier@cnexlabs.com>
Signed-off-by: Igor Konopko <igor.j.konopko@intel.com>
Signed-off-by: Matias Bjørling <mb@lightnvm.io>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

drivers/lightnvm/pblk-core.c

@@ -376,7 +376,7 @@ void pblk_write_should_kick(struct pblk *pblk)
 {
 	unsigned int secs_avail = pblk_rb_read_count(&pblk->rwb);
 
-	if (secs_avail >= pblk->min_write_pgs)
+	if (secs_avail >= pblk->min_write_pgs_data)
 		pblk_write_kick(pblk);
 }
 
@@ -407,7 +407,9 @@ struct list_head *pblk_line_gc_list(struct pblk *pblk, struct pblk_line *line)
 	struct pblk_line_meta *lm = &pblk->lm;
 	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
 	struct list_head *move_list = NULL;
-	int vsc = le32_to_cpu(*line->vsc);
+	int packed_meta = (le32_to_cpu(*line->vsc) / pblk->min_write_pgs_data)
+			* (pblk->min_write_pgs - pblk->min_write_pgs_data);
+	int vsc = le32_to_cpu(*line->vsc) + packed_meta;
 
 	lockdep_assert_held(&line->lock);
 
@@ -620,12 +622,15 @@ struct bio *pblk_bio_map_addr(struct pblk *pblk, void *data,
 }
 
 int pblk_calc_secs(struct pblk *pblk, unsigned long secs_avail,
-		   unsigned long secs_to_flush)
+		   unsigned long secs_to_flush, bool skip_meta)
 {
 	int max = pblk->sec_per_write;
 	int min = pblk->min_write_pgs;
 	int secs_to_sync = 0;
 
+	if (skip_meta && pblk->min_write_pgs_data != pblk->min_write_pgs)
+		min = max = pblk->min_write_pgs_data;
+
 	if (secs_avail >= max)
 		secs_to_sync = max;
 	else if (secs_avail >= min)
@@ -852,7 +857,7 @@ int pblk_line_emeta_read(struct pblk *pblk, struct pblk_line *line,
 next_rq:
 	memset(&rqd, 0, sizeof(struct nvm_rq));
 
-	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
+	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0, false);
 	rq_len = rq_ppas * geo->csecs;
 
 	bio = pblk_bio_map_addr(pblk, emeta_buf, rq_ppas, rq_len,
@@ -2169,3 +2174,38 @@ void pblk_lookup_l2p_rand(struct pblk *pblk, struct ppa_addr *ppas,
 	}
 	spin_unlock(&pblk->trans_lock);
 }
+
+void *pblk_get_meta_for_writes(struct pblk *pblk, struct nvm_rq *rqd)
+{
+	void *buffer;
+
+	if (pblk_is_oob_meta_supported(pblk)) {
+		/* Just use OOB metadata buffer as always */
+		buffer = rqd->meta_list;
+	} else {
+		/* We need to reuse last page of request (packed metadata)
+		 * in similar way as traditional oob metadata
+		 */
+		buffer = page_to_virt(
+			rqd->bio->bi_io_vec[rqd->bio->bi_vcnt - 1].bv_page);
+	}
+
+	return buffer;
+}
+
+void pblk_get_packed_meta(struct pblk *pblk, struct nvm_rq *rqd)
+{
+	void *meta_list = rqd->meta_list;
+	void *page;
+	int i = 0;
+
+	if (pblk_is_oob_meta_supported(pblk))
+		return;
+
+	page = page_to_virt(rqd->bio->bi_io_vec[rqd->bio->bi_vcnt - 1].bv_page);
+	/* We need to fill oob meta buffer with data from packed metadata */
+	for (; i < rqd->nr_ppas; i++)
+		memcpy(pblk_get_meta(pblk, meta_list, i),
+			page + (i * sizeof(struct pblk_sec_meta)),
+			sizeof(struct pblk_sec_meta));
+}

drivers/lightnvm/pblk-init.c

@@ -399,6 +399,7 @@ static int pblk_core_init(struct pblk *pblk)
 	pblk->nr_flush_rst = 0;
 
 	pblk->min_write_pgs = geo->ws_opt;
+	pblk->min_write_pgs_data = pblk->min_write_pgs;
 	max_write_ppas = pblk->min_write_pgs * geo->all_luns;
 	pblk->max_write_pgs = min_t(int, max_write_ppas, NVM_MAX_VLBA);
 	pblk->max_write_pgs = min_t(int, pblk->max_write_pgs,
@@ -406,10 +407,36 @@ static int pblk_core_init(struct pblk *pblk)
 	pblk_set_sec_per_write(pblk, pblk->min_write_pgs);
 
 	pblk->oob_meta_size = geo->sos;
-	if (pblk->oob_meta_size < sizeof(struct pblk_sec_meta)) {
-		pblk_err(pblk, "Unsupported metadata size\n");
+	if (!pblk_is_oob_meta_supported(pblk)) {
+		/* For drives which does not have OOB metadata feature
+		 * in order to support recovery feature we need to use
+		 * so called packed metadata. Packed metada will store
+		 * the same information as OOB metadata (l2p table mapping,
+		 * but in the form of the single page at the end of
+		 * every write request.
+		 */
+		if (pblk->min_write_pgs
+			* sizeof(struct pblk_sec_meta) > PAGE_SIZE) {
+			/* We want to keep all the packed metadata on single
+			 * page per write requests. So we need to ensure that
+			 * it will fit.
+			 *
+			 * This is more like sanity check, since there is
+			 * no device with such a big minimal write size
+			 * (above 1 metabytes).
+			 */
+			pblk_err(pblk, "Not supported min write size\n");
 			return -EINVAL;
 		}
+		/* For packed meta approach we do some simplification.
+		 * On read path we always issue requests which size
+		 * equal to max_write_pgs, with all pages filled with
+		 * user payload except of last one page which will be
+		 * filled with packed metadata.
+		 */
+		pblk->max_write_pgs = pblk->min_write_pgs;
+		pblk->min_write_pgs_data = pblk->min_write_pgs - 1;
+	}
 
 	pblk->pad_dist = kcalloc(pblk->min_write_pgs - 1, sizeof(atomic64_t),
 								GFP_KERNEL);
@@ -641,7 +668,7 @@ static int pblk_set_provision(struct pblk *pblk, int nr_free_chks)
 	struct pblk_line_meta *lm = &pblk->lm;
 	struct nvm_geo *geo = &dev->geo;
 	sector_t provisioned;
-	int sec_meta, blk_meta;
+	int sec_meta, blk_meta, clba;
 	int minimum;
 
 	if (geo->op == NVM_TARGET_DEFAULT_OP)
@@ -682,7 +709,8 @@ static int pblk_set_provision(struct pblk *pblk, int nr_free_chks)
 	sec_meta = (lm->smeta_sec + lm->emeta_sec[0]) * l_mg->nr_free_lines;
 	blk_meta = DIV_ROUND_UP(sec_meta, geo->clba);
 
-	pblk->capacity = (provisioned - blk_meta) * geo->clba;
+	clba = (geo->clba / pblk->min_write_pgs) * pblk->min_write_pgs_data;
+	pblk->capacity = (provisioned - blk_meta) * clba;
 
 	atomic_set(&pblk->rl.free_blocks, nr_free_chks);
 	atomic_set(&pblk->rl.free_user_blocks, nr_free_chks);

drivers/lightnvm/pblk-map.c

@@ -96,7 +96,7 @@ int pblk_map_rq(struct pblk *pblk, struct nvm_rq *rqd, unsigned int sentry,
 		 unsigned long *lun_bitmap, unsigned int valid_secs,
 		 unsigned int off)
 {
-	void *meta_list = rqd->meta_list;
+	void *meta_list = pblk_get_meta_for_writes(pblk, rqd);
 	void *meta_buffer;
 	struct ppa_addr *ppa_list = nvm_rq_to_ppa_list(rqd);
 	unsigned int map_secs;
@@ -125,7 +125,7 @@ int pblk_map_erase_rq(struct pblk *pblk, struct nvm_rq *rqd,
 	struct nvm_tgt_dev *dev = pblk->dev;
 	struct nvm_geo *geo = &dev->geo;
 	struct pblk_line_meta *lm = &pblk->lm;
-	void *meta_list = rqd->meta_list;
+	void *meta_list = pblk_get_meta_for_writes(pblk, rqd);
 	void *meta_buffer;
 	struct ppa_addr *ppa_list = nvm_rq_to_ppa_list(rqd);
 	struct pblk_line *e_line, *d_line;

drivers/lightnvm/pblk-rb.c

@@ -552,6 +552,9 @@ unsigned int pblk_rb_read_to_bio(struct pblk_rb *rb, struct nvm_rq *rqd,
 		to_read = count;
 	}
 
+	/* Add space for packed metadata if in use*/
+	pad += (pblk->min_write_pgs - pblk->min_write_pgs_data);
+
 	c_ctx->sentry = pos;
 	c_ctx->nr_valid = to_read;
 	c_ctx->nr_padded = pad;

drivers/lightnvm/pblk-read.c

@@ -112,6 +112,9 @@ static void pblk_read_check_seq(struct pblk *pblk, struct nvm_rq *rqd,
 	int nr_lbas = rqd->nr_ppas;
 	int i;
 
+	if (!pblk_is_oob_meta_supported(pblk))
+		return;
+
 	for (i = 0; i < nr_lbas; i++) {
 		struct pblk_sec_meta *meta = pblk_get_meta(pblk, meta_list, i);
 		u64 lba = le64_to_cpu(meta->lba);
@@ -141,6 +144,9 @@ static void pblk_read_check_rand(struct pblk *pblk, struct nvm_rq *rqd,
 	void *meta_lba_list = rqd->meta_list;
 	int i, j;
 
+	if (!pblk_is_oob_meta_supported(pblk))
+		return;
+
 	for (i = 0, j = 0; i < nr_lbas; i++) {
 		struct pblk_sec_meta *meta = pblk_get_meta(pblk,
 							   meta_lba_list, j);

drivers/lightnvm/pblk-recovery.c

@@ -191,7 +191,7 @@ static int pblk_recov_pad_line(struct pblk *pblk, struct pblk_line *line,
 	kref_init(&pad_rq->ref);
 
 next_pad_rq:
-	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
+	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0, false);
 	if (rq_ppas < pblk->min_write_pgs) {
 		pblk_err(pblk, "corrupted pad line %d\n", line->id);
 		goto fail_free_pad;
@@ -371,17 +371,19 @@ static int pblk_recov_scan_oob(struct pblk *pblk, struct pblk_line *line,
 next_rq:
 	memset(rqd, 0, pblk_g_rq_size);
 
-	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
+	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0, false);
 	if (!rq_ppas)
 		rq_ppas = pblk->min_write_pgs;
 	rq_len = rq_ppas * geo->csecs;
 
+retry_rq:
 	bio = bio_map_kern(dev->q, data, rq_len, GFP_KERNEL);
 	if (IS_ERR(bio))
 		return PTR_ERR(bio);
 
 	bio->bi_iter.bi_sector = 0; /* internal bio */
 	bio_set_op_attrs(bio, REQ_OP_READ, 0);
+	bio_get(bio);
 
 	rqd->bio = bio;
 	rqd->opcode = NVM_OP_PREAD;
@@ -394,7 +396,6 @@ static int pblk_recov_scan_oob(struct pblk *pblk, struct pblk_line *line,
 	if (pblk_io_aligned(pblk, rq_ppas))
 		rqd->is_seq = 1;
 
-retry_rq:
 	for (i = 0; i < rqd->nr_ppas; ) {
 		struct ppa_addr ppa;
 		int pos;
@@ -417,6 +418,7 @@ static int pblk_recov_scan_oob(struct pblk *pblk, struct pblk_line *line,
 	if (ret) {
 		pblk_err(pblk, "I/O submission failed: %d\n", ret);
 		bio_put(bio);
+		bio_put(bio);
 		return ret;
 	}
 
@@ -428,18 +430,25 @@ static int pblk_recov_scan_oob(struct pblk *pblk, struct pblk_line *line,
 
 		if (padded) {
 			pblk_log_read_err(pblk, rqd);
+			bio_put(bio);
 			return -EINTR;
 		}
 
 		pad_distance = pblk_pad_distance(pblk, line);
 		ret = pblk_recov_pad_line(pblk, line, pad_distance);
-		if (ret)
+		if (ret) {
+			bio_put(bio);
 			return ret;
+		}
 
 		padded = true;
+		bio_put(bio);
 		goto retry_rq;
 	}
 
+	pblk_get_packed_meta(pblk, rqd);
+	bio_put(bio);
+
 	for (i = 0; i < rqd->nr_ppas; i++) {
 		struct pblk_sec_meta *meta = pblk_get_meta(pblk, meta_list, i);
 		u64 lba = le64_to_cpu(meta->lba);

drivers/lightnvm/pblk-sysfs.c

@@ -479,6 +479,13 @@ static ssize_t pblk_sysfs_set_sec_per_write(struct pblk *pblk,
 	if (kstrtouint(page, 0, &sec_per_write))
 		return -EINVAL;
 
+	if (!pblk_is_oob_meta_supported(pblk)) {
+		/* For packed metadata case it is
+		 * not allowed to change sec_per_write.
+		 */
+		return -EINVAL;
+	}
+
 	if (sec_per_write < pblk->min_write_pgs
 				|| sec_per_write > pblk->max_write_pgs
 				|| sec_per_write % pblk->min_write_pgs != 0)

drivers/lightnvm/pblk-write.c

@@ -348,7 +348,7 @@ static int pblk_calc_secs_to_sync(struct pblk *pblk, unsigned int secs_avail,
 {
 	int secs_to_sync;
 
-	secs_to_sync = pblk_calc_secs(pblk, secs_avail, secs_to_flush);
+	secs_to_sync = pblk_calc_secs(pblk, secs_avail, secs_to_flush, true);
 
 #ifdef CONFIG_NVM_PBLK_DEBUG
 	if ((!secs_to_sync && secs_to_flush)
@@ -569,7 +569,7 @@ static int pblk_submit_write(struct pblk *pblk, int *secs_left)
 	struct bio *bio;
 	struct nvm_rq *rqd;
 	unsigned int secs_avail, secs_to_sync, secs_to_com;
-	unsigned int secs_to_flush;
+	unsigned int secs_to_flush, packed_meta_pgs;
 	unsigned long pos;
 	unsigned int resubmit;
 
@@ -607,7 +607,7 @@ static int pblk_submit_write(struct pblk *pblk, int *secs_left)
 			return 0;
 
 		secs_to_flush = pblk_rb_flush_point_count(&pblk->rwb);
-		if (!secs_to_flush && secs_avail < pblk->min_write_pgs)
+		if (!secs_to_flush && secs_avail < pblk->min_write_pgs_data)
 			return 0;
 
 		secs_to_sync = pblk_calc_secs_to_sync(pblk, secs_avail,
@@ -622,7 +622,8 @@ static int pblk_submit_write(struct pblk *pblk, int *secs_left)
 		pos = pblk_rb_read_commit(&pblk->rwb, secs_to_com);
 	}
 
-	bio = bio_alloc(GFP_KERNEL, secs_to_sync);
+	packed_meta_pgs = (pblk->min_write_pgs - pblk->min_write_pgs_data);
+	bio = bio_alloc(GFP_KERNEL, secs_to_sync + packed_meta_pgs);
 
 	bio->bi_iter.bi_sector = 0; /* internal bio */
 	bio_set_op_attrs(bio, REQ_OP_WRITE, 0);

drivers/lightnvm/pblk.h

@@ -632,6 +632,7 @@ struct pblk {
 	int state;			/* pblk line state */
 
 	int min_write_pgs; /* Minimum amount of pages required by controller */
+	int min_write_pgs_data; /* Minimum amount of payload pages */
 	int max_write_pgs; /* Maximum amount of pages supported by controller */
 	int oob_meta_size; /* Size of OOB sector metadata */
 
@@ -838,7 +839,7 @@ void pblk_dealloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs);
 u64 pblk_alloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs);
 u64 __pblk_alloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs);
 int pblk_calc_secs(struct pblk *pblk, unsigned long secs_avail,
-		   unsigned long secs_to_flush);
+		   unsigned long secs_to_flush, bool skip_meta);
 void pblk_down_rq(struct pblk *pblk, struct ppa_addr ppa,
 		  unsigned long *lun_bitmap);
 void pblk_down_chunk(struct pblk *pblk, struct ppa_addr ppa);
@@ -862,6 +863,8 @@ void pblk_lookup_l2p_rand(struct pblk *pblk, struct ppa_addr *ppas,
 			  u64 *lba_list, int nr_secs);
 void pblk_lookup_l2p_seq(struct pblk *pblk, struct ppa_addr *ppas,
 			 sector_t blba, int nr_secs);
+void *pblk_get_meta_for_writes(struct pblk *pblk, struct nvm_rq *rqd);
+void pblk_get_packed_meta(struct pblk *pblk, struct nvm_rq *rqd);
 
 /*
  * pblk user I/O write path
@@ -1392,4 +1395,9 @@ static inline int pblk_dma_meta_size(struct pblk *pblk)
 {
 	return pblk->oob_meta_size * NVM_MAX_VLBA;
 }
+
+static inline int pblk_is_oob_meta_supported(struct pblk *pblk)
+{
+	return pblk->oob_meta_size >= sizeof(struct pblk_sec_meta);
+}
 #endif /* PBLK_H_ */