[PATCH 00/28] LightNVM fixes for 4.7

[PATCH 00/28] LightNVM fixes for 4.7

[Matias Bjørling]

Hi Jens,

A collection of fixes and features destined for 4.7. They contain:

  Fixes all over the place from Arnd, Jeff, Sagi, Wenwei, Javier and me.

  Preparation patches from Simon to enable sysfs support. The final
  patch is delayed to 4.8. We need to integrate it correctly with
  device drivers.

  A patch from Javier to allow a target to access the metadata field.
  Typically it maps to the flash page out of band area.

  Preparation patches for pblk and persistent block management
  information from me.

Please pick up.

-Matias

Arnd Bergmann (1):
  lightnvm: pass dma address to hardware rather than pointer

Javier González (6):
  lightnvm: do not free unused metadata on rrpc
  lightnvm: enable metadata to be sent to device
  lightnvm: rename dma helper functions
  lightnvm: do not assume sequential lun alloc.
  lightnvm: rename nr_pages to nr_ppas on nvm_rq
  lightnvm: reserved space calculation incorrect

Jeff Mahoney (1):
  lightnvm: fix "warning: ‘ret’ may be used uninitialized"

Matias Bjørling (15):
  lightnvm: handle submit_io failure
  lightnvm: implement nvm_submit_ppa_list
  lightnvm: add fpg_size and pfpg_size to struct nvm_dev
  lightnvm: move block fold outside of get_bb_tbl()
  lightnvm: avoid memory leak when lun_map kcalloc fails
  lightnvm: introduce nvm_for_each_lun_ppa() macro
  lightnvm: refactor device ops->get_bb_tbl()
  lightnvm: remove struct factory_blks
  lightnvm: make nvm_set_rqd_ppalist() aware of vblks
  lightnvm: move responsibility for bad blk mgmt to target
  lightnvm: refactor set_bb_tbl for accepting ppa list
  lightnvm: fix out of bound ppa lun id on bb tbl
  lightnvm: remove mgt targets on mgt removal
  lightnvm: expose gennvm_mark_blk to targets
  lightnvm: add is_cached entry to struct ppa_addr

Sagi Grimberg (1):
  nvme/lightnvm: Log using the ctrl named device

Simon A. F. Lund (2):
  lightnvm: rename nvm_targets to nvm_tgt_type
  lightnvm: refactor dev->online_target to global nvm_targets

Wenwei Tao (2):
  lightnvm: calculate rrpc total blocks and sectors up front
  lightnvm: store rrpc->soffset in device sector size

 drivers/lightnvm/core.c      | 370 +++++++++++++++++++++++++++++--------------
 drivers/lightnvm/gennvm.c    | 100 ++++++------
 drivers/lightnvm/rrpc.c      |  42 +++--
 drivers/lightnvm/rrpc.h      |   2 +-
 drivers/lightnvm/sysblk.c    | 276 ++++++++++++++++----------------
 drivers/nvme/host/lightnvm.c |  82 +++-------
 include/linux/lightnvm.h     |  48 ++++--
 7 files changed, 521 insertions(+), 399 deletions(-)

-- 
2.1.4

[PATCH 01/28] lightnvm: fix "warning: ‘ret’ may be used uninitialized"

[Matias Bjørling]

From: Jeff Mahoney <jeffm@suse.com>

This fixes the following warnings:
drivers/lightnvm/sysblk.c:125:9: warning: ‘ret’ may be used
uninitialized in this function

drivers/lightnvm/sysblk.c:275:15: warning: ‘ret’ may be used
uninitialized in this function

In both cases, ret is only set from within a loop that may not be entered.

Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/sysblk.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index 321de1f..b1e1404 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -122,7 +122,7 @@ static int nvm_get_all_sysblks(struct nvm_dev *dev, struct sysblk_scan *s,
 				struct ppa_addr *ppas, nvm_bb_update_fn *fn)
 {
 	struct ppa_addr dppa;
-	int i, ret;
+	int i, ret = 0;
 
 	s->nr_ppas = 0;
 
@@ -272,7 +272,7 @@ static int nvm_write_and_verify(struct nvm_dev *dev, struct nvm_sb_info *info,
 {
 	struct nvm_system_block nvmsb;
 	void *buf;
-	int i, sect, ret, bufsz;
+	int i, sect, ret = 0, bufsz;
 	struct ppa_addr *ppas;
 
 	nvm_cpu_to_sysblk(&nvmsb, info);
-- 
2.1.4

[PATCH 02/28] lightnvm: handle submit_io failure

[Matias Bjørling]

The device ->submit_io() callback might fail to submit I/O to device.
In that case, the nvm_submit_ppa function should not wait for
completion. Instead return the ->submit_io() error.

Reviewed by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 0dc9a80..c2ef53a 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -351,6 +351,11 @@ int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
 	nvm_generic_to_addr_mode(dev, &rqd);
 
 	ret = dev->ops->submit_io(dev, &rqd);
+	if (ret) {
+		nvm_free_rqd_ppalist(dev, &rqd);
+		bio_put(bio);
+		return ret;
+	}
 
 	/* Prevent hang_check timer from firing at us during very long I/O */
 	hang_check = sysctl_hung_task_timeout_secs;
-- 
2.1.4

[PATCH 03/28] lightnvm: implement nvm_submit_ppa_list

[Matias Bjørling]

The nvm_submit_ppa function assumes that users manage all plane
blocks as a single block. Extend the API with nvm_submit_ppa_list
to allow the user to send its own ppa list. If the user submits more
than a single PPA, the user must take care to allocate and free
the corresponding ppa list.

Reviewed by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c  | 88 +++++++++++++++++++++++++++++++++++++-----------
 include/linux/lightnvm.h |  2 ++
 2 files changed, 71 insertions(+), 19 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index c2ef53a..f4e04a5 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -322,11 +322,10 @@ static void nvm_end_io_sync(struct nvm_rq *rqd)
 	complete(waiting);
 }
 
-int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
-				int opcode, int flags, void *buf, int len)
+int __nvm_submit_ppa(struct nvm_dev *dev, struct nvm_rq *rqd, int opcode,
+						int flags, void *buf, int len)
 {
 	DECLARE_COMPLETION_ONSTACK(wait);
-	struct nvm_rq rqd;
 	struct bio *bio;
 	int ret;
 	unsigned long hang_check;
@@ -335,24 +334,17 @@ int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
 	if (IS_ERR_OR_NULL(bio))
 		return -ENOMEM;
 
-	memset(&rqd, 0, sizeof(struct nvm_rq));
-	ret = nvm_set_rqd_ppalist(dev, &rqd, ppa, nr_ppas);
-	if (ret) {
-		bio_put(bio);
-		return ret;
-	}
+	nvm_generic_to_addr_mode(dev, rqd);
 
-	rqd.opcode = opcode;
-	rqd.bio = bio;
-	rqd.wait = &wait;
-	rqd.dev = dev;
-	rqd.end_io = nvm_end_io_sync;
-	rqd.flags = flags;
-	nvm_generic_to_addr_mode(dev, &rqd);
+	rqd->dev = dev;
+	rqd->opcode = opcode;
+	rqd->flags = flags;
+	rqd->bio = bio;
+	rqd->wait = &wait;
+	rqd->end_io = nvm_end_io_sync;
 
-	ret = dev->ops->submit_io(dev, &rqd);
+	ret = dev->ops->submit_io(dev, rqd);
 	if (ret) {
-		nvm_free_rqd_ppalist(dev, &rqd);
 		bio_put(bio);
 		return ret;
 	}
@@ -364,9 +356,67 @@ int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
 	else
 		wait_for_completion_io(&wait);
 
+	return rqd->error;
+}
+
+/**
+ * nvm_submit_ppa_list - submit user-defined ppa list to device. The user must
+ *			 take to free ppa list if necessary.
+ * @dev:	device
+ * @ppa_list:	user created ppa_list
+ * @nr_ppas:	length of ppa_list
+ * @opcode:	device opcode
+ * @flags:	device flags
+ * @buf:	data buffer
+ * @len:	data buffer length
+ */
+int nvm_submit_ppa_list(struct nvm_dev *dev, struct ppa_addr *ppa_list,
+			int nr_ppas, int opcode, int flags, void *buf, int len)
+{
+	struct nvm_rq rqd;
+
+	if (dev->ops->max_phys_sect < nr_ppas)
+		return -EINVAL;
+
+	memset(&rqd, 0, sizeof(struct nvm_rq));
+
+	rqd.nr_pages = nr_ppas;
+	if (nr_ppas > 1)
+		rqd.ppa_list = ppa_list;
+	else
+		rqd.ppa_addr = ppa_list[0];
+
+	return __nvm_submit_ppa(dev, &rqd, opcode, flags, buf, len);
+}
+EXPORT_SYMBOL(nvm_submit_ppa_list);
+
+/**
+ * nvm_submit_ppa - submit PPAs to device. PPAs will automatically be unfolded
+ *		    as single, dual, quad plane PPAs depending on device type.
+ * @dev:	device
+ * @ppa:	user created ppa_list
+ * @nr_ppas:	length of ppa_list
+ * @opcode:	device opcode
+ * @flags:	device flags
+ * @buf:	data buffer
+ * @len:	data buffer length
+ */
+int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
+				int opcode, int flags, void *buf, int len)
+{
+	struct nvm_rq rqd;
+	int ret;
+
+	memset(&rqd, 0, sizeof(struct nvm_rq));
+	ret = nvm_set_rqd_ppalist(dev, &rqd, ppa, nr_ppas);
+	if (ret)
+		return ret;
+
+	ret = __nvm_submit_ppa(dev, &rqd, opcode, flags, buf, len);
+
 	nvm_free_rqd_ppalist(dev, &rqd);
 
-	return rqd.error;
+	return ret;
 }
 EXPORT_SYMBOL(nvm_submit_ppa);
 
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index cdcb2cc..38814e2 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -534,6 +534,8 @@ extern int nvm_erase_blk(struct nvm_dev *, struct nvm_block *);
 extern void nvm_end_io(struct nvm_rq *, int);
 extern int nvm_submit_ppa(struct nvm_dev *, struct ppa_addr *, int, int, int,
 								void *, int);
+extern int nvm_submit_ppa_list(struct nvm_dev *, struct ppa_addr *, int, int,
+							int, void *, int);
 
 /* sysblk.c */
 #define NVM_SYSBLK_MAGIC 0x4E564D53 /* "NVMS" */
-- 
2.1.4

[PATCH 04/28] lightnvm: add fpg_size and pfpg_size to struct nvm_dev

[Matias Bjørling]

The flash page size (fpg) and size across planes (pfpg) are convenient
to know when allocating buffer sizes. This has previously been a
calculated in various places. Replace with the pre-calculated values.

Reviewed by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c   |  2 ++
 drivers/lightnvm/sysblk.c | 17 +++++++----------
 include/linux/lightnvm.h  |  2 ++
 3 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index f4e04a5..652b8c7 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -476,6 +476,8 @@ static int nvm_core_init(struct nvm_dev *dev)
 	dev->pgs_per_blk = grp->num_pg;
 	dev->blks_per_lun = grp->num_blk;
 	dev->nr_planes = grp->num_pln;
+	dev->fpg_size = grp->fpg_sz;
+	dev->pfpg_size = grp->fpg_sz * grp->num_pln;
 	dev->sec_size = grp->csecs;
 	dev->oob_size = grp->sos;
 	dev->sec_per_pg = grp->fpg_sz / grp->csecs;
diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index b1e1404..b6971f8 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -154,13 +154,12 @@ static int nvm_scan_block(struct nvm_dev *dev, struct ppa_addr *ppa,
 						struct nvm_system_block *sblk)
 {
 	struct nvm_system_block *cur;
-	int pg, cursz, ret, found = 0;
+	int pg, ret, found = 0;
 
 	/* the full buffer for a flash page is allocated. Only the first of it
 	 * contains the system block information
 	 */
-	cursz = dev->sec_size * dev->sec_per_pg * dev->nr_planes;
-	cur = kmalloc(cursz, GFP_KERNEL);
+	cur = kmalloc(dev->pfpg_size, GFP_KERNEL);
 	if (!cur)
 		return -ENOMEM;
 
@@ -169,7 +168,7 @@ static int nvm_scan_block(struct nvm_dev *dev, struct ppa_addr *ppa,
 		ppa->g.pg = ppa_to_slc(dev, pg);
 
 		ret = nvm_submit_ppa(dev, ppa, 1, NVM_OP_PREAD, NVM_IO_SLC_MODE,
-								cur, cursz);
+							cur, dev->pfpg_size);
 		if (ret) {
 			if (ret == NVM_RSP_ERR_EMPTYPAGE) {
 				pr_debug("nvm: sysblk scan empty ppa (%u %u %u %u)\n",
@@ -272,14 +271,12 @@ static int nvm_write_and_verify(struct nvm_dev *dev, struct nvm_sb_info *info,
 {
 	struct nvm_system_block nvmsb;
 	void *buf;
-	int i, sect, ret = 0, bufsz;
+	int i, sect, ret = 0;
 	struct ppa_addr *ppas;
 
 	nvm_cpu_to_sysblk(&nvmsb, info);
 
-	/* buffer for flash page */
-	bufsz = dev->sec_size * dev->sec_per_pg * dev->nr_planes;
-	buf = kzalloc(bufsz, GFP_KERNEL);
+	buf = kzalloc(dev->pfpg_size, GFP_KERNEL);
 	if (!buf)
 		return -ENOMEM;
 	memcpy(buf, &nvmsb, sizeof(struct nvm_system_block));
@@ -309,7 +306,7 @@ static int nvm_write_and_verify(struct nvm_dev *dev, struct nvm_sb_info *info,
 		}
 
 		ret = nvm_submit_ppa(dev, ppas, dev->sec_per_pg, NVM_OP_PWRITE,
-						NVM_IO_SLC_MODE, buf, bufsz);
+					NVM_IO_SLC_MODE, buf, dev->pfpg_size);
 		if (ret) {
 			pr_err("nvm: sysblk failed program (%u %u %u)\n",
 							ppas[0].g.ch,
@@ -319,7 +316,7 @@ static int nvm_write_and_verify(struct nvm_dev *dev, struct nvm_sb_info *info,
 		}
 
 		ret = nvm_submit_ppa(dev, ppas, dev->sec_per_pg, NVM_OP_PREAD,
-						NVM_IO_SLC_MODE, buf, bufsz);
+					NVM_IO_SLC_MODE, buf, dev->pfpg_size);
 		if (ret) {
 			pr_err("nvm: sysblk failed read (%u %u %u)\n",
 							ppas[0].g.ch,
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 38814e2..f7c607f 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -323,6 +323,8 @@ struct nvm_dev {
 	int sec_per_pg; /* only sectors for a single page */
 	int pgs_per_blk;
 	int blks_per_lun;
+	int fpg_size;
+	int pfpg_size; /* size of buffer if all pages are to be read */
 	int sec_size;
 	int oob_size;
 	int mccap;
-- 
2.1.4

[PATCH 05/28] lightnvm: move block fold outside of get_bb_tbl()

[Matias Bjørling]

The get block table command returns a list of blocks and planes
with their associated state. Users, such as gennvm and sysblk,
manages all planes as a single virtual block.

It was therefore  natural to fold the bad block list before it is
returned. However, to allow users, which manages on a per-plane
block level, to also use the interface, the get_bb_tbl interface is
changed to not fold by default and instead let the caller fold if
necessary.

Reviewed by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c      | 35 +++++++++++++++++++++++++++++++++
 drivers/lightnvm/gennvm.c    | 14 +++++++------
 drivers/lightnvm/sysblk.c    | 29 ++++++++++++++++++---------
 drivers/nvme/host/lightnvm.c | 47 ++++++--------------------------------------
 include/linux/lightnvm.h     |  6 ++++--
 5 files changed, 73 insertions(+), 58 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 652b8c7..4cadbe0 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -420,6 +420,41 @@ int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
 }
 EXPORT_SYMBOL(nvm_submit_ppa);
 
+/*
+ * folds a bad block list from its plane representation to its virtual
+ * block representation. The fold is done in place and reduced size is
+ * returned.
+ *
+ * If any of the planes status are bad or grown bad block, the virtual block
+ * is marked bad. If not bad, the first plane state acts as the block state.
+ */
+int nvm_bb_tbl_fold(struct nvm_dev *dev, u8 *blks, int nr_blks)
+{
+	int blk, offset, pl, blktype;
+
+	if (nr_blks != dev->blks_per_lun * dev->plane_mode)
+		return -EINVAL;
+
+	for (blk = 0; blk < dev->blks_per_lun; blk++) {
+		offset = blk * dev->plane_mode;
+		blktype = blks[offset];
+
+		/* Bad blocks on any planes take precedence over other types */
+		for (pl = 0; pl < dev->plane_mode; pl++) {
+			if (blks[offset + pl] &
+					(NVM_BLK_T_BAD|NVM_BLK_T_GRWN_BAD)) {
+				blktype = blks[offset + pl];
+				break;
+			}
+		}
+
+		blks[blk] = blktype;
+	}
+
+	return dev->blks_per_lun;
+}
+EXPORT_SYMBOL(nvm_bb_tbl_fold);
+
 static int nvm_init_slc_tbl(struct nvm_dev *dev, struct nvm_id_group *grp)
 {
 	int i;
diff --git a/drivers/lightnvm/gennvm.c b/drivers/lightnvm/gennvm.c
index 72e124a..6096077 100644
--- a/drivers/lightnvm/gennvm.c
+++ b/drivers/lightnvm/gennvm.c
@@ -129,18 +129,21 @@ static int gennvm_luns_init(struct nvm_dev *dev, struct gen_nvm *gn)
 	return 0;
 }
 
-static int gennvm_block_bb(struct ppa_addr ppa, int nr_blocks, u8 *blks,
-								void *private)
+static int gennvm_block_bb(struct nvm_dev *dev, struct ppa_addr ppa,
+					u8 *blks, int nr_blks, void *private)
 {
 	struct gen_nvm *gn = private;
-	struct nvm_dev *dev = gn->dev;
 	struct gen_lun *lun;
 	struct nvm_block *blk;
 	int i;
 
+	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
+	if (nr_blks < 0)
+		return nr_blks;
+
 	lun = &gn->luns[(dev->luns_per_chnl * ppa.g.ch) + ppa.g.lun];
 
-	for (i = 0; i < nr_blocks; i++) {
+	for (i = 0; i < nr_blks; i++) {
 		if (blks[i] == 0)
 			continue;
 
@@ -250,8 +253,7 @@ static int gennvm_blocks_init(struct nvm_dev *dev, struct gen_nvm *gn)
 			ppa = generic_to_dev_addr(dev, ppa);
 
 			ret = dev->ops->get_bb_tbl(dev, ppa,
-						dev->blks_per_lun,
-						gennvm_block_bb, gn);
+							gennvm_block_bb, gn);
 			if (ret)
 				pr_err("gennvm: could not read BB table\n");
 		}
diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index b6971f8..7fce588 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -93,12 +93,16 @@ void nvm_setup_sysblk_scan(struct nvm_dev *dev, struct sysblk_scan *s,
 	s->nr_rows = nvm_setup_sysblks(dev, sysblk_ppas);
 }
 
-static int sysblk_get_host_blks(struct ppa_addr ppa, int nr_blks, u8 *blks,
-								void *private)
+static int sysblk_get_host_blks(struct nvm_dev *dev, struct ppa_addr ppa,
+					u8 *blks, int nr_blks, void *private)
 {
 	struct sysblk_scan *s = private;
 	int i, nr_sysblk = 0;
 
+	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
+	if (nr_blks < 0)
+		return nr_blks;
+
 	for (i = 0; i < nr_blks; i++) {
 		if (blks[i] != NVM_BLK_T_HOST)
 			continue;
@@ -130,7 +134,7 @@ static int nvm_get_all_sysblks(struct nvm_dev *dev, struct sysblk_scan *s,
 		dppa = generic_to_dev_addr(dev, ppas[i]);
 		s->row = i;
 
-		ret = dev->ops->get_bb_tbl(dev, dppa, dev->blks_per_lun, fn, s);
+		ret = dev->ops->get_bb_tbl(dev, dppa, fn, s);
 		if (ret) {
 			pr_err("nvm: failed bb tbl for ppa (%u %u)\n",
 							ppas[i].g.ch,
@@ -235,13 +239,17 @@ static int nvm_set_bb_tbl(struct nvm_dev *dev, struct sysblk_scan *s, int type)
 	return 0;
 }
 
-static int sysblk_get_free_blks(struct ppa_addr ppa, int nr_blks, u8 *blks,
-								void *private)
+static int sysblk_get_free_blks(struct nvm_dev *dev, struct ppa_addr ppa,
+					u8 *blks, int nr_blks, void *private)
 {
 	struct sysblk_scan *s = private;
 	struct ppa_addr *sppa;
 	int i, blkid = 0;
 
+	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
+	if (nr_blks < 0)
+		return nr_blks;
+
 	for (i = 0; i < nr_blks; i++) {
 		if (blks[i] == NVM_BLK_T_HOST)
 			return -EEXIST;
@@ -578,13 +586,16 @@ static unsigned int factory_blk_offset(struct nvm_dev *dev, int ch, int lun)
 								BITS_PER_LONG;
 }
 
-static int nvm_factory_blks(struct ppa_addr ppa, int nr_blks, u8 *blks,
-								void *private)
+static int nvm_factory_blks(struct nvm_dev *dev, struct ppa_addr ppa,
+					u8 *blks, int nr_blks, void *private)
 {
 	struct factory_blks *f = private;
-	struct nvm_dev *dev = f->dev;
 	int i, lunoff;
 
+	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
+	if (nr_blks < 0)
+		return nr_blks;
+
 	lunoff = factory_blk_offset(dev, ppa.g.ch, ppa.g.lun);
 
 	/* non-set bits correspond to the block must be erased */
@@ -661,7 +672,7 @@ static int nvm_fact_get_bb_tbl(struct nvm_dev *dev, struct ppa_addr ppa,
 
 	dev_ppa = generic_to_dev_addr(dev, ppa);
 
-	ret = dev->ops->get_bb_tbl(dev, dev_ppa, dev->blks_per_lun, fn, priv);
+	ret = dev->ops->get_bb_tbl(dev, dev_ppa, fn, priv);
 	if (ret)
 		pr_err("nvm: failed bb tbl for ch%u lun%u\n",
 							ppa.g.ch, ppa.g.blk);
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 9461dd6..d289980 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -387,41 +387,16 @@ out:
 	return ret;
 }
 
-static void nvme_nvm_bb_tbl_fold(struct nvm_dev *nvmdev,
-						int nr_dst_blks, u8 *dst_blks,
-						int nr_src_blks, u8 *src_blks)
-{
-	int blk, offset, pl, blktype;
-
-	for (blk = 0; blk < nr_dst_blks; blk++) {
-		offset = blk * nvmdev->plane_mode;
-		blktype = src_blks[offset];
-
-		/* Bad blocks on any planes take precedence over other types */
-		for (pl = 0; pl < nvmdev->plane_mode; pl++) {
-			if (src_blks[offset + pl] &
-					(NVM_BLK_T_BAD|NVM_BLK_T_GRWN_BAD)) {
-				blktype = src_blks[offset + pl];
-				break;
-			}
-		}
-
-		dst_blks[blk] = blktype;
-	}
-}
-
 static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
-				int nr_dst_blks, nvm_bb_update_fn *update_bbtbl,
-				void *priv)
+				nvm_bb_update_fn *update_bbtbl, void *priv)
 {
 	struct request_queue *q = nvmdev->q;
 	struct nvme_ns *ns = q->queuedata;
 	struct nvme_ctrl *ctrl = ns->ctrl;
 	struct nvme_nvm_command c = {};
 	struct nvme_nvm_bb_tbl *bb_tbl;
-	u8 *dst_blks = NULL;
-	int nr_src_blks = nr_dst_blks * nvmdev->plane_mode;
-	int tblsz = sizeof(struct nvme_nvm_bb_tbl) + nr_src_blks;
+	int nr_blks = nvmdev->blks_per_lun * nvmdev->plane_mode;
+	int tblsz = sizeof(struct nvme_nvm_bb_tbl) + nr_blks;
 	int ret = 0;
 
 	c.get_bb.opcode = nvme_nvm_admin_get_bb_tbl;
@@ -432,12 +407,6 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 	if (!bb_tbl)
 		return -ENOMEM;
 
-	dst_blks = kzalloc(nr_dst_blks, GFP_KERNEL);
-	if (!dst_blks) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
 	ret = nvme_submit_sync_cmd(ctrl->admin_q, (struct nvme_command *)&c,
 								bb_tbl, tblsz);
 	if (ret) {
@@ -459,21 +428,17 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 		goto out;
 	}
 
-	if (le32_to_cpu(bb_tbl->tblks) != nr_src_blks) {
+	if (le32_to_cpu(bb_tbl->tblks) != nr_blks) {
 		ret = -EINVAL;
 		dev_err(ctrl->dev, "bbt unsuspected blocks returned (%u!=%u)",
-				le32_to_cpu(bb_tbl->tblks), nr_src_blks);
+				le32_to_cpu(bb_tbl->tblks), nr_blks);
 		goto out;
 	}
 
-	nvme_nvm_bb_tbl_fold(nvmdev, nr_dst_blks, dst_blks,
-						nr_src_blks, bb_tbl->blk);
-
 	ppa = dev_to_generic_addr(nvmdev, ppa);
-	ret = update_bbtbl(ppa, nr_dst_blks, dst_blks, priv);
+	ret = update_bbtbl(nvmdev, ppa, bb_tbl->blk, nr_blks, priv);
 
 out:
-	kfree(dst_blks);
 	kfree(bb_tbl);
 	return ret;
 }
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index f7c607f..dacaa28 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -41,11 +41,12 @@ struct nvm_id;
 struct nvm_dev;
 
 typedef int (nvm_l2p_update_fn)(u64, u32, __le64 *, void *);
-typedef int (nvm_bb_update_fn)(struct ppa_addr, int, u8 *, void *);
+typedef int (nvm_bb_update_fn)(struct nvm_dev *, struct ppa_addr, u8 *, int,
+								void *);
 typedef int (nvm_id_fn)(struct nvm_dev *, struct nvm_id *);
 typedef int (nvm_get_l2p_tbl_fn)(struct nvm_dev *, u64, u32,
 				nvm_l2p_update_fn *, void *);
-typedef int (nvm_op_bb_tbl_fn)(struct nvm_dev *, struct ppa_addr, int,
+typedef int (nvm_op_bb_tbl_fn)(struct nvm_dev *, struct ppa_addr,
 				nvm_bb_update_fn *, void *);
 typedef int (nvm_op_set_bb_fn)(struct nvm_dev *, struct nvm_rq *, int);
 typedef int (nvm_submit_io_fn)(struct nvm_dev *, struct nvm_rq *);
@@ -538,6 +539,7 @@ extern int nvm_submit_ppa(struct nvm_dev *, struct ppa_addr *, int, int, int,
 								void *, int);
 extern int nvm_submit_ppa_list(struct nvm_dev *, struct ppa_addr *, int, int,
 							int, void *, int);
+extern int nvm_bb_tbl_fold(struct nvm_dev *, u8 *, int);
 
 /* sysblk.c */
 #define NVM_SYSBLK_MAGIC 0x4E564D53 /* "NVMS" */
-- 
2.1.4

[PATCH 06/28] lightnvm: avoid memory leak when lun_map kcalloc fails

[Matias Bjørling]

A memory leak occurs if the lower page table is initialized and the
following dev->lun_map fails on allocation.

Rearrange the initialization of lower page table to allow dev->lun_map
to fail gracefully without memory leak.

Reviewed by: Johannes Thumshirn <jthumshirn@suse.de>
Move kfree of dev->lun_map to nvm_free()
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c | 53 ++++++++++++++++++++++++++++---------------------
 1 file changed, 30 insertions(+), 23 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 4cadbe0..74fb049 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -504,6 +504,7 @@ static int nvm_core_init(struct nvm_dev *dev)
 {
 	struct nvm_id *id = &dev->identity;
 	struct nvm_id_group *grp = &id->groups[0];
+	int ret;
 
 	/* device values */
 	dev->nr_chnls = grp->num_ch;
@@ -522,33 +523,16 @@ static int nvm_core_init(struct nvm_dev *dev)
 	dev->plane_mode = NVM_PLANE_SINGLE;
 	dev->max_rq_size = dev->ops->max_phys_sect * dev->sec_size;
 
-	if (grp->mtype != 0) {
-		pr_err("nvm: memory type not supported\n");
-		return -EINVAL;
-	}
-
-	switch (grp->fmtype) {
-	case NVM_ID_FMTYPE_SLC:
-		if (nvm_init_slc_tbl(dev, grp))
-			return -ENOMEM;
-		break;
-	case NVM_ID_FMTYPE_MLC:
-		if (nvm_init_mlc_tbl(dev, grp))
-			return -ENOMEM;
-		break;
-	default:
-		pr_err("nvm: flash type not supported\n");
-		return -EINVAL;
-	}
-
-	if (!dev->lps_per_blk)
-		pr_info("nvm: lower page programming table missing\n");
-
 	if (grp->mpos & 0x020202)
 		dev->plane_mode = NVM_PLANE_DOUBLE;
 	if (grp->mpos & 0x040404)
 		dev->plane_mode = NVM_PLANE_QUAD;
 
+	if (grp->mtype != 0) {
+		pr_err("nvm: memory type not supported\n");
+		return -EINVAL;
+	}
+
 	/* calculated values */
 	dev->sec_per_pl = dev->sec_per_pg * dev->nr_planes;
 	dev->sec_per_blk = dev->sec_per_pl * dev->pgs_per_blk;
@@ -560,11 +544,34 @@ static int nvm_core_init(struct nvm_dev *dev)
 					sizeof(unsigned long), GFP_KERNEL);
 	if (!dev->lun_map)
 		return -ENOMEM;
+
+	switch (grp->fmtype) {
+	case NVM_ID_FMTYPE_SLC:
+		if (nvm_init_slc_tbl(dev, grp)) {
+			ret = -ENOMEM;
+			goto err_fmtype;
+		}
+		break;
+	case NVM_ID_FMTYPE_MLC:
+		if (nvm_init_mlc_tbl(dev, grp)) {
+			ret = -ENOMEM;
+			goto err_fmtype;
+		}
+		break;
+	default:
+		pr_err("nvm: flash type not supported\n");
+		ret = -EINVAL;
+		goto err_fmtype;
+	}
+
 	INIT_LIST_HEAD(&dev->online_targets);
 	mutex_init(&dev->mlock);
 	spin_lock_init(&dev->lock);
 
 	return 0;
+err_fmtype:
+	kfree(dev->lun_map);
+	return ret;
 }
 
 static void nvm_free(struct nvm_dev *dev)
@@ -576,6 +583,7 @@ static void nvm_free(struct nvm_dev *dev)
 		dev->mt->unregister_mgr(dev);
 
 	kfree(dev->lptbl);
+	kfree(dev->lun_map);
 }
 
 static int nvm_init(struct nvm_dev *dev)
@@ -705,7 +713,6 @@ void nvm_unregister(char *disk_name)
 	up_write(&nvm_lock);
 
 	nvm_exit(dev);
-	kfree(dev->lun_map);
 	kfree(dev);
 }
 EXPORT_SYMBOL(nvm_unregister);
-- 
2.1.4

[PATCH 07/28] lightnvm: calculate rrpc total blocks and sectors up front

[Matias Bjørling]

From: Wenwei Tao <ww.tao0320@gmail.com>

Calculate rrpc total blocks and sectors up front, make sense
to use them. For example, we use rrpc->nr_sects to calculate rrpc
area size, but it makes no sense if we don't initialize it up front,
since it would be zero until we finish rrpc luns init.

Signed-off-by: Wenwei Tao <ww.tao0320@gmail.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/rrpc.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index 3ab6495..c0e303c 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -1207,10 +1207,6 @@ static int rrpc_luns_init(struct rrpc *rrpc, int lun_begin, int lun_end)
 
 		INIT_WORK(&rlun->ws_gc, rrpc_lun_gc);
 		spin_lock_init(&rlun->lock);
-
-		rrpc->total_blocks += dev->blks_per_lun;
-		rrpc->nr_sects += dev->sec_per_lun;
-
 	}
 
 	return 0;
@@ -1388,6 +1384,8 @@ static void *rrpc_init(struct nvm_dev *dev, struct gendisk *tdisk,
 	INIT_WORK(&rrpc->ws_requeue, rrpc_requeue);
 
 	rrpc->nr_luns = lun_end - lun_begin + 1;
+	rrpc->total_blocks = (unsigned long)dev->blks_per_lun * rrpc->nr_luns;
+	rrpc->nr_sects = (unsigned long long)dev->sec_per_lun * rrpc->nr_luns;
 
 	/* simple round-robin strategy */
 	atomic_set(&rrpc->next_lun, -1);
-- 
2.1.4

[PATCH 08/28] lightnvm: store rrpc->soffset in device sector size

[Matias Bjørling]

From: Wenwei Tao <ww.tao0320@gmail.com>

Since we mainly use soffset in device sector size, we therefore store
this value in rrpc->soffset, instead of the offset in 512byte sector
size. This eliminates the "(ilog2(dev->sec_size) - 9)" calculation on
each I/O.

Signed-off-by: Wenwei Tao <ww.tao0320@gmail.com>
Updated patch description.
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/rrpc.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index c0e303c..3143b98 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -1039,11 +1039,8 @@ static int rrpc_map_init(struct rrpc *rrpc)
 {
 	struct nvm_dev *dev = rrpc->dev;
 	sector_t i;
-	u64 slba;
 	int ret;
 
-	slba = rrpc->soffset >> (ilog2(dev->sec_size) - 9);
-
 	rrpc->trans_map = vzalloc(sizeof(struct rrpc_addr) * rrpc->nr_sects);
 	if (!rrpc->trans_map)
 		return -ENOMEM;
@@ -1065,8 +1062,8 @@ static int rrpc_map_init(struct rrpc *rrpc)
 		return 0;
 
 	/* Bring up the mapping table from device */
-	ret = dev->ops->get_l2p_tbl(dev, slba, rrpc->nr_sects, rrpc_l2p_update,
-									rrpc);
+	ret = dev->ops->get_l2p_tbl(dev, rrpc->soffset, rrpc->nr_sects,
+					rrpc_l2p_update, rrpc);
 	if (ret) {
 		pr_err("nvm: rrpc: could not read L2P table.\n");
 		return -EINVAL;
@@ -1220,18 +1217,24 @@ static int rrpc_area_init(struct rrpc *rrpc, sector_t *begin)
 	struct nvm_dev *dev = rrpc->dev;
 	struct nvmm_type *mt = dev->mt;
 	sector_t size = rrpc->nr_sects * dev->sec_size;
+	int ret;
 
 	size >>= 9;
 
-	return mt->get_area(dev, begin, size);
+	ret = mt->get_area(dev, begin, size);
+	if (!ret)
+		*begin >>= (ilog2(dev->sec_size) - 9);
+
+	return ret;
 }
 
 static void rrpc_area_free(struct rrpc *rrpc)
 {
 	struct nvm_dev *dev = rrpc->dev;
 	struct nvmm_type *mt = dev->mt;
+	sector_t begin = rrpc->soffset << (ilog2(dev->sec_size) - 9);
 
-	mt->put_area(dev, rrpc->soffset);
+	mt->put_area(dev, begin);
 }
 
 static void rrpc_free(struct rrpc *rrpc)
-- 
2.1.4

[PATCH 09/28] lightnvm: rename nvm_targets to nvm_tgt_type

[Matias Bjørling]

From: "Simon A. F. Lund" <slund@cnexlabs.com>

The functions nvm_register_target(), nvm_unregister_target() and
associated list refers to a target type that is being registered by a
target type module. Rename nvm_*_targets() to nvm_*_tgt_type(), so that
the intension is clear.

This enables target instances to use the _nvm_*_targets() naming.

Signed-off-by: Simon A. F. Lund <slund@cnexlabs.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c  | 16 ++++++++--------
 drivers/lightnvm/rrpc.c  |  4 ++--
 include/linux/lightnvm.h |  4 ++--
 3 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 74fb049..240b473 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -30,7 +30,7 @@
 #include <linux/sched/sysctl.h>
 #include <uapi/linux/lightnvm.h>
 
-static LIST_HEAD(nvm_targets);
+static LIST_HEAD(nvm_tgt_types);
 static LIST_HEAD(nvm_mgrs);
 static LIST_HEAD(nvm_devices);
 static DECLARE_RWSEM(nvm_lock);
@@ -39,14 +39,14 @@ static struct nvm_tgt_type *nvm_find_target_type(const char *name)
 {
 	struct nvm_tgt_type *tt;
 
-	list_for_each_entry(tt, &nvm_targets, list)
+	list_for_each_entry(tt, &nvm_tgt_types, list)
 		if (!strcmp(name, tt->name))
 			return tt;
 
 	return NULL;
 }
 
-int nvm_register_target(struct nvm_tgt_type *tt)
+int nvm_register_tgt_type(struct nvm_tgt_type *tt)
 {
 	int ret = 0;
 
@@ -54,14 +54,14 @@ int nvm_register_target(struct nvm_tgt_type *tt)
 	if (nvm_find_target_type(tt->name))
 		ret = -EEXIST;
 	else
-		list_add(&tt->list, &nvm_targets);
+		list_add(&tt->list, &nvm_tgt_types);
 	up_write(&nvm_lock);
 
 	return ret;
 }
-EXPORT_SYMBOL(nvm_register_target);
+EXPORT_SYMBOL(nvm_register_tgt_type);
 
-void nvm_unregister_target(struct nvm_tgt_type *tt)
+void nvm_unregister_tgt_type(struct nvm_tgt_type *tt)
 {
 	if (!tt)
 		return;
@@ -70,7 +70,7 @@ void nvm_unregister_target(struct nvm_tgt_type *tt)
 	list_del(&tt->list);
 	up_write(&nvm_lock);
 }
-EXPORT_SYMBOL(nvm_unregister_target);
+EXPORT_SYMBOL(nvm_unregister_tgt_type);
 
 void *nvm_dev_dma_alloc(struct nvm_dev *dev, gfp_t mem_flags,
 							dma_addr_t *dma_handler)
@@ -1020,7 +1020,7 @@ static long nvm_ioctl_info(struct file *file, void __user *arg)
 	info->version[2] = NVM_VERSION_PATCH;
 
 	down_write(&nvm_lock);
-	list_for_each_entry(tt, &nvm_targets, list) {
+	list_for_each_entry(tt, &nvm_tgt_types, list) {
 		struct nvm_ioctl_info_tgt *tgt = &info->tgts[tgt_iter];
 
 		tgt->version[0] = tt->version[0];
diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index 3143b98..c7fef71 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -1469,12 +1469,12 @@ static struct nvm_tgt_type tt_rrpc = {
 
 static int __init rrpc_module_init(void)
 {
-	return nvm_register_target(&tt_rrpc);
+	return nvm_register_tgt_type(&tt_rrpc);
 }
 
 static void rrpc_module_exit(void)
 {
-	nvm_unregister_target(&tt_rrpc);
+	nvm_unregister_tgt_type(&tt_rrpc);
 }
 
 module_init(rrpc_module_init);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index dacaa28..497da91 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -453,8 +453,8 @@ struct nvm_tgt_type {
 	struct list_head list;
 };
 
-extern int nvm_register_target(struct nvm_tgt_type *);
-extern void nvm_unregister_target(struct nvm_tgt_type *);
+extern int nvm_register_tgt_type(struct nvm_tgt_type *);
+extern void nvm_unregister_tgt_type(struct nvm_tgt_type *);
 
 extern void *nvm_dev_dma_alloc(struct nvm_dev *, gfp_t, dma_addr_t *);
 extern void nvm_dev_dma_free(struct nvm_dev *, void *, dma_addr_t);
-- 
2.1.4

[PATCH 10/28] lightnvm: refactor dev->online_target to global nvm_targets

[Matias Bjørling]

From: "Simon A. F. Lund" <slund@cnexlabs.com>

A target name must be unique. However, a per-device registration of
targets is maintained on a dev->online_targets list, with a per-device
search for targets upon registration.

This results in a name collision when two targets, with the same name,
are created on two different targets, where the per-device list is not
shared.

Signed-off-by: Simon A. F. Lund <slund@cnexlabs.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c  | 47 +++++++++++++++++++++++++----------------------
 include/linux/lightnvm.h |  1 -
 2 files changed, 25 insertions(+), 23 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 240b473..0296223 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -33,8 +33,20 @@
 static LIST_HEAD(nvm_tgt_types);
 static LIST_HEAD(nvm_mgrs);
 static LIST_HEAD(nvm_devices);
+static LIST_HEAD(nvm_targets);
 static DECLARE_RWSEM(nvm_lock);
 
+static struct nvm_target *nvm_find_target(const char *name)
+{
+	struct nvm_target *tgt;
+
+	list_for_each_entry(tgt, &nvm_targets, list)
+		if (!strcmp(name, tgt->disk->disk_name))
+			return tgt;
+
+	return NULL;
+}
+
 static struct nvm_tgt_type *nvm_find_target_type(const char *name)
 {
 	struct nvm_tgt_type *tt;
@@ -564,7 +576,6 @@ static int nvm_core_init(struct nvm_dev *dev)
 		goto err_fmtype;
 	}
 
-	INIT_LIST_HEAD(&dev->online_targets);
 	mutex_init(&dev->mlock);
 	spin_lock_init(&dev->lock);
 
@@ -744,12 +755,11 @@ static int nvm_create_target(struct nvm_dev *dev,
 		return -EINVAL;
 	}
 
-	list_for_each_entry(t, &dev->online_targets, list) {
-		if (!strcmp(create->tgtname, t->disk->disk_name)) {
-			pr_err("nvm: target name already exists.\n");
-			up_write(&nvm_lock);
-			return -EINVAL;
-		}
+	t = nvm_find_target(create->tgtname);
+	if (t) {
+		pr_err("nvm: target name already exists.\n");
+		up_write(&nvm_lock);
+		return -EINVAL;
 	}
 	up_write(&nvm_lock);
 
@@ -789,7 +799,7 @@ static int nvm_create_target(struct nvm_dev *dev,
 	t->disk = tdisk;
 
 	down_write(&nvm_lock);
-	list_add_tail(&t->list, &dev->online_targets);
+	list_add_tail(&t->list, &nvm_targets);
 	up_write(&nvm_lock);
 
 	return 0;
@@ -852,26 +862,19 @@ static int __nvm_configure_create(struct nvm_ioctl_create *create)
 
 static int __nvm_configure_remove(struct nvm_ioctl_remove *remove)
 {
-	struct nvm_target *t = NULL;
-	struct nvm_dev *dev;
-	int ret = -1;
+	struct nvm_target *t;
 
 	down_write(&nvm_lock);
-	list_for_each_entry(dev, &nvm_devices, devices)
-		list_for_each_entry(t, &dev->online_targets, list) {
-			if (!strcmp(remove->tgtname, t->disk->disk_name)) {
-				nvm_remove_target(t);
-				ret = 0;
-				break;
-			}
-		}
-	up_write(&nvm_lock);
-
-	if (ret) {
+	t = nvm_find_target(remove->tgtname);
+	if (!t) {
 		pr_err("nvm: target \"%s\" doesn't exist.\n", remove->tgtname);
+		up_write(&nvm_lock);
 		return -EINVAL;
 	}
 
+	nvm_remove_target(t);
+	up_write(&nvm_lock);
+
 	return 0;
 }
 
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 497da91..5eabdba 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -308,7 +308,6 @@ struct nvm_dev {
 	struct nvm_dev_ops *ops;
 
 	struct list_head devices;
-	struct list_head online_targets;
 
 	/* Media manager */
 	struct nvmm_type *mt;
-- 
2.1.4

[PATCH 11/28] lightnvm: introduce nvm_for_each_lun_ppa() macro

[Matias Bjørling]

Users that wish to iterate all luns on a device. Must create a
struct ppa_addr and separate iterators for channels and luns. To set the
iterators, two loops are required, one to iterate channels, and another
to iterate luns. This leads to decrease in readability.

Introduce nvm_for_each_lun_ppa, which implements the nested loop and
sets ppa, channel, and lun variable for each loop body, eliminating
the boilerplate code.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/sysblk.c | 56 +++++++++++++++++++----------------------------
 include/linux/lightnvm.h  |  7 ++++++
 2 files changed, 29 insertions(+), 34 deletions(-)

diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index 7fce588..a48093d 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -631,33 +631,28 @@ static int nvm_fact_get_blks(struct nvm_dev *dev, struct ppa_addr *erase_list,
 
 	while (!done) {
 		done = 1;
-		for (ch = 0; ch < dev->nr_chnls; ch++) {
-			for (lun = 0; lun < dev->luns_per_chnl; lun++) {
-				idx = factory_blk_offset(dev, ch, lun);
-				offset = &f->blks[idx];
+		nvm_for_each_lun_ppa(dev, ppa, ch, lun) {
+			idx = factory_blk_offset(dev, ch, lun);
+			offset = &f->blks[idx];
 
-				blkid = find_first_zero_bit(offset,
-							dev->blks_per_lun);
-				if (blkid >= dev->blks_per_lun)
-					continue;
-				set_bit(blkid, offset);
+			blkid = find_first_zero_bit(offset,
+						dev->blks_per_lun);
+			if (blkid >= dev->blks_per_lun)
+				continue;
+			set_bit(blkid, offset);
 
-				ppa.ppa = 0;
-				ppa.g.ch = ch;
-				ppa.g.lun = lun;
-				ppa.g.blk = blkid;
-				pr_debug("nvm: erase ppa (%u %u %u)\n",
-								ppa.g.ch,
-								ppa.g.lun,
-								ppa.g.blk);
+			ppa.g.blk = blkid;
+			pr_debug("nvm: erase ppa (%u %u %u)\n",
+							ppa.g.ch,
+							ppa.g.lun,
+							ppa.g.blk);
 
-				erase_list[ppa_cnt] = ppa;
-				ppa_cnt++;
-				done = 0;
+			erase_list[ppa_cnt] = ppa;
+			ppa_cnt++;
+			done = 0;
 
-				if (ppa_cnt == max_ppas)
-					return ppa_cnt;
-			}
+			if (ppa_cnt == max_ppas)
+				return ppa_cnt;
 		}
 	}
 
@@ -684,17 +679,10 @@ static int nvm_fact_select_blks(struct nvm_dev *dev, struct factory_blks *f)
 	int ch, lun, ret;
 	struct ppa_addr ppa;
 
-	ppa.ppa = 0;
-	for (ch = 0; ch < dev->nr_chnls; ch++) {
-		for (lun = 0; lun < dev->luns_per_chnl; lun++) {
-			ppa.g.ch = ch;
-			ppa.g.lun = lun;
-
-			ret = nvm_fact_get_bb_tbl(dev, ppa, nvm_factory_blks,
-									f);
-			if (ret)
-				return ret;
-		}
+	nvm_for_each_lun_ppa(dev, ppa, ch, lun) {
+		ret = nvm_fact_get_bb_tbl(dev, ppa, nvm_factory_blks, f);
+		if (ret)
+			return ret;
 	}
 
 	return 0;
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 5eabdba..3f25635 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -559,6 +559,13 @@ extern int nvm_update_sysblock(struct nvm_dev *, struct nvm_sb_info *);
 extern int nvm_init_sysblock(struct nvm_dev *, struct nvm_sb_info *);
 
 extern int nvm_dev_factory(struct nvm_dev *, int flags);
+
+#define nvm_for_each_lun_ppa(dev, ppa, chid, lunid)			\
+	for ((chid) = 0, (ppa).ppa = 0; (chid) < (dev)->nr_chnls;	\
+					(chid)++, (ppa).g.ch = (chid))	\
+		for ((lunid) = 0; (lunid) < (dev)->luns_per_chnl;	\
+					(lunid)++, (ppa).g.lun = (lunid))
+
 #else /* CONFIG_NVM */
 struct nvm_dev_ops;
 
-- 
2.1.4

[PATCH 12/28] lightnvm: refactor device ops->get_bb_tbl()

[Matias Bjørling]

The device ops->get_bb_tbl() takes a callback, that allows the caller
to use its own callback function to update its data structures in the
returning function.

This makes it difficult to send parameters to the callback, and usually
is circumvented by small private structures, that both carry the callers
state and any flags needed to fulfill the update.

Refactor ops->get_bb_tbl() to fill a data buffer with the status of the
blocks returned, and let the user call the callback function manually.
That will provide the necessary flags and data structures and simplify
the logic around ops->get_bb_tbl().

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c      |   8 +++
 drivers/lightnvm/gennvm.c    |  29 ++++++---
 drivers/lightnvm/sysblk.c    | 146 +++++++++++++++++++++++--------------------
 drivers/nvme/host/lightnvm.c |   6 +-
 include/linux/lightnvm.h     |   6 +-
 5 files changed, 111 insertions(+), 84 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 0296223..e6d7a98 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -467,6 +467,14 @@ int nvm_bb_tbl_fold(struct nvm_dev *dev, u8 *blks, int nr_blks)
 }
 EXPORT_SYMBOL(nvm_bb_tbl_fold);
 
+int nvm_get_bb_tbl(struct nvm_dev *dev, struct ppa_addr ppa, u8 *blks)
+{
+	ppa = generic_to_dev_addr(dev, ppa);
+
+	return dev->ops->get_bb_tbl(dev, ppa, blks);
+}
+EXPORT_SYMBOL(nvm_get_bb_tbl);
+
 static int nvm_init_slc_tbl(struct nvm_dev *dev, struct nvm_id_group *grp)
 {
 	int i;
diff --git a/drivers/lightnvm/gennvm.c b/drivers/lightnvm/gennvm.c
index 6096077..9c6b141 100644
--- a/drivers/lightnvm/gennvm.c
+++ b/drivers/lightnvm/gennvm.c
@@ -129,10 +129,10 @@ static int gennvm_luns_init(struct nvm_dev *dev, struct gen_nvm *gn)
 	return 0;
 }
 
-static int gennvm_block_bb(struct nvm_dev *dev, struct ppa_addr ppa,
-					u8 *blks, int nr_blks, void *private)
+static int gennvm_block_bb(struct gen_nvm *gn, struct ppa_addr ppa,
+							u8 *blks, int nr_blks)
 {
-	struct gen_nvm *gn = private;
+	struct nvm_dev *dev = gn->dev;
 	struct gen_lun *lun;
 	struct nvm_block *blk;
 	int i;
@@ -219,13 +219,21 @@ static int gennvm_blocks_init(struct nvm_dev *dev, struct gen_nvm *gn)
 	struct gen_lun *lun;
 	struct nvm_block *block;
 	sector_t lun_iter, blk_iter, cur_block_id = 0;
-	int ret;
+	int ret, nr_blks;
+	u8 *blks;
+
+	nr_blks = dev->blks_per_lun * dev->plane_mode;
+	blks = kmalloc(nr_blks, GFP_KERNEL);
+	if (!blks)
+		return -ENOMEM;
 
 	gennvm_for_each_lun(gn, lun, lun_iter) {
 		lun->vlun.blocks = vzalloc(sizeof(struct nvm_block) *
 							dev->blks_per_lun);
-		if (!lun->vlun.blocks)
+		if (!lun->vlun.blocks) {
+			kfree(blks);
 			return -ENOMEM;
+		}
 
 		for (blk_iter = 0; blk_iter < dev->blks_per_lun; blk_iter++) {
 			block = &lun->vlun.blocks[blk_iter];
@@ -250,12 +258,14 @@ static int gennvm_blocks_init(struct nvm_dev *dev, struct gen_nvm *gn)
 			ppa.ppa = 0;
 			ppa.g.ch = lun->vlun.chnl_id;
 			ppa.g.lun = lun->vlun.id;
-			ppa = generic_to_dev_addr(dev, ppa);
 
-			ret = dev->ops->get_bb_tbl(dev, ppa,
-							gennvm_block_bb, gn);
+			ret = nvm_get_bb_tbl(dev, ppa, blks);
 			if (ret)
-				pr_err("gennvm: could not read BB table\n");
+				pr_err("gennvm: could not get BB table\n");
+
+			ret = gennvm_block_bb(gn, ppa, blks, nr_blks);
+			if (ret)
+				pr_err("gennvm: BB table map failed\n");
 		}
 	}
 
@@ -268,6 +278,7 @@ static int gennvm_blocks_init(struct nvm_dev *dev, struct gen_nvm *gn)
 		}
 	}
 
+	kfree(blks);
 	return 0;
 }
 
diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index a48093d..dc99c0a 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -93,10 +93,45 @@ void nvm_setup_sysblk_scan(struct nvm_dev *dev, struct sysblk_scan *s,
 	s->nr_rows = nvm_setup_sysblks(dev, sysblk_ppas);
 }
 
+static int sysblk_get_free_blks(struct nvm_dev *dev, struct ppa_addr ppa,
+					u8 *blks, int nr_blks,
+					struct sysblk_scan *s)
+{
+	struct ppa_addr *sppa;
+	int i, blkid = 0;
+
+	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
+	if (nr_blks < 0)
+		return nr_blks;
+
+	for (i = 0; i < nr_blks; i++) {
+		if (blks[i] == NVM_BLK_T_HOST)
+			return -EEXIST;
+
+		if (blks[i] != NVM_BLK_T_FREE)
+			continue;
+
+		sppa = &s->ppas[scan_ppa_idx(s->row, blkid)];
+		sppa->g.ch = ppa.g.ch;
+		sppa->g.lun = ppa.g.lun;
+		sppa->g.blk = i;
+		s->nr_ppas++;
+		blkid++;
+
+		pr_debug("nvm: use (%u %u %u) as sysblk\n",
+					sppa->g.ch, sppa->g.lun, sppa->g.blk);
+		if (blkid > MAX_BLKS_PR_SYSBLK - 1)
+			return 0;
+	}
+
+	pr_err("nvm: sysblk failed get sysblk\n");
+	return -EINVAL;
+}
+
 static int sysblk_get_host_blks(struct nvm_dev *dev, struct ppa_addr ppa,
-					u8 *blks, int nr_blks, void *private)
+					u8 *blks, int nr_blks,
+					struct sysblk_scan *s)
 {
-	struct sysblk_scan *s = private;
 	int i, nr_sysblk = 0;
 
 	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
@@ -123,26 +158,42 @@ static int sysblk_get_host_blks(struct nvm_dev *dev, struct ppa_addr ppa,
 }
 
 static int nvm_get_all_sysblks(struct nvm_dev *dev, struct sysblk_scan *s,
-				struct ppa_addr *ppas, nvm_bb_update_fn *fn)
+				struct ppa_addr *ppas, int get_free)
 {
-	struct ppa_addr dppa;
-	int i, ret = 0;
+	int i, nr_blks, ret = 0;
+	u8 *blks;
 
 	s->nr_ppas = 0;
+	nr_blks = dev->blks_per_lun * dev->plane_mode;
+
+	blks = kmalloc(nr_blks, GFP_KERNEL);
+	if (!blks)
+		return -ENOMEM;
 
 	for (i = 0; i < s->nr_rows; i++) {
-		dppa = generic_to_dev_addr(dev, ppas[i]);
 		s->row = i;
 
-		ret = dev->ops->get_bb_tbl(dev, dppa, fn, s);
+		ret = nvm_get_bb_tbl(dev, ppas[i], blks);
 		if (ret) {
 			pr_err("nvm: failed bb tbl for ppa (%u %u)\n",
 							ppas[i].g.ch,
 							ppas[i].g.blk);
-			return ret;
+			goto err_get;
 		}
+
+		if (get_free)
+			ret = sysblk_get_free_blks(dev, ppas[i], blks, nr_blks,
+									s);
+		else
+			ret = sysblk_get_host_blks(dev, ppas[i], blks, nr_blks,
+									s);
+
+		if (ret)
+			goto err_get;
 	}
 
+err_get:
+	kfree(blks);
 	return ret;
 }
 
@@ -239,41 +290,6 @@ static int nvm_set_bb_tbl(struct nvm_dev *dev, struct sysblk_scan *s, int type)
 	return 0;
 }
 
-static int sysblk_get_free_blks(struct nvm_dev *dev, struct ppa_addr ppa,
-					u8 *blks, int nr_blks, void *private)
-{
-	struct sysblk_scan *s = private;
-	struct ppa_addr *sppa;
-	int i, blkid = 0;
-
-	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
-	if (nr_blks < 0)
-		return nr_blks;
-
-	for (i = 0; i < nr_blks; i++) {
-		if (blks[i] == NVM_BLK_T_HOST)
-			return -EEXIST;
-
-		if (blks[i] != NVM_BLK_T_FREE)
-			continue;
-
-		sppa = &s->ppas[scan_ppa_idx(s->row, blkid)];
-		sppa->g.ch = ppa.g.ch;
-		sppa->g.lun = ppa.g.lun;
-		sppa->g.blk = i;
-		s->nr_ppas++;
-		blkid++;
-
-		pr_debug("nvm: use (%u %u %u) as sysblk\n",
-					sppa->g.ch, sppa->g.lun, sppa->g.blk);
-		if (blkid > MAX_BLKS_PR_SYSBLK - 1)
-			return 0;
-	}
-
-	pr_err("nvm: sysblk failed get sysblk\n");
-	return -EINVAL;
-}
-
 static int nvm_write_and_verify(struct nvm_dev *dev, struct nvm_sb_info *info,
 							struct sysblk_scan *s)
 {
@@ -393,7 +409,7 @@ int nvm_get_sysblock(struct nvm_dev *dev, struct nvm_sb_info *info)
 	nvm_setup_sysblk_scan(dev, &s, sysblk_ppas);
 
 	mutex_lock(&dev->mlock);
-	ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, sysblk_get_host_blks);
+	ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, 0);
 	if (ret)
 		goto err_sysblk;
 
@@ -453,7 +469,7 @@ int nvm_update_sysblock(struct nvm_dev *dev, struct nvm_sb_info *new)
 	nvm_setup_sysblk_scan(dev, &s, sysblk_ppas);
 
 	mutex_lock(&dev->mlock);
-	ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, sysblk_get_host_blks);
+	ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, 0);
 	if (ret)
 		goto err_sysblk;
 
@@ -551,7 +567,7 @@ int nvm_init_sysblock(struct nvm_dev *dev, struct nvm_sb_info *info)
 	nvm_setup_sysblk_scan(dev, &s, sysblk_ppas);
 
 	mutex_lock(&dev->mlock);
-	ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, sysblk_get_free_blks);
+	ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, 1);
 	if (ret)
 		goto err_mark;
 
@@ -587,9 +603,9 @@ static unsigned int factory_blk_offset(struct nvm_dev *dev, int ch, int lun)
 }
 
 static int nvm_factory_blks(struct nvm_dev *dev, struct ppa_addr ppa,
-					u8 *blks, int nr_blks, void *private)
+					u8 *blks, int nr_blks,
+					struct factory_blks *f)
 {
-	struct factory_blks *f = private;
 	int i, lunoff;
 
 	nr_blks = nvm_bb_tbl_fold(dev, blks, nr_blks);
@@ -659,32 +675,29 @@ static int nvm_fact_get_blks(struct nvm_dev *dev, struct ppa_addr *erase_list,
 	return ppa_cnt;
 }
 
-static int nvm_fact_get_bb_tbl(struct nvm_dev *dev, struct ppa_addr ppa,
-					nvm_bb_update_fn *fn, void *priv)
-{
-	struct ppa_addr dev_ppa;
-	int ret;
-
-	dev_ppa = generic_to_dev_addr(dev, ppa);
-
-	ret = dev->ops->get_bb_tbl(dev, dev_ppa, fn, priv);
-	if (ret)
-		pr_err("nvm: failed bb tbl for ch%u lun%u\n",
-							ppa.g.ch, ppa.g.blk);
-	return ret;
-}
-
 static int nvm_fact_select_blks(struct nvm_dev *dev, struct factory_blks *f)
 {
-	int ch, lun, ret;
 	struct ppa_addr ppa;
+	int ch, lun, nr_blks, ret;
+	u8 *blks;
+
+	nr_blks = dev->blks_per_lun * dev->plane_mode;
+	blks = kmalloc(nr_blks, GFP_KERNEL);
+	if (!blks)
+		return -ENOMEM;
 
 	nvm_for_each_lun_ppa(dev, ppa, ch, lun) {
-		ret = nvm_fact_get_bb_tbl(dev, ppa, nvm_factory_blks, f);
+		ret = nvm_get_bb_tbl(dev, ppa, blks);
+		if (ret)
+			pr_err("nvm: failed bb tbl for ch%u lun%u\n",
+							ppa.g.ch, ppa.g.blk);
+
+		ret = nvm_factory_blks(dev, ppa, blks, nr_blks, f);
 		if (ret)
 			return ret;
 	}
 
+	kfree(blks);
 	return 0;
 }
 
@@ -722,8 +735,7 @@ int nvm_dev_factory(struct nvm_dev *dev, int flags)
 	if (flags & NVM_FACTORY_RESET_HOST_BLKS) {
 		nvm_setup_sysblk_scan(dev, &s, sysblk_ppas);
 		mutex_lock(&dev->mlock);
-		ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas,
-							sysblk_get_host_blks);
+		ret = nvm_get_all_sysblks(dev, &s, sysblk_ppas, 0);
 		if (!ret)
 			ret = nvm_set_bb_tbl(dev, &s, NVM_BLK_T_FREE);
 		mutex_unlock(&dev->mlock);
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index d289980..45e3511 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -388,7 +388,7 @@ out:
 }
 
 static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
-				nvm_bb_update_fn *update_bbtbl, void *priv)
+								u8 *blks)
 {
 	struct request_queue *q = nvmdev->q;
 	struct nvme_ns *ns = q->queuedata;
@@ -435,9 +435,7 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 		goto out;
 	}
 
-	ppa = dev_to_generic_addr(nvmdev, ppa);
-	ret = update_bbtbl(nvmdev, ppa, bb_tbl->blk, nr_blks, priv);
-
+	memcpy(blks, bb_tbl->blk, nvmdev->blks_per_lun * nvmdev->plane_mode);
 out:
 	kfree(bb_tbl);
 	return ret;
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 3f25635..16d4f2e 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -41,13 +41,10 @@ struct nvm_id;
 struct nvm_dev;
 
 typedef int (nvm_l2p_update_fn)(u64, u32, __le64 *, void *);
-typedef int (nvm_bb_update_fn)(struct nvm_dev *, struct ppa_addr, u8 *, int,
-								void *);
 typedef int (nvm_id_fn)(struct nvm_dev *, struct nvm_id *);
 typedef int (nvm_get_l2p_tbl_fn)(struct nvm_dev *, u64, u32,
 				nvm_l2p_update_fn *, void *);
-typedef int (nvm_op_bb_tbl_fn)(struct nvm_dev *, struct ppa_addr,
-				nvm_bb_update_fn *, void *);
+typedef int (nvm_op_bb_tbl_fn)(struct nvm_dev *, struct ppa_addr, u8 *);
 typedef int (nvm_op_set_bb_fn)(struct nvm_dev *, struct nvm_rq *, int);
 typedef int (nvm_submit_io_fn)(struct nvm_dev *, struct nvm_rq *);
 typedef int (nvm_erase_blk_fn)(struct nvm_dev *, struct nvm_rq *);
@@ -539,6 +536,7 @@ extern int nvm_submit_ppa(struct nvm_dev *, struct ppa_addr *, int, int, int,
 extern int nvm_submit_ppa_list(struct nvm_dev *, struct ppa_addr *, int, int,
 							int, void *, int);
 extern int nvm_bb_tbl_fold(struct nvm_dev *, u8 *, int);
+extern int nvm_get_bb_tbl(struct nvm_dev *, struct ppa_addr, u8 *);
 
 /* sysblk.c */
 #define NVM_SYSBLK_MAGIC 0x4E564D53 /* "NVMS" */
-- 
2.1.4

[PATCH 13/28] lightnvm: remove struct factory_blks

[Matias Bjørling]

Now that device ops->get_bb_table no longer uses a callback, the
struct factory_blks can be removed.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/sysblk.c | 62 +++++++++++++++++++++--------------------------
 1 file changed, 28 insertions(+), 34 deletions(-)

diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index dc99c0a..bca6902 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -582,29 +582,23 @@ err_mark:
 	return ret;
 }
 
-struct factory_blks {
-	struct nvm_dev *dev;
-	int flags;
-	unsigned long *blks;
-};
-
 static int factory_nblks(int nblks)
 {
 	/* Round up to nearest BITS_PER_LONG */
 	return (nblks + (BITS_PER_LONG - 1)) & ~(BITS_PER_LONG - 1);
 }
 
-static unsigned int factory_blk_offset(struct nvm_dev *dev, int ch, int lun)
+static unsigned int factory_blk_offset(struct nvm_dev *dev, struct ppa_addr ppa)
 {
 	int nblks = factory_nblks(dev->blks_per_lun);
 
-	return ((ch * dev->luns_per_chnl * nblks) + (lun * nblks)) /
+	return ((ppa.g.ch * dev->luns_per_chnl * nblks) + (ppa.g.lun * nblks)) /
 								BITS_PER_LONG;
 }
 
 static int nvm_factory_blks(struct nvm_dev *dev, struct ppa_addr ppa,
 					u8 *blks, int nr_blks,
-					struct factory_blks *f)
+					unsigned long *blk_bitmap, int flags)
 {
 	int i, lunoff;
 
@@ -612,25 +606,25 @@ static int nvm_factory_blks(struct nvm_dev *dev, struct ppa_addr ppa,
 	if (nr_blks < 0)
 		return nr_blks;
 
-	lunoff = factory_blk_offset(dev, ppa.g.ch, ppa.g.lun);
+	lunoff = factory_blk_offset(dev, ppa);
 
 	/* non-set bits correspond to the block must be erased */
 	for (i = 0; i < nr_blks; i++) {
 		switch (blks[i]) {
 		case NVM_BLK_T_FREE:
-			if (f->flags & NVM_FACTORY_ERASE_ONLY_USER)
-				set_bit(i, &f->blks[lunoff]);
+			if (flags & NVM_FACTORY_ERASE_ONLY_USER)
+				set_bit(i, &blk_bitmap[lunoff]);
 			break;
 		case NVM_BLK_T_HOST:
-			if (!(f->flags & NVM_FACTORY_RESET_HOST_BLKS))
-				set_bit(i, &f->blks[lunoff]);
+			if (!(flags & NVM_FACTORY_RESET_HOST_BLKS))
+				set_bit(i, &blk_bitmap[lunoff]);
 			break;
 		case NVM_BLK_T_GRWN_BAD:
-			if (!(f->flags & NVM_FACTORY_RESET_GRWN_BBLKS))
-				set_bit(i, &f->blks[lunoff]);
+			if (!(flags & NVM_FACTORY_RESET_GRWN_BBLKS))
+				set_bit(i, &blk_bitmap[lunoff]);
 			break;
 		default:
-			set_bit(i, &f->blks[lunoff]);
+			set_bit(i, &blk_bitmap[lunoff]);
 			break;
 		}
 	}
@@ -639,7 +633,7 @@ static int nvm_factory_blks(struct nvm_dev *dev, struct ppa_addr ppa,
 }
 
 static int nvm_fact_get_blks(struct nvm_dev *dev, struct ppa_addr *erase_list,
-					int max_ppas, struct factory_blks *f)
+					int max_ppas, unsigned long *blk_bitmap)
 {
 	struct ppa_addr ppa;
 	int ch, lun, blkid, idx, done = 0, ppa_cnt = 0;
@@ -648,8 +642,8 @@ static int nvm_fact_get_blks(struct nvm_dev *dev, struct ppa_addr *erase_list,
 	while (!done) {
 		done = 1;
 		nvm_for_each_lun_ppa(dev, ppa, ch, lun) {
-			idx = factory_blk_offset(dev, ch, lun);
-			offset = &f->blks[idx];
+			idx = factory_blk_offset(dev, ppa);
+			offset = &blk_bitmap[idx];
 
 			blkid = find_first_zero_bit(offset,
 						dev->blks_per_lun);
@@ -675,10 +669,11 @@ static int nvm_fact_get_blks(struct nvm_dev *dev, struct ppa_addr *erase_list,
 	return ppa_cnt;
 }
 
-static int nvm_fact_select_blks(struct nvm_dev *dev, struct factory_blks *f)
+static int nvm_fact_select_blks(struct nvm_dev *dev, unsigned long *blk_bitmap,
+								int flags)
 {
 	struct ppa_addr ppa;
-	int ch, lun, nr_blks, ret;
+	int ch, lun, nr_blks, ret = 0;
 	u8 *blks;
 
 	nr_blks = dev->blks_per_lun * dev->plane_mode;
@@ -692,43 +687,42 @@ static int nvm_fact_select_blks(struct nvm_dev *dev, struct factory_blks *f)
 			pr_err("nvm: failed bb tbl for ch%u lun%u\n",
 							ppa.g.ch, ppa.g.blk);
 
-		ret = nvm_factory_blks(dev, ppa, blks, nr_blks, f);
+		ret = nvm_factory_blks(dev, ppa, blks, nr_blks, blk_bitmap,
+									flags);
 		if (ret)
-			return ret;
+			break;
 	}
 
 	kfree(blks);
-	return 0;
+	return ret;
 }
 
 int nvm_dev_factory(struct nvm_dev *dev, int flags)
 {
-	struct factory_blks f;
 	struct ppa_addr *ppas;
 	int ppa_cnt, ret = -ENOMEM;
 	int max_ppas = dev->ops->max_phys_sect / dev->nr_planes;
 	struct ppa_addr sysblk_ppas[MAX_SYSBLKS];
 	struct sysblk_scan s;
+	unsigned long *blk_bitmap;
 
-	f.blks = kzalloc(factory_nblks(dev->blks_per_lun) * dev->nr_luns,
+	blk_bitmap = kzalloc(factory_nblks(dev->blks_per_lun) * dev->nr_luns,
 								GFP_KERNEL);
-	if (!f.blks)
+	if (!blk_bitmap)
 		return ret;
 
 	ppas = kcalloc(max_ppas, sizeof(struct ppa_addr), GFP_KERNEL);
 	if (!ppas)
 		goto err_blks;
 
-	f.dev = dev;
-	f.flags = flags;
-
 	/* create list of blks to be erased */
-	ret = nvm_fact_select_blks(dev, &f);
+	ret = nvm_fact_select_blks(dev, blk_bitmap, flags);
 	if (ret)
 		goto err_ppas;
 
 	/* continue to erase until list of blks until empty */
-	while ((ppa_cnt = nvm_fact_get_blks(dev, ppas, max_ppas, &f)) > 0)
+	while ((ppa_cnt =
+			nvm_fact_get_blks(dev, ppas, max_ppas, blk_bitmap)) > 0)
 		nvm_erase_ppa(dev, ppas, ppa_cnt);
 
 	/* mark host reserved blocks free */
@@ -743,7 +737,7 @@ int nvm_dev_factory(struct nvm_dev *dev, int flags)
 err_ppas:
 	kfree(ppas);
 err_blks:
-	kfree(f.blks);
+	kfree(blk_bitmap);
 	return ret;
 }
 EXPORT_SYMBOL(nvm_dev_factory);
-- 
2.1.4

[PATCH 14/28] lightnvm: make nvm_set_rqd_ppalist() aware of vblks

[Matias Bjørling]

A virtual block enables a block to identify multiple physical blocks.
This is useful for metadata where a device media supports multiple
planes. In that case, a block, with multiple planes can be managed
as a single vblk. Reducing the metadata required by one forth.

nvm_set_rqd_ppalist() takes care of expanding a ppa_list with vblks
automatically. However, for some use-cases, where only a single physical
block is required, the ppa_list should not be expanded.

Therefore, add a vblk parameter to nvm_set_rqd_ppalist(), and only
expand the ppa_list if vblk is set.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c   | 31 +++++++++++++++++--------------
 drivers/lightnvm/sysblk.c |  2 +-
 include/linux/lightnvm.h  |  2 +-
 3 files changed, 19 insertions(+), 16 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index e6d7a98..de5db7b 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -251,33 +251,36 @@ void nvm_generic_to_addr_mode(struct nvm_dev *dev, struct nvm_rq *rqd)
 EXPORT_SYMBOL(nvm_generic_to_addr_mode);
 
 int nvm_set_rqd_ppalist(struct nvm_dev *dev, struct nvm_rq *rqd,
-					struct ppa_addr *ppas, int nr_ppas)
+				struct ppa_addr *ppas, int nr_ppas, int vblk)
 {
 	int i, plane_cnt, pl_idx;
 
-	if (dev->plane_mode == NVM_PLANE_SINGLE && nr_ppas == 1) {
-		rqd->nr_pages = 1;
+	if ((!vblk || dev->plane_mode == NVM_PLANE_SINGLE) && nr_ppas == 1) {
+		rqd->nr_pages = nr_ppas;
 		rqd->ppa_addr = ppas[0];
 
 		return 0;
 	}
 
-	plane_cnt = dev->plane_mode;
-	rqd->nr_pages = plane_cnt * nr_ppas;
-
-	if (dev->ops->max_phys_sect < rqd->nr_pages)
-		return -EINVAL;
-
+	rqd->nr_pages = nr_ppas;
 	rqd->ppa_list = nvm_dev_dma_alloc(dev, GFP_KERNEL, &rqd->dma_ppa_list);
 	if (!rqd->ppa_list) {
 		pr_err("nvm: failed to allocate dma memory\n");
 		return -ENOMEM;
 	}
 
-	for (pl_idx = 0; pl_idx < plane_cnt; pl_idx++) {
+	if (!vblk) {
+		for (i = 0; i < nr_ppas; i++)
+			rqd->ppa_list[i] = ppas[i];
+	} else {
+		plane_cnt = dev->plane_mode;
+		rqd->nr_pages *= plane_cnt;
+
 		for (i = 0; i < nr_ppas; i++) {
-			ppas[i].g.pl = pl_idx;
-			rqd->ppa_list[(pl_idx * nr_ppas) + i] = ppas[i];
+			for (pl_idx = 0; pl_idx < plane_cnt; pl_idx++) {
+				ppas[i].g.pl = pl_idx;
+				rqd->ppa_list[(pl_idx * nr_ppas) + i] = ppas[i];
+			}
 		}
 	}
 
@@ -304,7 +307,7 @@ int nvm_erase_ppa(struct nvm_dev *dev, struct ppa_addr *ppas, int nr_ppas)
 
 	memset(&rqd, 0, sizeof(struct nvm_rq));
 
-	ret = nvm_set_rqd_ppalist(dev, &rqd, ppas, nr_ppas);
+	ret = nvm_set_rqd_ppalist(dev, &rqd, ppas, nr_ppas, 1);
 	if (ret)
 		return ret;
 
@@ -420,7 +423,7 @@ int nvm_submit_ppa(struct nvm_dev *dev, struct ppa_addr *ppa, int nr_ppas,
 	int ret;
 
 	memset(&rqd, 0, sizeof(struct nvm_rq));
-	ret = nvm_set_rqd_ppalist(dev, &rqd, ppa, nr_ppas);
+	ret = nvm_set_rqd_ppalist(dev, &rqd, ppa, nr_ppas, 1);
 	if (ret)
 		return ret;
 
diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index bca6902..737fbc3 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -277,7 +277,7 @@ static int nvm_set_bb_tbl(struct nvm_dev *dev, struct sysblk_scan *s, int type)
 
 	memset(&rqd, 0, sizeof(struct nvm_rq));
 
-	nvm_set_rqd_ppalist(dev, &rqd, s->ppas, s->nr_ppas);
+	nvm_set_rqd_ppalist(dev, &rqd, s->ppas, s->nr_ppas, 1);
 	nvm_generic_to_addr_mode(dev, &rqd);
 
 	ret = dev->ops->set_bb_tbl(dev, &rqd, type);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 16d4f2e..9ae0b7c 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -526,7 +526,7 @@ extern int nvm_submit_io(struct nvm_dev *, struct nvm_rq *);
 extern void nvm_generic_to_addr_mode(struct nvm_dev *, struct nvm_rq *);
 extern void nvm_addr_to_generic_mode(struct nvm_dev *, struct nvm_rq *);
 extern int nvm_set_rqd_ppalist(struct nvm_dev *, struct nvm_rq *,
-							struct ppa_addr *, int);
+						struct ppa_addr *, int, int);
 extern void nvm_free_rqd_ppalist(struct nvm_dev *, struct nvm_rq *);
 extern int nvm_erase_ppa(struct nvm_dev *, struct ppa_addr *, int);
 extern int nvm_erase_blk(struct nvm_dev *, struct nvm_block *);
-- 
2.1.4

[PATCH 15/28] lightnvm: move responsibility for bad blk mgmt to target

[Matias Bjørling]

We move the responsibility of managing the persistent bad block table to
the target. The target may choose to mark a block bad or retry writing
to it. Never the less, it should be the target that makes the decision
and not the media manager.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/gennvm.c | 35 ++++++++++++++++-------------------
 1 file changed, 16 insertions(+), 19 deletions(-)

diff --git a/drivers/lightnvm/gennvm.c b/drivers/lightnvm/gennvm.c
index 9c6b141..89b880a 100644
--- a/drivers/lightnvm/gennvm.c
+++ b/drivers/lightnvm/gennvm.c
@@ -419,6 +419,9 @@ static void gennvm_blk_set_type(struct nvm_dev *dev, struct ppa_addr *ppa,
 	struct gen_lun *lun;
 	struct nvm_block *blk;
 
+	pr_debug("gennvm: ppa  (ch: %u lun: %u blk: %u pg: %u) -> %u\n",
+			ppa->g.ch, ppa->g.lun, ppa->g.blk, ppa->g.pg, type);
+
 	if (unlikely(ppa->g.ch > dev->nr_chnls ||
 					ppa->g.lun > dev->luns_per_chnl ||
 					ppa->g.blk > dev->blks_per_lun)) {
@@ -437,39 +440,33 @@ static void gennvm_blk_set_type(struct nvm_dev *dev, struct ppa_addr *ppa,
 	blk->state = type;
 }
 
-/* mark block bad. It is expected the target recover from the error. */
+/*
+ * mark block bad in gennvm. It is expected that the target recovers separately
+ */
 static void gennvm_mark_blk_bad(struct nvm_dev *dev, struct nvm_rq *rqd)
 {
-	int i;
-
-	if (!dev->ops->set_bb_tbl)
-		return;
-
-	if (dev->ops->set_bb_tbl(dev, rqd, 1))
-		return;
+	int bit = -1;
+	int max_secs = dev->ops->max_phys_sect;
+	void *comp_bits = &rqd->ppa_status;
 
 	nvm_addr_to_generic_mode(dev, rqd);
 
 	/* look up blocks and mark them as bad */
-	if (rqd->nr_pages > 1)
-		for (i = 0; i < rqd->nr_pages; i++)
-			gennvm_blk_set_type(dev, &rqd->ppa_list[i],
-						NVM_BLK_ST_BAD);
-	else
+	if (rqd->nr_pages == 1) {
 		gennvm_blk_set_type(dev, &rqd->ppa_addr, NVM_BLK_ST_BAD);
+		return;
+	}
+
+	while ((bit = find_next_bit(comp_bits, max_secs, bit + 1)) < max_secs)
+		gennvm_blk_set_type(dev, &rqd->ppa_list[bit], NVM_BLK_ST_BAD);
 }
 
 static void gennvm_end_io(struct nvm_rq *rqd)
 {
 	struct nvm_tgt_instance *ins = rqd->ins;
 
-	switch (rqd->error) {
-	case NVM_RSP_SUCCESS:
-	case NVM_RSP_ERR_EMPTYPAGE:
-		break;
-	case NVM_RSP_ERR_FAILWRITE:
+	if (rqd->error == NVM_RSP_ERR_FAILWRITE)
 		gennvm_mark_blk_bad(rqd->dev, rqd);
-	}
 
 	ins->tt->end_io(rqd);
 }
-- 
2.1.4

[PATCH 16/28] lightnvm: refactor set_bb_tbl for accepting ppa list

[Matias Bjørling]

The set_bb_tbl takes struct nvm_rq and only uses its ppa_list and
nr_pages internally. Instead, make these two variables explicit.
This allows a user to call it without initializing a struct nvm_rq
first.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/sysblk.c    | 2 +-
 drivers/nvme/host/lightnvm.c | 8 ++++----
 include/linux/lightnvm.h     | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index 737fbc3..b98ca19 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -280,7 +280,7 @@ static int nvm_set_bb_tbl(struct nvm_dev *dev, struct sysblk_scan *s, int type)
 	nvm_set_rqd_ppalist(dev, &rqd, s->ppas, s->nr_ppas, 1);
 	nvm_generic_to_addr_mode(dev, &rqd);
 
-	ret = dev->ops->set_bb_tbl(dev, &rqd, type);
+	ret = dev->ops->set_bb_tbl(dev, &rqd.ppa_addr, rqd.nr_pages, type);
 	nvm_free_rqd_ppalist(dev, &rqd);
 	if (ret) {
 		pr_err("nvm: sysblk failed bb mark\n");
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 45e3511..d426d95 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -441,8 +441,8 @@ out:
 	return ret;
 }
 
-static int nvme_nvm_set_bb_tbl(struct nvm_dev *nvmdev, struct nvm_rq *rqd,
-								int type)
+static int nvme_nvm_set_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr *ppas,
+							int nr_ppas, int type)
 {
 	struct nvme_ns *ns = nvmdev->q->queuedata;
 	struct nvme_nvm_command c = {};
@@ -450,8 +450,8 @@ static int nvme_nvm_set_bb_tbl(struct nvm_dev *nvmdev, struct nvm_rq *rqd,
 
 	c.set_bb.opcode = nvme_nvm_admin_set_bb_tbl;
 	c.set_bb.nsid = cpu_to_le32(ns->ns_id);
-	c.set_bb.spba = cpu_to_le64(rqd->ppa_addr.ppa);
-	c.set_bb.nlb = cpu_to_le16(rqd->nr_pages - 1);
+	c.set_bb.spba = cpu_to_le64(ppas->ppa);
+	c.set_bb.nlb = cpu_to_le16(nr_ppas - 1);
 	c.set_bb.value = type;
 
 	ret = nvme_submit_sync_cmd(ns->ctrl->admin_q, (struct nvme_command *)&c,
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 9ae0b7c..af72ca7 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -45,7 +45,7 @@ typedef int (nvm_id_fn)(struct nvm_dev *, struct nvm_id *);
 typedef int (nvm_get_l2p_tbl_fn)(struct nvm_dev *, u64, u32,
 				nvm_l2p_update_fn *, void *);
 typedef int (nvm_op_bb_tbl_fn)(struct nvm_dev *, struct ppa_addr, u8 *);
-typedef int (nvm_op_set_bb_fn)(struct nvm_dev *, struct nvm_rq *, int);
+typedef int (nvm_op_set_bb_fn)(struct nvm_dev *, struct ppa_addr *, int, int);
 typedef int (nvm_submit_io_fn)(struct nvm_dev *, struct nvm_rq *);
 typedef int (nvm_erase_blk_fn)(struct nvm_dev *, struct nvm_rq *);
 typedef void *(nvm_create_dma_pool_fn)(struct nvm_dev *, char *);
-- 
2.1.4

[PATCH 17/28] lightnvm: fix out of bound ppa lun id on bb tbl

[Matias Bjørling]

The ppa configured for retrieving the bad block table uses the internal
lun id to setup the get bad block ppa. This increases monotonically
with the number luns available. When configuring a ppa, the channel and
lun must be specified separately, leading to an out of bound memory
access in gennvm_block_bb when lun id goes beyond the luns available
within a channel.

Additional, remove out of bound check in gennvm_block_bb(), as it was a
buggy to begin with.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/gennvm.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/drivers/lightnvm/gennvm.c b/drivers/lightnvm/gennvm.c
index 89b880a..61790ae 100644
--- a/drivers/lightnvm/gennvm.c
+++ b/drivers/lightnvm/gennvm.c
@@ -148,11 +148,6 @@ static int gennvm_block_bb(struct gen_nvm *gn, struct ppa_addr ppa,
 			continue;
 
 		blk = &lun->vlun.blocks[i];
-		if (!blk) {
-			pr_err("gennvm: BB data is out of bounds.\n");
-			return -EINVAL;
-		}
-
 		list_move_tail(&blk->list, &lun->bb_list);
 		lun->vlun.nr_bad_blocks++;
 		lun->vlun.nr_free_blocks--;
@@ -257,7 +252,7 @@ static int gennvm_blocks_init(struct nvm_dev *dev, struct gen_nvm *gn)
 
 			ppa.ppa = 0;
 			ppa.g.ch = lun->vlun.chnl_id;
-			ppa.g.lun = lun->vlun.id;
+			ppa.g.lun = lun->vlun.lun_id;
 
 			ret = nvm_get_bb_tbl(dev, ppa, blks);
 			if (ret)
-- 
2.1.4

[PATCH 18/28] lightnvm: do not free unused metadata on rrpc

[Matias Bjørling]

From: Javier González <javier@cnexlabs.com>

rrpc does not save any metadata on a given request. Thus, do not attempt
to free the metadata dma region.

Signed-off-by: Javier González <javier@cnexlabs.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/rrpc.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index c7fef71..ffcfee6 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -711,8 +711,6 @@ static void rrpc_end_io(struct nvm_rq *rqd)
 
 	if (npages > 1)
 		nvm_dev_dma_free(rrpc->dev, rqd->ppa_list, rqd->dma_ppa_list);
-	if (rqd->metadata)
-		nvm_dev_dma_free(rrpc->dev, rqd->metadata, rqd->dma_metadata);
 
 	mempool_free(rqd, rrpc->rq_pool);
 }
-- 
2.1.4

[PATCH 19/28] lightnvm: enable metadata to be sent to device

[Matias Bjørling]

From: Javier González <javier@cnexlabs.com>

Enable metadata buffer to be sent to the device through the metadata
field on the physical rw nvme command. The size of the metadata buffer
must follow dev->oob_size * # of PPAs.

Signed-off-by: Javier González <javier@cnexlabs.com>
Updated description.
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/nvme/host/lightnvm.c | 1 +
 include/linux/lightnvm.h     | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index d426d95..69a47fb 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -467,6 +467,7 @@ static inline void nvme_nvm_rqtocmd(struct request *rq, struct nvm_rq *rqd,
 	c->ph_rw.opcode = rqd->opcode;
 	c->ph_rw.nsid = cpu_to_le32(ns->ns_id);
 	c->ph_rw.spba = cpu_to_le64(rqd->ppa_addr.ppa);
+	c->ph_rw.metadata = cpu_to_le64(rqd->meta_list);
 	c->ph_rw.control = cpu_to_le16(rqd->flags);
 	c->ph_rw.length = cpu_to_le16(rqd->nr_pages - 1);
 
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index af72ca7..678df4d 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -230,8 +230,8 @@ struct nvm_rq {
 
 	struct ppa_addr *ppa_list;
 
-	void *metadata;
-	dma_addr_t dma_metadata;
+	void *meta_list;
+	dma_addr_t dma_meta_list;
 
 	struct completion *wait;
 	nvm_end_io_fn *end_io;
-- 
2.1.4

[PATCH 20/28] lightnvm: rename dma helper functions

[Matias Bjørling]

From: Javier González <javier@cnexlabs.com>

Until now, the dma pool have been exclusively used to allocate the ppa
list being sent to the device. In pblk (upcoming), we use these pools to
allocate metadata too. Thus, we generalize the names of some variables
on the dma helper functions to make the code more readable.

Signed-off-by: Javier González <javier@cnexlabs.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c      | 16 ++++++++--------
 drivers/nvme/host/lightnvm.c |  4 ++--
 include/linux/lightnvm.h     |  2 +-
 3 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index de5db7b..1873a3b 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -87,15 +87,15 @@ EXPORT_SYMBOL(nvm_unregister_tgt_type);
 void *nvm_dev_dma_alloc(struct nvm_dev *dev, gfp_t mem_flags,
 							dma_addr_t *dma_handler)
 {
-	return dev->ops->dev_dma_alloc(dev, dev->ppalist_pool, mem_flags,
+	return dev->ops->dev_dma_alloc(dev, dev->dma_pool, mem_flags,
 								dma_handler);
 }
 EXPORT_SYMBOL(nvm_dev_dma_alloc);
 
-void nvm_dev_dma_free(struct nvm_dev *dev, void *ppa_list,
+void nvm_dev_dma_free(struct nvm_dev *dev, void *addr,
 							dma_addr_t dma_handler)
 {
-	dev->ops->dev_dma_free(dev->ppalist_pool, ppa_list, dma_handler);
+	dev->ops->dev_dma_free(dev->dma_pool, addr, dma_handler);
 }
 EXPORT_SYMBOL(nvm_dev_dma_free);
 
@@ -652,8 +652,8 @@ err:
 
 static void nvm_exit(struct nvm_dev *dev)
 {
-	if (dev->ppalist_pool)
-		dev->ops->destroy_dma_pool(dev->ppalist_pool);
+	if (dev->dma_pool)
+		dev->ops->destroy_dma_pool(dev->dma_pool);
 	nvm_free(dev);
 
 	pr_info("nvm: successfully unloaded\n");
@@ -687,9 +687,9 @@ int nvm_register(struct request_queue *q, char *disk_name,
 	}
 
 	if (dev->ops->max_phys_sect > 1) {
-		dev->ppalist_pool = dev->ops->create_dma_pool(dev, "ppalist");
-		if (!dev->ppalist_pool) {
-			pr_err("nvm: could not create ppa pool\n");
+		dev->dma_pool = dev->ops->create_dma_pool(dev, "ppalist");
+		if (!dev->dma_pool) {
+			pr_err("nvm: could not create dma pool\n");
 			ret = -ENOMEM;
 			goto err_init;
 		}
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 69a47fb..8461f5a 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -565,10 +565,10 @@ static void *nvme_nvm_dev_dma_alloc(struct nvm_dev *dev, void *pool,
 	return dma_pool_alloc(pool, mem_flags, dma_handler);
 }
 
-static void nvme_nvm_dev_dma_free(void *pool, void *ppa_list,
+static void nvme_nvm_dev_dma_free(void *pool, void *addr,
 							dma_addr_t dma_handler)
 {
-	dma_pool_free(pool, ppa_list, dma_handler);
+	dma_pool_free(pool, addr, dma_handler);
 }
 
 static struct nvm_dev_ops nvme_nvm_dev_ops = {
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 678df4d..0e8e019 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -347,7 +347,7 @@ struct nvm_dev {
 	unsigned max_pages_per_blk;
 
 	unsigned long *lun_map;
-	void *ppalist_pool;
+	void *dma_pool;
 
 	struct nvm_id identity;
 
-- 
2.1.4

[PATCH 21/28] nvme/lightnvm: Log using the ctrl named device

[Matias Bjørling]

From: Sagi Grimberg <sagi@grimberg.me>

Align with the rest of the nvme subsystem.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/nvme/host/lightnvm.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 8461f5a..fadeb54 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -367,8 +367,8 @@ static int nvme_nvm_get_l2p_tbl(struct nvm_dev *nvmdev, u64 slba, u32 nlb,
 		ret = nvme_submit_sync_cmd(ns->ctrl->admin_q,
 				(struct nvme_command *)&c, entries, len);
 		if (ret) {
-			dev_err(ns->ctrl->dev, "L2P table transfer failed (%d)\n",
-									ret);
+			dev_err(ns->ctrl->device,
+				"L2P table transfer failed (%d)\n", ret);
 			ret = -EIO;
 			goto out;
 		}
@@ -410,27 +410,28 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
 	ret = nvme_submit_sync_cmd(ctrl->admin_q, (struct nvme_command *)&c,
 								bb_tbl, tblsz);
 	if (ret) {
-		dev_err(ctrl->dev, "get bad block table failed (%d)\n", ret);
+		dev_err(ctrl->device, "get bad block table failed (%d)\n", ret);
 		ret = -EIO;
 		goto out;
 	}
 
 	if (bb_tbl->tblid[0] != 'B' || bb_tbl->tblid[1] != 'B' ||
 		bb_tbl->tblid[2] != 'L' || bb_tbl->tblid[3] != 'T') {
-		dev_err(ctrl->dev, "bbt format mismatch\n");
+		dev_err(ctrl->device, "bbt format mismatch\n");
 		ret = -EINVAL;
 		goto out;
 	}
 
 	if (le16_to_cpu(bb_tbl->verid) != 1) {
 		ret = -EINVAL;
-		dev_err(ctrl->dev, "bbt version not supported\n");
+		dev_err(ctrl->device, "bbt version not supported\n");
 		goto out;
 	}
 
 	if (le32_to_cpu(bb_tbl->tblks) != nr_blks) {
 		ret = -EINVAL;
-		dev_err(ctrl->dev, "bbt unsuspected blocks returned (%u!=%u)",
+		dev_err(ctrl->device,
+				"bbt unsuspected blocks returned (%u!=%u)",
 				le32_to_cpu(bb_tbl->tblks), nr_blks);
 		goto out;
 	}
@@ -457,7 +458,8 @@ static int nvme_nvm_set_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr *ppas,
 	ret = nvme_submit_sync_cmd(ns->ctrl->admin_q, (struct nvme_command *)&c,
 								NULL, 0);
 	if (ret)
-		dev_err(ns->ctrl->dev, "set bad block table failed (%d)\n", ret);
+		dev_err(ns->ctrl->device, "set bad block table failed (%d)\n",
+									ret);
 	return ret;
 }
 
-- 
2.1.4

[PATCH 22/28] lightnvm: do not assume sequential lun alloc.

[Matias Bjørling]

From: Javier González <jg@lightnvm.io>

When doing GC, rrpc calculates the physical LUN to which the rrpc block
belongs too. This calculation is based on the assumption that LUNs are
assigned sequentially to the LUN list. Use the reference to the LUN
instead. This saves us the calculation and allows us to align LUNs in a
different manner to, for example, take advantage of devide parallelism.

Signed-off-by: Javier González <javier@cnexlabs.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/rrpc.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index ffcfee6..48862ead 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -405,9 +405,8 @@ static void rrpc_block_gc(struct work_struct *work)
 									ws_gc);
 	struct rrpc *rrpc = gcb->rrpc;
 	struct rrpc_block *rblk = gcb->rblk;
+	struct rrpc_lun *rlun = rblk->rlun;
 	struct nvm_dev *dev = rrpc->dev;
-	struct nvm_lun *lun = rblk->parent->lun;
-	struct rrpc_lun *rlun = &rrpc->luns[lun->id - rrpc->lun_offset];
 
 	mempool_free(gcb, rrpc->gcb_pool);
 	pr_debug("nvm: block '%lu' being reclaimed\n", rblk->parent->id);
@@ -508,9 +507,9 @@ static void rrpc_gc_queue(struct work_struct *work)
 									ws_gc);
 	struct rrpc *rrpc = gcb->rrpc;
 	struct rrpc_block *rblk = gcb->rblk;
+	struct rrpc_lun *rlun = rblk->rlun;
 	struct nvm_lun *lun = rblk->parent->lun;
 	struct nvm_block *blk = rblk->parent;
-	struct rrpc_lun *rlun = &rrpc->luns[lun->id - rrpc->lun_offset];
 
 	spin_lock(&rlun->lock);
 	list_add_tail(&rblk->prio, &rlun->prio_list);
-- 
2.1.4

[PATCH 23/28] lightnvm: pass dma address to hardware rather than pointer

[Matias Bjørling]

From: Arnd Bergmann <arnd@arndb.de>

A recent change to lightnvm added code to pass a kernel pointer
to the hardware, which gcc complained about:

drivers/nvme/host/lightnvm.c: In function 'nvme_nvm_rqtocmd':
drivers/nvme/host/lightnvm.c:472:32: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
  c->ph_rw.metadata = cpu_to_le64(rqd->meta_list);

It looks like this has no way of working anyway, so this changes
the code to pass the dma_address instead. This was most likely
what was intended here. Neither of the two are currently ever
written to, so the effect is the same for now.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: a34b1eb78e21 ("lightnvm: enable metadata to be sent to device")
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/nvme/host/lightnvm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index fadeb54..65de1e5 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -469,7 +469,7 @@ static inline void nvme_nvm_rqtocmd(struct request *rq, struct nvm_rq *rqd,
 	c->ph_rw.opcode = rqd->opcode;
 	c->ph_rw.nsid = cpu_to_le32(ns->ns_id);
 	c->ph_rw.spba = cpu_to_le64(rqd->ppa_addr.ppa);
-	c->ph_rw.metadata = cpu_to_le64(rqd->meta_list);
+	c->ph_rw.metadata = cpu_to_le64(rqd->dma_meta_list);
 	c->ph_rw.control = cpu_to_le16(rqd->flags);
 	c->ph_rw.length = cpu_to_le16(rqd->nr_pages - 1);
 
-- 
2.1.4

[PATCH 24/28] lightnvm: remove mgt targets on mgt removal

[Matias Bjørling]

Targets associated with a device manager are not freed on device
removal. They have to be manually removed before shutdown. Make sure
any outstanding targets are freed upon shutdown.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c  | 69 ++++++++++++++++++++++++++++++------------------
 include/linux/lightnvm.h |  1 +
 2 files changed, 44 insertions(+), 26 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 1873a3b..d3af771 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -596,13 +596,52 @@ err_fmtype:
 	return ret;
 }
 
+static void nvm_remove_target(struct nvm_target *t)
+{
+	struct nvm_tgt_type *tt = t->type;
+	struct gendisk *tdisk = t->disk;
+	struct request_queue *q = tdisk->queue;
+
+	lockdep_assert_held(&nvm_lock);
+
+	del_gendisk(tdisk);
+	blk_cleanup_queue(q);
+
+	if (tt->exit)
+		tt->exit(tdisk->private_data);
+
+	put_disk(tdisk);
+
+	list_del(&t->list);
+	kfree(t);
+}
+
+static void nvm_free_mgr(struct nvm_dev *dev)
+{
+	struct nvm_target *tgt, *tmp;
+
+	if (!dev->mt)
+		return;
+
+	down_write(&nvm_lock);
+	list_for_each_entry_safe(tgt, tmp, &nvm_targets, list) {
+		if (tgt->dev != dev)
+			continue;
+
+		nvm_remove_target(tgt);
+	}
+	up_write(&nvm_lock);
+
+	dev->mt->unregister_mgr(dev);
+	dev->mt = NULL;
+}
+
 static void nvm_free(struct nvm_dev *dev)
 {
 	if (!dev)
 		return;
 
-	if (dev->mt)
-		dev->mt->unregister_mgr(dev);
+	nvm_free_mgr(dev);
 
 	kfree(dev->lptbl);
 	kfree(dev->lun_map);
@@ -808,6 +847,7 @@ static int nvm_create_target(struct nvm_dev *dev,
 
 	t->type = tt;
 	t->disk = tdisk;
+	t->dev = dev;
 
 	down_write(&nvm_lock);
 	list_add_tail(&t->list, &nvm_targets);
@@ -823,26 +863,6 @@ err_t:
 	return -ENOMEM;
 }
 
-static void nvm_remove_target(struct nvm_target *t)
-{
-	struct nvm_tgt_type *tt = t->type;
-	struct gendisk *tdisk = t->disk;
-	struct request_queue *q = tdisk->queue;
-
-	lockdep_assert_held(&nvm_lock);
-
-	del_gendisk(tdisk);
-	blk_cleanup_queue(q);
-
-	if (tt->exit)
-		tt->exit(tdisk->private_data);
-
-	put_disk(tdisk);
-
-	list_del(&t->list);
-	kfree(t);
-}
-
 static int __nvm_configure_create(struct nvm_ioctl_create *create)
 {
 	struct nvm_dev *dev;
@@ -1231,10 +1251,7 @@ static long nvm_ioctl_dev_factory(struct file *file, void __user *arg)
 		return -EINVAL;
 	}
 
-	if (dev->mt) {
-		dev->mt->unregister_mgr(dev);
-		dev->mt = NULL;
-	}
+	nvm_free_mgr(dev);
 
 	if (dev->identity.cap & NVM_ID_DCAP_BBLKMGMT)
 		return nvm_dev_factory(dev, fact.flags);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 0e8e019..cde31ff 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -200,6 +200,7 @@ struct nvm_id {
 
 struct nvm_target {
 	struct list_head list;
+	struct nvm_dev *dev;
 	struct nvm_tgt_type *type;
 	struct gendisk *disk;
 };
-- 
2.1.4

[PATCH 25/28] lightnvm: expose gennvm_mark_blk to targets

[Matias Bjørling]

Targets can update a block state when having a reference to an
in-memory virtual block. In the case that a target does not keep the
block metadata in memory, it does not have a way to update this
structure.

Therefore, expose gennvm_mark_blk() through the media managers
->mark_blk() callback and let targets update the state structure through
this callback.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/gennvm.c | 27 ++++++++++++++-------------
 include/linux/lightnvm.h  |  4 ++++
 2 files changed, 18 insertions(+), 13 deletions(-)

diff --git a/drivers/lightnvm/gennvm.c b/drivers/lightnvm/gennvm.c
index 61790ae..33a66d8c0 100644
--- a/drivers/lightnvm/gennvm.c
+++ b/drivers/lightnvm/gennvm.c
@@ -407,29 +407,28 @@ static void gennvm_put_blk(struct nvm_dev *dev, struct nvm_block *blk)
 	spin_unlock(&vlun->lock);
 }
 
-static void gennvm_blk_set_type(struct nvm_dev *dev, struct ppa_addr *ppa,
-								int type)
+static void gennvm_mark_blk(struct nvm_dev *dev, struct ppa_addr ppa, int type)
 {
 	struct gen_nvm *gn = dev->mp;
 	struct gen_lun *lun;
 	struct nvm_block *blk;
 
 	pr_debug("gennvm: ppa  (ch: %u lun: %u blk: %u pg: %u) -> %u\n",
-			ppa->g.ch, ppa->g.lun, ppa->g.blk, ppa->g.pg, type);
+			ppa.g.ch, ppa.g.lun, ppa.g.blk, ppa.g.pg, type);
 
-	if (unlikely(ppa->g.ch > dev->nr_chnls ||
-					ppa->g.lun > dev->luns_per_chnl ||
-					ppa->g.blk > dev->blks_per_lun)) {
+	if (unlikely(ppa.g.ch > dev->nr_chnls ||
+					ppa.g.lun > dev->luns_per_chnl ||
+					ppa.g.blk > dev->blks_per_lun)) {
 		WARN_ON_ONCE(1);
 		pr_err("gennvm: ppa broken (ch: %u > %u lun: %u > %u blk: %u > %u",
-				ppa->g.ch, dev->nr_chnls,
-				ppa->g.lun, dev->luns_per_chnl,
-				ppa->g.blk, dev->blks_per_lun);
+				ppa.g.ch, dev->nr_chnls,
+				ppa.g.lun, dev->luns_per_chnl,
+				ppa.g.blk, dev->blks_per_lun);
 		return;
 	}
 
-	lun = &gn->luns[ppa->g.lun * ppa->g.ch];
-	blk = &lun->vlun.blocks[ppa->g.blk];
+	lun = &gn->luns[ppa.g.lun * ppa.g.ch];
+	blk = &lun->vlun.blocks[ppa.g.blk];
 
 	/* will be moved to bb list on put_blk from target */
 	blk->state = type;
@@ -448,12 +447,12 @@ static void gennvm_mark_blk_bad(struct nvm_dev *dev, struct nvm_rq *rqd)
 
 	/* look up blocks and mark them as bad */
 	if (rqd->nr_pages == 1) {
-		gennvm_blk_set_type(dev, &rqd->ppa_addr, NVM_BLK_ST_BAD);
+		gennvm_mark_blk(dev, rqd->ppa_addr, NVM_BLK_ST_BAD);
 		return;
 	}
 
 	while ((bit = find_next_bit(comp_bits, max_secs, bit + 1)) < max_secs)
-		gennvm_blk_set_type(dev, &rqd->ppa_list[bit], NVM_BLK_ST_BAD);
+		gennvm_mark_blk(dev, rqd->ppa_list[bit], NVM_BLK_ST_BAD);
 }
 
 static void gennvm_end_io(struct nvm_rq *rqd)
@@ -544,6 +543,8 @@ static struct nvmm_type gennvm = {
 	.submit_io		= gennvm_submit_io,
 	.erase_blk		= gennvm_erase_blk,
 
+	.mark_blk		= gennvm_mark_blk,
+
 	.get_lun		= gennvm_get_lun,
 	.reserve_lun		= gennvm_reserve_lun,
 	.release_lun		= gennvm_release_lun,
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index cde31ff..3c53911 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -467,6 +467,7 @@ typedef void (nvmm_flush_blk_fn)(struct nvm_dev *, struct nvm_block *);
 typedef int (nvmm_submit_io_fn)(struct nvm_dev *, struct nvm_rq *);
 typedef int (nvmm_erase_blk_fn)(struct nvm_dev *, struct nvm_block *,
 								unsigned long);
+typedef void (nvmm_mark_blk_fn)(struct nvm_dev *, struct ppa_addr, int);
 typedef struct nvm_lun *(nvmm_get_lun_fn)(struct nvm_dev *, int);
 typedef int (nvmm_reserve_lun)(struct nvm_dev *, int);
 typedef void (nvmm_release_lun)(struct nvm_dev *, int);
@@ -494,6 +495,9 @@ struct nvmm_type {
 	nvmm_submit_io_fn *submit_io;
 	nvmm_erase_blk_fn *erase_blk;
 
+	/* Bad block mgmt */
+	nvmm_mark_blk_fn *mark_blk;
+
 	/* Configuration management */
 	nvmm_get_lun_fn *get_lun;
 	nvmm_reserve_lun *reserve_lun;
-- 
2.1.4

[PATCH 26/28] lightnvm: add is_cached entry to struct ppa_addr

[Matias Bjørling]

A target requires a method to identify PPAs that are either cached in
memory or on disk. This can efficiently be maintained within the PPA.
The target host-side translation table can then lookup a PPA and know
from the PPA if it is cached or on disk. In the case it is cached, it is
the responsibility of the target to maintain this cache.

Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 include/linux/lightnvm.h | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 3c53911..3a810d7 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -18,7 +18,7 @@ enum {
 #define NVM_SEC_BITS (8)
 #define NVM_PL_BITS  (8)
 #define NVM_LUN_BITS (8)
-#define NVM_CH_BITS  (8)
+#define NVM_CH_BITS  (7)
 
 struct ppa_addr {
 	/* Generic structure for all addresses */
@@ -30,8 +30,14 @@ struct ppa_addr {
 			u64 pl		: NVM_PL_BITS;
 			u64 lun		: NVM_LUN_BITS;
 			u64 ch		: NVM_CH_BITS;
+			u64 reserved	: 1;
 		} g;
 
+		struct {
+			u64 line	: 63;
+			u64 is_cached	: 1;
+		} c;
+
 		u64 ppa;
 	};
 };
-- 
2.1.4

[PATCH 27/28] lightnvm: rename nr_pages to nr_ppas on nvm_rq

[Matias Bjørling]

From: Javier González <jg@lightnvm.io>

The number of ppas contained on a request is not necessarily the number
of pages that it maps to neither on the target nor on the device side.
In order to avoid confusion, rename nr_pages to nr_ppas since it is what
the variable actually contains.

Signed-off-by: Javier González <javier@cnexlabs.com>
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/core.c      | 16 ++++++++--------
 drivers/lightnvm/gennvm.c    |  2 +-
 drivers/lightnvm/rrpc.c      |  6 +++---
 drivers/lightnvm/rrpc.h      |  2 +-
 drivers/lightnvm/sysblk.c    |  2 +-
 drivers/nvme/host/lightnvm.c |  4 ++--
 include/linux/lightnvm.h     |  2 +-
 7 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index d3af771..160c1a6 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -226,8 +226,8 @@ void nvm_addr_to_generic_mode(struct nvm_dev *dev, struct nvm_rq *rqd)
 {
 	int i;
 
-	if (rqd->nr_pages > 1) {
-		for (i = 0; i < rqd->nr_pages; i++)
+	if (rqd->nr_ppas > 1) {
+		for (i = 0; i < rqd->nr_ppas; i++)
 			rqd->ppa_list[i] = dev_to_generic_addr(dev,
 							rqd->ppa_list[i]);
 	} else {
@@ -240,8 +240,8 @@ void nvm_generic_to_addr_mode(struct nvm_dev *dev, struct nvm_rq *rqd)
 {
 	int i;
 
-	if (rqd->nr_pages > 1) {
-		for (i = 0; i < rqd->nr_pages; i++)
+	if (rqd->nr_ppas > 1) {
+		for (i = 0; i < rqd->nr_ppas; i++)
 			rqd->ppa_list[i] = generic_to_dev_addr(dev,
 							rqd->ppa_list[i]);
 	} else {
@@ -256,13 +256,13 @@ int nvm_set_rqd_ppalist(struct nvm_dev *dev, struct nvm_rq *rqd,
 	int i, plane_cnt, pl_idx;
 
 	if ((!vblk || dev->plane_mode == NVM_PLANE_SINGLE) && nr_ppas == 1) {
-		rqd->nr_pages = nr_ppas;
+		rqd->nr_ppas = nr_ppas;
 		rqd->ppa_addr = ppas[0];
 
 		return 0;
 	}
 
-	rqd->nr_pages = nr_ppas;
+	rqd->nr_ppas = nr_ppas;
 	rqd->ppa_list = nvm_dev_dma_alloc(dev, GFP_KERNEL, &rqd->dma_ppa_list);
 	if (!rqd->ppa_list) {
 		pr_err("nvm: failed to allocate dma memory\n");
@@ -274,7 +274,7 @@ int nvm_set_rqd_ppalist(struct nvm_dev *dev, struct nvm_rq *rqd,
 			rqd->ppa_list[i] = ppas[i];
 	} else {
 		plane_cnt = dev->plane_mode;
-		rqd->nr_pages *= plane_cnt;
+		rqd->nr_ppas *= plane_cnt;
 
 		for (i = 0; i < nr_ppas; i++) {
 			for (pl_idx = 0; pl_idx < plane_cnt; pl_idx++) {
@@ -395,7 +395,7 @@ int nvm_submit_ppa_list(struct nvm_dev *dev, struct ppa_addr *ppa_list,
 
 	memset(&rqd, 0, sizeof(struct nvm_rq));
 
-	rqd.nr_pages = nr_ppas;
+	rqd.nr_ppas = nr_ppas;
 	if (nr_ppas > 1)
 		rqd.ppa_list = ppa_list;
 	else
diff --git a/drivers/lightnvm/gennvm.c b/drivers/lightnvm/gennvm.c
index 33a66d8c0..ec9fb68 100644
--- a/drivers/lightnvm/gennvm.c
+++ b/drivers/lightnvm/gennvm.c
@@ -446,7 +446,7 @@ static void gennvm_mark_blk_bad(struct nvm_dev *dev, struct nvm_rq *rqd)
 	nvm_addr_to_generic_mode(dev, rqd);
 
 	/* look up blocks and mark them as bad */
-	if (rqd->nr_pages == 1) {
+	if (rqd->nr_ppas == 1) {
 		gennvm_mark_blk(dev, rqd->ppa_addr, NVM_BLK_ST_BAD);
 		return;
 	}
diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index 48862ead..72aca96 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -695,7 +695,7 @@ static void rrpc_end_io(struct nvm_rq *rqd)
 {
 	struct rrpc *rrpc = container_of(rqd->ins, struct rrpc, instance);
 	struct rrpc_rq *rrqd = nvm_rq_to_pdu(rqd);
-	uint8_t npages = rqd->nr_pages;
+	uint8_t npages = rqd->nr_ppas;
 	sector_t laddr = rrpc_get_laddr(rqd->bio) - npages;
 
 	if (bio_data_dir(rqd->bio) == WRITE)
@@ -883,7 +883,7 @@ static int rrpc_submit_io(struct rrpc *rrpc, struct bio *bio,
 	bio_get(bio);
 	rqd->bio = bio;
 	rqd->ins = &rrpc->instance;
-	rqd->nr_pages = nr_pages;
+	rqd->nr_ppas = nr_pages;
 	rrq->flags = flags;
 
 	err = nvm_submit_io(rrpc->dev, rqd);
@@ -892,7 +892,7 @@ static int rrpc_submit_io(struct rrpc *rrpc, struct bio *bio,
 		bio_put(bio);
 		if (!(flags & NVM_IOTYPE_GC)) {
 			rrpc_unlock_rq(rrpc, rqd);
-			if (rqd->nr_pages > 1)
+			if (rqd->nr_ppas > 1)
 				nvm_dev_dma_free(rrpc->dev,
 			rqd->ppa_list, rqd->dma_ppa_list);
 		}
diff --git a/drivers/lightnvm/rrpc.h b/drivers/lightnvm/rrpc.h
index 2653484..87e84b5 100644
--- a/drivers/lightnvm/rrpc.h
+++ b/drivers/lightnvm/rrpc.h
@@ -251,7 +251,7 @@ static inline void rrpc_unlock_laddr(struct rrpc *rrpc,
 static inline void rrpc_unlock_rq(struct rrpc *rrpc, struct nvm_rq *rqd)
 {
 	struct rrpc_inflight_rq *r = rrpc_get_inflight_rq(rqd);
-	uint8_t pages = rqd->nr_pages;
+	uint8_t pages = rqd->nr_ppas;
 
 	BUG_ON((r->l_start + pages) > rrpc->nr_sects);
 
diff --git a/drivers/lightnvm/sysblk.c b/drivers/lightnvm/sysblk.c
index b98ca19..994697a 100644
--- a/drivers/lightnvm/sysblk.c
+++ b/drivers/lightnvm/sysblk.c
@@ -280,7 +280,7 @@ static int nvm_set_bb_tbl(struct nvm_dev *dev, struct sysblk_scan *s, int type)
 	nvm_set_rqd_ppalist(dev, &rqd, s->ppas, s->nr_ppas, 1);
 	nvm_generic_to_addr_mode(dev, &rqd);
 
-	ret = dev->ops->set_bb_tbl(dev, &rqd.ppa_addr, rqd.nr_pages, type);
+	ret = dev->ops->set_bb_tbl(dev, &rqd.ppa_addr, rqd.nr_ppas, type);
 	nvm_free_rqd_ppalist(dev, &rqd);
 	if (ret) {
 		pr_err("nvm: sysblk failed bb mark\n");
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 65de1e5..a0af055 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -471,7 +471,7 @@ static inline void nvme_nvm_rqtocmd(struct request *rq, struct nvm_rq *rqd,
 	c->ph_rw.spba = cpu_to_le64(rqd->ppa_addr.ppa);
 	c->ph_rw.metadata = cpu_to_le64(rqd->dma_meta_list);
 	c->ph_rw.control = cpu_to_le16(rqd->flags);
-	c->ph_rw.length = cpu_to_le16(rqd->nr_pages - 1);
+	c->ph_rw.length = cpu_to_le16(rqd->nr_ppas - 1);
 
 	if (rqd->opcode == NVM_OP_HBWRITE || rqd->opcode == NVM_OP_HBREAD)
 		c->hb_rw.slba = cpu_to_le64(nvme_block_nr(ns,
@@ -542,7 +542,7 @@ static int nvme_nvm_erase_block(struct nvm_dev *dev, struct nvm_rq *rqd)
 	c.erase.opcode = NVM_OP_ERASE;
 	c.erase.nsid = cpu_to_le32(ns->ns_id);
 	c.erase.spba = cpu_to_le64(rqd->ppa_addr.ppa);
-	c.erase.length = cpu_to_le16(rqd->nr_pages - 1);
+	c.erase.length = cpu_to_le16(rqd->nr_ppas - 1);
 
 	return nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
 }
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 3a810d7..b2991c7 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -244,7 +244,7 @@ struct nvm_rq {
 	nvm_end_io_fn *end_io;
 
 	uint8_t opcode;
-	uint16_t nr_pages;
+	uint16_t nr_ppas;
 	uint16_t flags;
 
 	u64 ppa_status; /* ppa media status */
-- 
2.1.4

[PATCH 28/28] lightnvm: reserved space calculation incorrect

[Matias Bjørling]

From: Javier González <jg@lightnvm.io>

The nvm_dev->max_pages_per_blk variable was removed in favor of the new
nvm->sec_per_blk variable. The ->max_pages_per_blk variable was still
used in rrpc_capacity, reporting the reserved capacity to zero. Replace
with ->sec_per_blk to calculate the reserved area again.

Signed-off-by: Javier González <javier@cnexlabs.com>
Updated patch description. Was "lightnvm: eliminate redundant variable"
Signed-off-by: Matias Bjørling <m@bjorling.me>
---
 drivers/lightnvm/rrpc.c  | 2 +-
 include/linux/lightnvm.h | 1 -
 2 files changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index 72aca96..2103e97 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -1264,7 +1264,7 @@ static sector_t rrpc_capacity(void *private)
 	sector_t reserved, provisioned;
 
 	/* cur, gc, and two emergency blocks for each lun */
-	reserved = rrpc->nr_luns * dev->max_pages_per_blk * 4;
+	reserved = rrpc->nr_luns * dev->sec_per_blk * 4;
 	provisioned = rrpc->nr_sects - reserved;
 
 	if (reserved > rrpc->nr_sects) {
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index b2991c7..ef2c7d2 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -351,7 +351,6 @@ struct nvm_dev {
 	unsigned long total_blocks;
 	unsigned long total_secs;
 	int nr_luns;
-	unsigned max_pages_per_blk;
 
 	unsigned long *lun_map;
 	void *dma_pool;
-- 
2.1.4

Re: [PATCH 00/28] LightNVM fixes for 4.7

[Jens Axboe]

On 05/06/2016 12:02 PM, Matias Bjørling wrote:
> Hi Jens,
>
> A collection of fixes and features destined for 4.7. They contain:
>
>    Fixes all over the place from Arnd, Jeff, Sagi, Wenwei, Javier and me.
>
>    Preparation patches from Simon to enable sysfs support. The final
>    patch is delayed to 4.8. We need to integrate it correctly with
>    device drivers.
>
>    A patch from Javier to allow a target to access the metadata field.
>    Typically it maps to the flash page out of band area.
>
>    Preparation patches for pblk and persistent block management
>    information from me.
>
> Please pick up.

Added for 4.7, thanks.

-- 
Jens Axboe