* [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support
@ 2018-02-26 13:16 Javier González
2018-02-26 13:17 ` [PATCH 01/19] lightnvm: simplify geometry structure Javier González
` (19 more replies)
0 siblings, 20 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:16 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
# Changes since V2:
Apply Matias' feedback:
- Remove generic nvm_id identify structure.
- Do not remap capabilities (cap) to media and controlled capabilities
(mccap). Instead, add a comment to prevent confusion when
crosschecking with 2.0 spec.
- Change maxoc and maxocpu defaults from 1 block to the max number of
blocks.
- Re-implement the generic geometry to use nvm_geo on both device and
targets. Maintain nvm_common_geo to make it easier to copy the common
part of the geometry (without having to overwrite target-specific
fields, which is ugly and error prone). Matias, if you still want to
get rid of this, we can do it.
- Re-order patches with renaming to make them more meaningful. These
belong to the series, since the name changes are motivated by 2.0
inclusions. The only exception would be 36d10bfd3234, but I hope it
is OK I include it here.
Also,
- Eliminate a dependency between luns and lines in the init/exit
refactoring.
- Use the global address format when possible to avoid defaulting on
the 1.2 path. This will safe headaches if the address format changes
at some point.
I took out the patch allowing to do bit shifts on non power-of-2 media
formats on pblk's mapping since it requires touching many places that
are not 2.0 related. I'll submit this separately.
# Changes since V1:
Apply Matias' feedback:
- Rebase on top of Matias' latest patches.
- Use nvme_get_log_ext to submit report chunk and export it.
- Re-write report chunk based on Matias' suggestions. Here, I
maintained the lba interface, but it was necessary to redo the
address formatting to match the chunk log page format. For pblk,
this means a double address transformation, but it enables the
standard path to use lbas, plus, this is not in the fast path.
- Fold address format together with address transformations.
- Split the generic geometry patch in different patches.
- Remove refactoring of ligthnvm's core sysfs.
Feedback not applied:
- Not letting pblk know about 1.2 and 2.0 bad block paths.
Since the interfaces for get/set bad block and report chunk are so
different, moving this logic to core adds assumptions on how the
targets would want to get the data back. A way of doing this is
creating a logical report chunk on the 1.2 path, but this would
mean that values like the wear-index are invalid, which requires
the target knowledge. I'm open to suggestions here.
Also:
- Do some further renamings
- Create a generic address format to make it explicit where we share
1.2 and 2.0 fields to avoid address formatting in the fast path.
- Add new fields to sysfs to complete spec and show major/minor
versions (version and subversion to respect current interface).
Implement 2.0 support in pblk. This includes the address formatting and
mapping paths, as well as the sysfs entries for them.
Javier
Javier González (19):
lightnvm: simplify geometry structure.
lightnvm: add controller capabilities to 2.0
lightnvm: add minor version to generic geometry
lightnvm: add shorten OCSSD version in geo
lightnvm: complete geo structure with maxoc*
lightnvm: pblk: check for supported version
lightnvm: complete 2.0 values in sysfs
lightnvm: rename number of channels and luns
lightnvm: rename sect_* to sec_*
lightnvm: add support for 2.0 address format
lightnvm: pblk: rename ppaf* to addrf*
lightnvn: pblk: use generic address format
lightnvm: make address conversions depend on generic device
nvme: make nvme_get_log_ext available
lightnvm: implement get log report chunk helpers
lightnvm: define chunk states
lightnvm: pblk: implement get log report chunk
lightnvm: pblk: refactor init/exit sequences
lightnvm: pblk: implement 2.0 support
drivers/lightnvm/core.c | 199 +++++-----
drivers/lightnvm/pblk-core.c | 151 +++++--
drivers/lightnvm/pblk-gc.c | 2 +-
drivers/lightnvm/pblk-init.c | 833 +++++++++++++++++++++++----------------
drivers/lightnvm/pblk-map.c | 4 +-
drivers/lightnvm/pblk-read.c | 2 +-
drivers/lightnvm/pblk-recovery.c | 14 +-
drivers/lightnvm/pblk-rl.c | 2 +-
drivers/lightnvm/pblk-sysfs.c | 134 ++++++-
drivers/lightnvm/pblk-write.c | 2 +-
drivers/lightnvm/pblk.h | 255 ++++++++----
drivers/nvme/host/lightnvm.c | 458 ++++++++++++++-------
drivers/nvme/host/nvme.h | 3 +
include/linux/lightnvm.h | 338 ++++++++++------
14 files changed, 1559 insertions(+), 838 deletions(-)
--
2.7.4
^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH 01/19] lightnvm: simplify geometry structure.
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 02/19] lightnvm: add controller capabilities to 2.0 Javier González
` (18 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Currently, the device geometry is stored redundantly in the nvm_id and
nvm_geo structures at a device level. Moreover, when instantiating
targets on a specific number of LUNs, these structures are replicated
and manually modified to fit the instance channel and LUN partitioning.
Instead, create a generic geometry around nvm_geo, which can be used by
(i) the underlying device to describe the geometry of the whole device,
and (ii) instances to describe their geometry independently.
Since these share a big part of the geometry, create a nvm_common_geo
structure that keeps the static geoometry values that are shared across
instances.
As we introduce support for 2.0, these structures allow to abstract
spec. specific values and present a common geometry to targets.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/core.c | 131 ++++++---------
drivers/lightnvm/pblk-core.c | 16 +-
drivers/lightnvm/pblk-gc.c | 2 +-
drivers/lightnvm/pblk-init.c | 123 +++++++-------
drivers/lightnvm/pblk-read.c | 2 +-
drivers/lightnvm/pblk-recovery.c | 14 +-
drivers/lightnvm/pblk-rl.c | 2 +-
drivers/lightnvm/pblk-sysfs.c | 39 +++--
drivers/lightnvm/pblk-write.c | 2 +-
drivers/lightnvm/pblk.h | 93 +++++------
drivers/nvme/host/lightnvm.c | 346 +++++++++++++++++++++++----------------
include/linux/lightnvm.h | 208 ++++++++++++-----------
12 files changed, 514 insertions(+), 464 deletions(-)
diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 689c97b97775..43e3d6bb5be6 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -111,6 +111,7 @@ static void nvm_release_luns_err(struct nvm_dev *dev, int lun_begin,
static void nvm_remove_tgt_dev(struct nvm_tgt_dev *tgt_dev, int clear)
{
struct nvm_dev *dev = tgt_dev->parent;
+ struct nvm_geo *dev_geo = &dev->geo;
struct nvm_dev_map *dev_map = tgt_dev->map;
int i, j;
@@ -122,7 +123,7 @@ static void nvm_remove_tgt_dev(struct nvm_tgt_dev *tgt_dev, int clear)
if (clear) {
for (j = 0; j < ch_map->nr_luns; j++) {
int lun = j + lun_offs[j];
- int lunid = (ch * dev->geo.nr_luns) + lun;
+ int lunid = (ch * dev_geo->nr_luns) + lun;
WARN_ON(!test_and_clear_bit(lunid,
dev->lun_map));
@@ -143,19 +144,20 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
u16 lun_begin, u16 lun_end,
u16 op)
{
+ struct nvm_geo *dev_geo = &dev->geo;
struct nvm_tgt_dev *tgt_dev = NULL;
struct nvm_dev_map *dev_rmap = dev->rmap;
struct nvm_dev_map *dev_map;
struct ppa_addr *luns;
int nr_luns = lun_end - lun_begin + 1;
int luns_left = nr_luns;
- int nr_chnls = nr_luns / dev->geo.nr_luns;
- int nr_chnls_mod = nr_luns % dev->geo.nr_luns;
- int bch = lun_begin / dev->geo.nr_luns;
- int blun = lun_begin % dev->geo.nr_luns;
+ int nr_chnls = nr_luns / dev_geo->nr_luns;
+ int nr_chnls_mod = nr_luns % dev_geo->nr_luns;
+ int bch = lun_begin / dev_geo->nr_luns;
+ int blun = lun_begin % dev_geo->nr_luns;
int lunid = 0;
int lun_balanced = 1;
- int prev_nr_luns;
+ int sec_per_lun, prev_nr_luns;
int i, j;
nr_chnls = (nr_chnls_mod == 0) ? nr_chnls : nr_chnls + 1;
@@ -173,15 +175,15 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
if (!luns)
goto err_luns;
- prev_nr_luns = (luns_left > dev->geo.nr_luns) ?
- dev->geo.nr_luns : luns_left;
+ prev_nr_luns = (luns_left > dev_geo->nr_luns) ?
+ dev_geo->nr_luns : luns_left;
for (i = 0; i < nr_chnls; i++) {
struct nvm_ch_map *ch_rmap = &dev_rmap->chnls[i + bch];
int *lun_roffs = ch_rmap->lun_offs;
struct nvm_ch_map *ch_map = &dev_map->chnls[i];
int *lun_offs;
- int luns_in_chnl = (luns_left > dev->geo.nr_luns) ?
- dev->geo.nr_luns : luns_left;
+ int luns_in_chnl = (luns_left > dev_geo->nr_luns) ?
+ dev_geo->nr_luns : luns_left;
if (lun_balanced && prev_nr_luns != luns_in_chnl)
lun_balanced = 0;
@@ -215,18 +217,22 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
if (!tgt_dev)
goto err_ch;
- memcpy(&tgt_dev->geo, &dev->geo, sizeof(struct nvm_geo));
/* Target device only owns a portion of the physical device */
tgt_dev->geo.nr_chnls = nr_chnls;
- tgt_dev->geo.all_luns = nr_luns;
tgt_dev->geo.nr_luns = (lun_balanced) ? prev_nr_luns : -1;
+ tgt_dev->geo.all_luns = nr_luns;
+ tgt_dev->geo.all_chunks = nr_luns * dev_geo->c.num_chk;
+
tgt_dev->geo.op = op;
- tgt_dev->total_secs = nr_luns * tgt_dev->geo.sec_per_lun;
+
+ sec_per_lun = dev_geo->c.clba * dev_geo->c.num_chk;
+ tgt_dev->geo.total_secs = nr_luns * sec_per_lun;
+
+ tgt_dev->geo.c = dev_geo->c;
+
tgt_dev->q = dev->q;
tgt_dev->map = dev_map;
tgt_dev->luns = luns;
- memcpy(&tgt_dev->identity, &dev->identity, sizeof(struct nvm_id));
-
tgt_dev->parent = dev;
return tgt_dev;
@@ -283,24 +289,24 @@ static int nvm_config_check_luns(struct nvm_geo *geo, int lun_begin,
static int __nvm_config_simple(struct nvm_dev *dev,
struct nvm_ioctl_create_simple *s)
{
- struct nvm_geo *geo = &dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
if (s->lun_begin == -1 && s->lun_end == -1) {
s->lun_begin = 0;
- s->lun_end = geo->all_luns - 1;
+ s->lun_end = dev_geo->all_luns - 1;
}
- return nvm_config_check_luns(geo, s->lun_begin, s->lun_end);
+ return nvm_config_check_luns(dev_geo, s->lun_begin, s->lun_end);
}
static int __nvm_config_extended(struct nvm_dev *dev,
struct nvm_ioctl_create_extended *e)
{
- struct nvm_geo *geo = &dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
if (e->lun_begin == 0xFFFF && e->lun_end == 0xFFFF) {
e->lun_begin = 0;
- e->lun_end = dev->geo.all_luns - 1;
+ e->lun_end = dev_geo->all_luns - 1;
}
/* op not set falls into target's default */
@@ -313,7 +319,7 @@ static int __nvm_config_extended(struct nvm_dev *dev,
return -EINVAL;
}
- return nvm_config_check_luns(geo, e->lun_begin, e->lun_end);
+ return nvm_config_check_luns(dev_geo, e->lun_begin, e->lun_end);
}
static int nvm_create_tgt(struct nvm_dev *dev, struct nvm_ioctl_create *create)
@@ -408,7 +414,7 @@ static int nvm_create_tgt(struct nvm_dev *dev, struct nvm_ioctl_create *create)
tqueue->queuedata = targetdata;
blk_queue_max_hw_sectors(tqueue,
- (dev->geo.sec_size >> 9) * NVM_MAX_VLBA);
+ (dev->geo.c.csecs >> 9) * NVM_MAX_VLBA);
set_capacity(tdisk, tt->capacity(targetdata));
add_disk(tdisk);
@@ -497,6 +503,7 @@ static int nvm_remove_tgt(struct nvm_dev *dev, struct nvm_ioctl_remove *remove)
static int nvm_register_map(struct nvm_dev *dev)
{
+ struct nvm_geo *dev_geo = &dev->geo;
struct nvm_dev_map *rmap;
int i, j;
@@ -504,15 +511,15 @@ static int nvm_register_map(struct nvm_dev *dev)
if (!rmap)
goto err_rmap;
- rmap->chnls = kcalloc(dev->geo.nr_chnls, sizeof(struct nvm_ch_map),
+ rmap->chnls = kcalloc(dev_geo->nr_chnls, sizeof(struct nvm_ch_map),
GFP_KERNEL);
if (!rmap->chnls)
goto err_chnls;
- for (i = 0; i < dev->geo.nr_chnls; i++) {
+ for (i = 0; i < dev_geo->nr_chnls; i++) {
struct nvm_ch_map *ch_rmap;
int *lun_roffs;
- int luns_in_chnl = dev->geo.nr_luns;
+ int luns_in_chnl = dev_geo->nr_luns;
ch_rmap = &rmap->chnls[i];
@@ -543,10 +550,11 @@ static int nvm_register_map(struct nvm_dev *dev)
static void nvm_unregister_map(struct nvm_dev *dev)
{
+ struct nvm_geo *dev_geo = &dev->geo;
struct nvm_dev_map *rmap = dev->rmap;
int i;
- for (i = 0; i < dev->geo.nr_chnls; i++)
+ for (i = 0; i < dev_geo->nr_chnls; i++)
kfree(rmap->chnls[i].lun_offs);
kfree(rmap->chnls);
@@ -675,7 +683,7 @@ static int nvm_set_rqd_ppalist(struct nvm_tgt_dev *tgt_dev, struct nvm_rq *rqd,
int i, plane_cnt, pl_idx;
struct ppa_addr ppa;
- if (geo->plane_mode == NVM_PLANE_SINGLE && nr_ppas == 1) {
+ if (geo->c.pln_mode == NVM_PLANE_SINGLE && nr_ppas == 1) {
rqd->nr_ppas = nr_ppas;
rqd->ppa_addr = ppas[0];
@@ -689,7 +697,7 @@ static int nvm_set_rqd_ppalist(struct nvm_tgt_dev *tgt_dev, struct nvm_rq *rqd,
return -ENOMEM;
}
- plane_cnt = geo->plane_mode;
+ plane_cnt = geo->c.pln_mode;
rqd->nr_ppas *= plane_cnt;
for (i = 0; i < nr_ppas; i++) {
@@ -804,18 +812,18 @@ EXPORT_SYMBOL(nvm_end_io);
*/
int nvm_bb_tbl_fold(struct nvm_dev *dev, u8 *blks, int nr_blks)
{
- struct nvm_geo *geo = &dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
int blk, offset, pl, blktype;
- if (nr_blks != geo->nr_chks * geo->plane_mode)
+ if (nr_blks != dev_geo->c.num_chk * dev_geo->c.pln_mode)
return -EINVAL;
- for (blk = 0; blk < geo->nr_chks; blk++) {
- offset = blk * geo->plane_mode;
+ for (blk = 0; blk < dev_geo->c.num_chk; blk++) {
+ offset = blk * dev_geo->c.pln_mode;
blktype = blks[offset];
/* Bad blocks on any planes take precedence over other types */
- for (pl = 0; pl < geo->plane_mode; pl++) {
+ for (pl = 0; pl < dev_geo->c.pln_mode; pl++) {
if (blks[offset + pl] &
(NVM_BLK_T_BAD|NVM_BLK_T_GRWN_BAD)) {
blktype = blks[offset + pl];
@@ -826,7 +834,7 @@ int nvm_bb_tbl_fold(struct nvm_dev *dev, u8 *blks, int nr_blks)
blks[blk] = blktype;
}
- return geo->nr_chks;
+ return dev_geo->c.num_chk;
}
EXPORT_SYMBOL(nvm_bb_tbl_fold);
@@ -843,41 +851,10 @@ EXPORT_SYMBOL(nvm_get_tgt_bb_tbl);
static int nvm_core_init(struct nvm_dev *dev)
{
- struct nvm_id *id = &dev->identity;
- struct nvm_geo *geo = &dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
int ret;
- memcpy(&geo->ppaf, &id->ppaf, sizeof(struct nvm_addr_format));
-
- if (id->mtype != 0) {
- pr_err("nvm: memory type not supported\n");
- return -EINVAL;
- }
-
- /* Whole device values */
- geo->nr_chnls = id->num_ch;
- geo->nr_luns = id->num_lun;
-
- /* Generic device geometry values */
- geo->ws_min = id->ws_min;
- geo->ws_opt = id->ws_opt;
- geo->ws_seq = id->ws_seq;
- geo->ws_per_chk = id->ws_per_chk;
- geo->nr_chks = id->num_chk;
- geo->mccap = id->mccap;
-
- geo->sec_per_chk = id->clba;
- geo->sec_per_lun = geo->sec_per_chk * geo->nr_chks;
- geo->all_luns = geo->nr_luns * geo->nr_chnls;
-
- /* 1.2 spec device geometry values */
- geo->plane_mode = 1 << geo->ws_seq;
- geo->nr_planes = geo->ws_opt / geo->ws_min;
- geo->sec_per_pg = geo->ws_min;
- geo->sec_per_pl = geo->sec_per_pg * geo->nr_planes;
-
- dev->total_secs = geo->all_luns * geo->sec_per_lun;
- dev->lun_map = kcalloc(BITS_TO_LONGS(geo->all_luns),
+ dev->lun_map = kcalloc(BITS_TO_LONGS(dev_geo->all_luns),
sizeof(unsigned long), GFP_KERNEL);
if (!dev->lun_map)
return -ENOMEM;
@@ -912,19 +889,17 @@ static void nvm_free(struct nvm_dev *dev)
static int nvm_init(struct nvm_dev *dev)
{
- struct nvm_geo *geo = &dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
int ret = -EINVAL;
- if (dev->ops->identity(dev, &dev->identity)) {
+ if (dev->ops->identity(dev)) {
pr_err("nvm: device could not be identified\n");
goto err;
}
- if (dev->identity.ver_id != 1 && dev->identity.ver_id != 2) {
- pr_err("nvm: device ver_id %d not supported by kernel.\n",
- dev->identity.ver_id);
- goto err;
- }
+ pr_debug("nvm: ver:%u nvm_vendor:%x\n",
+ dev_geo->ver_id,
+ dev_geo->c.vmnt);
ret = nvm_core_init(dev);
if (ret) {
@@ -932,10 +907,10 @@ static int nvm_init(struct nvm_dev *dev)
goto err;
}
- pr_info("nvm: registered %s [%u/%u/%u/%u/%u/%u]\n",
- dev->name, geo->sec_per_pg, geo->nr_planes,
- geo->ws_per_chk, geo->nr_chks,
- geo->all_luns, geo->nr_chnls);
+ pr_info("nvm: registered %s [%u/%u/%u/%u/%u]\n",
+ dev->name, dev_geo->c.ws_min, dev_geo->c.ws_opt,
+ dev_geo->c.num_chk, dev_geo->all_luns,
+ dev_geo->nr_chnls);
return 0;
err:
pr_err("nvm: failed to initialize nvm\n");
diff --git a/drivers/lightnvm/pblk-core.c b/drivers/lightnvm/pblk-core.c
index 22e61cd4f801..519af8b9eab7 100644
--- a/drivers/lightnvm/pblk-core.c
+++ b/drivers/lightnvm/pblk-core.c
@@ -613,7 +613,7 @@ static int pblk_line_submit_emeta_io(struct pblk *pblk, struct pblk_line *line,
memset(&rqd, 0, sizeof(struct nvm_rq));
rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
- rq_len = rq_ppas * geo->sec_size;
+ rq_len = rq_ppas * geo->c.csecs;
bio = pblk_bio_map_addr(pblk, emeta_buf, rq_ppas, rq_len,
l_mg->emeta_alloc_type, GFP_KERNEL);
@@ -722,7 +722,7 @@ u64 pblk_line_smeta_start(struct pblk *pblk, struct pblk_line *line)
if (bit >= lm->blk_per_line)
return -1;
- return bit * geo->sec_per_pl;
+ return bit * geo->c.ws_opt;
}
static int pblk_line_submit_smeta_io(struct pblk *pblk, struct pblk_line *line,
@@ -1035,19 +1035,19 @@ static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
/* Capture bad block information on line mapping bitmaps */
while ((bit = find_next_bit(line->blk_bitmap, lm->blk_per_line,
bit + 1)) < lm->blk_per_line) {
- off = bit * geo->sec_per_pl;
+ off = bit * geo->c.ws_opt;
bitmap_shift_left(l_mg->bb_aux, l_mg->bb_template, off,
lm->sec_per_line);
bitmap_or(line->map_bitmap, line->map_bitmap, l_mg->bb_aux,
lm->sec_per_line);
- line->sec_in_line -= geo->sec_per_chk;
+ line->sec_in_line -= geo->c.clba;
if (bit >= lm->emeta_bb)
nr_bb++;
}
/* Mark smeta metadata sectors as bad sectors */
bit = find_first_zero_bit(line->blk_bitmap, lm->blk_per_line);
- off = bit * geo->sec_per_pl;
+ off = bit * geo->c.ws_opt;
bitmap_set(line->map_bitmap, off, lm->smeta_sec);
line->sec_in_line -= lm->smeta_sec;
line->smeta_ssec = off;
@@ -1066,10 +1066,10 @@ static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
emeta_secs = lm->emeta_sec[0];
off = lm->sec_per_line;
while (emeta_secs) {
- off -= geo->sec_per_pl;
+ off -= geo->c.ws_opt;
if (!test_bit(off, line->invalid_bitmap)) {
- bitmap_set(line->invalid_bitmap, off, geo->sec_per_pl);
- emeta_secs -= geo->sec_per_pl;
+ bitmap_set(line->invalid_bitmap, off, geo->c.ws_opt);
+ emeta_secs -= geo->c.ws_opt;
}
}
diff --git a/drivers/lightnvm/pblk-gc.c b/drivers/lightnvm/pblk-gc.c
index 320f99af99e9..16afea3f5541 100644
--- a/drivers/lightnvm/pblk-gc.c
+++ b/drivers/lightnvm/pblk-gc.c
@@ -88,7 +88,7 @@ static void pblk_gc_line_ws(struct work_struct *work)
up(&gc->gc_sem);
- gc_rq->data = vmalloc(gc_rq->nr_secs * geo->sec_size);
+ gc_rq->data = vmalloc(gc_rq->nr_secs * geo->c.csecs);
if (!gc_rq->data) {
pr_err("pblk: could not GC line:%d (%d/%d)\n",
line->id, *line->vsc, gc_rq->nr_secs);
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index 5261702e9ff7..95ecb0ec736b 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -146,7 +146,7 @@ static int pblk_rwb_init(struct pblk *pblk)
return -ENOMEM;
power_size = get_count_order(nr_entries);
- power_seg_sz = get_count_order(geo->sec_size);
+ power_seg_sz = get_count_order(geo->c.csecs);
return pblk_rb_init(&pblk->rwb, entries, power_size, power_seg_sz);
}
@@ -154,11 +154,11 @@ static int pblk_rwb_init(struct pblk *pblk)
/* Minimum pages needed within a lun */
#define ADDR_POOL_SIZE 64
-static int pblk_set_ppaf(struct pblk *pblk)
+static int pblk_set_addrf_12(struct nvm_geo *geo,
+ struct nvm_addr_format_12 *dst)
{
- struct nvm_tgt_dev *dev = pblk->dev;
- struct nvm_geo *geo = &dev->geo;
- struct nvm_addr_format ppaf = geo->ppaf;
+ struct nvm_addr_format_12 *src =
+ (struct nvm_addr_format_12 *)&geo->c.addrf;
int power_len;
/* Re-calculate channel and lun format to adapt to configuration */
@@ -167,34 +167,50 @@ static int pblk_set_ppaf(struct pblk *pblk)
pr_err("pblk: supports only power-of-two channel config.\n");
return -EINVAL;
}
- ppaf.ch_len = power_len;
+ dst->ch_len = power_len;
power_len = get_count_order(geo->nr_luns);
if (1 << power_len != geo->nr_luns) {
pr_err("pblk: supports only power-of-two LUN config.\n");
return -EINVAL;
}
- ppaf.lun_len = power_len;
+ dst->lun_len = power_len;
- pblk->ppaf.sec_offset = 0;
- pblk->ppaf.pln_offset = ppaf.sect_len;
- pblk->ppaf.ch_offset = pblk->ppaf.pln_offset + ppaf.pln_len;
- pblk->ppaf.lun_offset = pblk->ppaf.ch_offset + ppaf.ch_len;
- pblk->ppaf.pg_offset = pblk->ppaf.lun_offset + ppaf.lun_len;
- pblk->ppaf.blk_offset = pblk->ppaf.pg_offset + ppaf.pg_len;
- pblk->ppaf.sec_mask = (1ULL << ppaf.sect_len) - 1;
- pblk->ppaf.pln_mask = ((1ULL << ppaf.pln_len) - 1) <<
- pblk->ppaf.pln_offset;
- pblk->ppaf.ch_mask = ((1ULL << ppaf.ch_len) - 1) <<
- pblk->ppaf.ch_offset;
- pblk->ppaf.lun_mask = ((1ULL << ppaf.lun_len) - 1) <<
- pblk->ppaf.lun_offset;
- pblk->ppaf.pg_mask = ((1ULL << ppaf.pg_len) - 1) <<
- pblk->ppaf.pg_offset;
- pblk->ppaf.blk_mask = ((1ULL << ppaf.blk_len) - 1) <<
- pblk->ppaf.blk_offset;
+ dst->blk_len = src->blk_len;
+ dst->pg_len = src->pg_len;
+ dst->pln_len = src->pln_len;
+ dst->sect_len = src->sect_len;
- pblk->ppaf_bitsize = pblk->ppaf.blk_offset + ppaf.blk_len;
+ dst->sect_offset = 0;
+ dst->pln_offset = dst->sect_len;
+ dst->ch_offset = dst->pln_offset + dst->pln_len;
+ dst->lun_offset = dst->ch_offset + dst->ch_len;
+ dst->pg_offset = dst->lun_offset + dst->lun_len;
+ dst->blk_offset = dst->pg_offset + dst->pg_len;
+
+ dst->sec_mask = ((1ULL << dst->sect_len) - 1) << dst->sect_offset;
+ dst->pln_mask = ((1ULL << dst->pln_len) - 1) << dst->pln_offset;
+ dst->ch_mask = ((1ULL << dst->ch_len) - 1) << dst->ch_offset;
+ dst->lun_mask = ((1ULL << dst->lun_len) - 1) << dst->lun_offset;
+ dst->pg_mask = ((1ULL << dst->pg_len) - 1) << dst->pg_offset;
+ dst->blk_mask = ((1ULL << dst->blk_len) - 1) << dst->blk_offset;
+
+ return dst->blk_offset + src->blk_len;
+}
+
+static int pblk_set_ppaf(struct pblk *pblk)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ int mod;
+
+ div_u64_rem(geo->c.clba, pblk->min_write_pgs, &mod);
+ if (mod) {
+ pr_err("pblk: bad configuration of sectors/pages\n");
+ return -EINVAL;
+ }
+
+ pblk->ppaf_bitsize = pblk_set_addrf_12(geo, (void *)&pblk->ppaf);
return 0;
}
@@ -253,8 +269,7 @@ static int pblk_core_init(struct pblk *pblk)
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
- pblk->pgs_in_buffer = NVM_MEM_PAGE_WRITE * geo->sec_per_pg *
- geo->nr_planes * geo->all_luns;
+ pblk->pgs_in_buffer = geo->c.mw_cunits * geo->c.ws_opt * geo->all_luns;
if (pblk_init_global_caches(pblk))
return -ENOMEM;
@@ -433,7 +448,7 @@ static void *pblk_bb_get_log(struct pblk *pblk)
int i, nr_blks, blk_per_lun;
int ret;
- blk_per_lun = geo->nr_chks * geo->plane_mode;
+ blk_per_lun = geo->c.num_chk * geo->c.pln_mode;
nr_blks = blk_per_lun * geo->all_luns;
log = kmalloc(nr_blks, GFP_KERNEL);
@@ -551,18 +566,18 @@ static unsigned int calc_emeta_len(struct pblk *pblk)
/* Round to sector size so that lba_list starts on its own sector */
lm->emeta_sec[1] = DIV_ROUND_UP(
sizeof(struct line_emeta) + lm->blk_bitmap_len +
- sizeof(struct wa_counters), geo->sec_size);
- lm->emeta_len[1] = lm->emeta_sec[1] * geo->sec_size;
+ sizeof(struct wa_counters), geo->c.csecs);
+ lm->emeta_len[1] = lm->emeta_sec[1] * geo->c.csecs;
/* Round to sector size so that vsc_list starts on its own sector */
lm->dsec_per_line = lm->sec_per_line - lm->emeta_sec[0];
lm->emeta_sec[2] = DIV_ROUND_UP(lm->dsec_per_line * sizeof(u64),
- geo->sec_size);
- lm->emeta_len[2] = lm->emeta_sec[2] * geo->sec_size;
+ geo->c.csecs);
+ lm->emeta_len[2] = lm->emeta_sec[2] * geo->c.csecs;
lm->emeta_sec[3] = DIV_ROUND_UP(l_mg->nr_lines * sizeof(u32),
- geo->sec_size);
- lm->emeta_len[3] = lm->emeta_sec[3] * geo->sec_size;
+ geo->c.csecs);
+ lm->emeta_len[3] = lm->emeta_sec[3] * geo->c.csecs;
lm->vsc_list_len = l_mg->nr_lines * sizeof(u32);
@@ -593,13 +608,13 @@ static void pblk_set_provision(struct pblk *pblk, long nr_free_blks)
* on user capacity consider only provisioned blocks
*/
pblk->rl.total_blocks = nr_free_blks;
- pblk->rl.nr_secs = nr_free_blks * geo->sec_per_chk;
+ pblk->rl.nr_secs = nr_free_blks * geo->c.clba;
/* Consider sectors used for metadata */
sec_meta = (lm->smeta_sec + lm->emeta_sec[0]) * l_mg->nr_free_lines;
- blk_meta = DIV_ROUND_UP(sec_meta, geo->sec_per_chk);
+ blk_meta = DIV_ROUND_UP(sec_meta, geo->c.clba);
- pblk->capacity = (provisioned - blk_meta) * geo->sec_per_chk;
+ pblk->capacity = (provisioned - blk_meta) * geo->c.clba;
atomic_set(&pblk->rl.free_blocks, nr_free_blks);
atomic_set(&pblk->rl.free_user_blocks, nr_free_blks);
@@ -710,10 +725,10 @@ static int pblk_lines_init(struct pblk *pblk)
void *chunk_log;
unsigned int smeta_len, emeta_len;
long nr_bad_blks = 0, nr_free_blks = 0;
- int bb_distance, max_write_ppas, mod;
+ int bb_distance, max_write_ppas;
int i, ret;
- pblk->min_write_pgs = geo->sec_per_pl * (geo->sec_size / PAGE_SIZE);
+ pblk->min_write_pgs = geo->c.ws_opt * (geo->c.csecs / PAGE_SIZE);
max_write_ppas = pblk->min_write_pgs * geo->all_luns;
pblk->max_write_pgs = min_t(int, max_write_ppas, NVM_MAX_VLBA);
pblk_set_sec_per_write(pblk, pblk->min_write_pgs);
@@ -724,19 +739,13 @@ static int pblk_lines_init(struct pblk *pblk)
return -EINVAL;
}
- div_u64_rem(geo->sec_per_chk, pblk->min_write_pgs, &mod);
- if (mod) {
- pr_err("pblk: bad configuration of sectors/pages\n");
- return -EINVAL;
- }
-
- l_mg->nr_lines = geo->nr_chks;
+ l_mg->nr_lines = geo->c.num_chk;
l_mg->log_line = l_mg->data_line = NULL;
l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
l_mg->nr_free_lines = 0;
bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
- lm->sec_per_line = geo->sec_per_chk * geo->all_luns;
+ lm->sec_per_line = geo->c.clba * geo->all_luns;
lm->blk_per_line = geo->all_luns;
lm->blk_bitmap_len = BITS_TO_LONGS(geo->all_luns) * sizeof(long);
lm->sec_bitmap_len = BITS_TO_LONGS(lm->sec_per_line) * sizeof(long);
@@ -750,8 +759,8 @@ static int pblk_lines_init(struct pblk *pblk)
*/
i = 1;
add_smeta_page:
- lm->smeta_sec = i * geo->sec_per_pl;
- lm->smeta_len = lm->smeta_sec * geo->sec_size;
+ lm->smeta_sec = i * geo->c.ws_opt;
+ lm->smeta_len = lm->smeta_sec * geo->c.csecs;
smeta_len = sizeof(struct line_smeta) + lm->lun_bitmap_len;
if (smeta_len > lm->smeta_len) {
@@ -764,8 +773,8 @@ static int pblk_lines_init(struct pblk *pblk)
*/
i = 1;
add_emeta_page:
- lm->emeta_sec[0] = i * geo->sec_per_pl;
- lm->emeta_len[0] = lm->emeta_sec[0] * geo->sec_size;
+ lm->emeta_sec[0] = i * geo->c.ws_opt;
+ lm->emeta_len[0] = lm->emeta_sec[0] * geo->c.csecs;
emeta_len = calc_emeta_len(pblk);
if (emeta_len > lm->emeta_len[0]) {
@@ -778,7 +787,7 @@ static int pblk_lines_init(struct pblk *pblk)
lm->min_blk_line = 1;
if (geo->all_luns > 1)
lm->min_blk_line += DIV_ROUND_UP(lm->smeta_sec +
- lm->emeta_sec[0], geo->sec_per_chk);
+ lm->emeta_sec[0], geo->c.clba);
if (lm->min_blk_line > lm->blk_per_line) {
pr_err("pblk: config. not supported. Min. LUN in line:%d\n",
@@ -802,9 +811,9 @@ static int pblk_lines_init(struct pblk *pblk)
goto fail_free_bb_template;
}
- bb_distance = (geo->all_luns) * geo->sec_per_pl;
+ bb_distance = (geo->all_luns) * geo->c.ws_opt;
for (i = 0; i < lm->sec_per_line; i += bb_distance)
- bitmap_set(l_mg->bb_template, i, geo->sec_per_pl);
+ bitmap_set(l_mg->bb_template, i, geo->c.ws_opt);
INIT_LIST_HEAD(&l_mg->free_list);
INIT_LIST_HEAD(&l_mg->corrupt_list);
@@ -981,9 +990,9 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
struct pblk *pblk;
int ret;
- if (dev->identity.dom & NVM_RSP_L2P) {
+ if (dev->geo.c.dom & NVM_RSP_L2P) {
pr_err("pblk: host-side L2P table not supported. (%x)\n",
- dev->identity.dom);
+ dev->geo.c.dom);
return ERR_PTR(-EINVAL);
}
@@ -1091,7 +1100,7 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
blk_queue_write_cache(tqueue, true, false);
- tqueue->limits.discard_granularity = geo->sec_per_chk * geo->sec_size;
+ tqueue->limits.discard_granularity = geo->c.clba * geo->c.csecs;
tqueue->limits.discard_alignment = 0;
blk_queue_max_discard_sectors(tqueue, UINT_MAX >> 9);
queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, tqueue);
diff --git a/drivers/lightnvm/pblk-read.c b/drivers/lightnvm/pblk-read.c
index 2f761283f43e..ebb6bae3a3b8 100644
--- a/drivers/lightnvm/pblk-read.c
+++ b/drivers/lightnvm/pblk-read.c
@@ -563,7 +563,7 @@ int pblk_submit_read_gc(struct pblk *pblk, struct pblk_gc_rq *gc_rq)
if (!(gc_rq->secs_to_gc))
goto out;
- data_len = (gc_rq->secs_to_gc) * geo->sec_size;
+ data_len = (gc_rq->secs_to_gc) * geo->c.csecs;
bio = pblk_bio_map_addr(pblk, gc_rq->data, gc_rq->secs_to_gc, data_len,
PBLK_VMALLOC_META, GFP_KERNEL);
if (IS_ERR(bio)) {
diff --git a/drivers/lightnvm/pblk-recovery.c b/drivers/lightnvm/pblk-recovery.c
index aaab9a5c17cc..1574dbbfbb1c 100644
--- a/drivers/lightnvm/pblk-recovery.c
+++ b/drivers/lightnvm/pblk-recovery.c
@@ -184,7 +184,7 @@ static int pblk_calc_sec_in_line(struct pblk *pblk, struct pblk_line *line)
int nr_bb = bitmap_weight(line->blk_bitmap, lm->blk_per_line);
return lm->sec_per_line - lm->smeta_sec - lm->emeta_sec[0] -
- nr_bb * geo->sec_per_chk;
+ nr_bb * geo->c.clba;
}
struct pblk_recov_alloc {
@@ -232,7 +232,7 @@ static int pblk_recov_read_oob(struct pblk *pblk, struct pblk_line *line,
rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
if (!rq_ppas)
rq_ppas = pblk->min_write_pgs;
- rq_len = rq_ppas * geo->sec_size;
+ rq_len = rq_ppas * geo->c.csecs;
bio = bio_map_kern(dev->q, data, rq_len, GFP_KERNEL);
if (IS_ERR(bio))
@@ -351,7 +351,7 @@ static int pblk_recov_pad_oob(struct pblk *pblk, struct pblk_line *line,
if (!pad_rq)
return -ENOMEM;
- data = vzalloc(pblk->max_write_pgs * geo->sec_size);
+ data = vzalloc(pblk->max_write_pgs * geo->c.csecs);
if (!data) {
ret = -ENOMEM;
goto free_rq;
@@ -368,7 +368,7 @@ static int pblk_recov_pad_oob(struct pblk *pblk, struct pblk_line *line,
goto fail_free_pad;
}
- rq_len = rq_ppas * geo->sec_size;
+ rq_len = rq_ppas * geo->c.csecs;
meta_list = nvm_dev_dma_alloc(dev->parent, GFP_KERNEL, &dma_meta_list);
if (!meta_list) {
@@ -509,7 +509,7 @@ static int pblk_recov_scan_all_oob(struct pblk *pblk, struct pblk_line *line,
rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
if (!rq_ppas)
rq_ppas = pblk->min_write_pgs;
- rq_len = rq_ppas * geo->sec_size;
+ rq_len = rq_ppas * geo->c.csecs;
bio = bio_map_kern(dev->q, data, rq_len, GFP_KERNEL);
if (IS_ERR(bio))
@@ -640,7 +640,7 @@ static int pblk_recov_scan_oob(struct pblk *pblk, struct pblk_line *line,
rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
if (!rq_ppas)
rq_ppas = pblk->min_write_pgs;
- rq_len = rq_ppas * geo->sec_size;
+ rq_len = rq_ppas * geo->c.csecs;
bio = bio_map_kern(dev->q, data, rq_len, GFP_KERNEL);
if (IS_ERR(bio))
@@ -745,7 +745,7 @@ static int pblk_recov_l2p_from_oob(struct pblk *pblk, struct pblk_line *line)
ppa_list = (void *)(meta_list) + pblk_dma_meta_size;
dma_ppa_list = dma_meta_list + pblk_dma_meta_size;
- data = kcalloc(pblk->max_write_pgs, geo->sec_size, GFP_KERNEL);
+ data = kcalloc(pblk->max_write_pgs, geo->c.csecs, GFP_KERNEL);
if (!data) {
ret = -ENOMEM;
goto free_meta_list;
diff --git a/drivers/lightnvm/pblk-rl.c b/drivers/lightnvm/pblk-rl.c
index 0d457b162f23..bcab203477ec 100644
--- a/drivers/lightnvm/pblk-rl.c
+++ b/drivers/lightnvm/pblk-rl.c
@@ -200,7 +200,7 @@ void pblk_rl_init(struct pblk_rl *rl, int budget)
/* Consider sectors used for metadata */
sec_meta = (lm->smeta_sec + lm->emeta_sec[0]) * l_mg->nr_free_lines;
- blk_meta = DIV_ROUND_UP(sec_meta, geo->sec_per_chk);
+ blk_meta = DIV_ROUND_UP(sec_meta, geo->c.clba);
rl->high = pblk->op_blks - blk_meta - lm->blk_per_line;
rl->high_pw = get_count_order(rl->high);
diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
index d93e9b1f083a..5eb21a279361 100644
--- a/drivers/lightnvm/pblk-sysfs.c
+++ b/drivers/lightnvm/pblk-sysfs.c
@@ -113,26 +113,31 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
+ struct nvm_addr_format_12 *ppaf;
+ struct nvm_addr_format_12 *geo_ppaf;
ssize_t sz = 0;
- sz = snprintf(page, PAGE_SIZE - sz,
- "g:(b:%d)blk:%d/%d,pg:%d/%d,lun:%d/%d,ch:%d/%d,pl:%d/%d,sec:%d/%d\n",
- pblk->ppaf_bitsize,
- pblk->ppaf.blk_offset, geo->ppaf.blk_len,
- pblk->ppaf.pg_offset, geo->ppaf.pg_len,
- pblk->ppaf.lun_offset, geo->ppaf.lun_len,
- pblk->ppaf.ch_offset, geo->ppaf.ch_len,
- pblk->ppaf.pln_offset, geo->ppaf.pln_len,
- pblk->ppaf.sec_offset, geo->ppaf.sect_len);
+ ppaf = (struct nvm_addr_format_12 *)&pblk->ppaf;
+ geo_ppaf = (struct nvm_addr_format_12 *)&geo->c.addrf;
+
+ sz = snprintf(page, PAGE_SIZE,
+ "pblk:(s:%d)ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
+ pblk->ppaf_bitsize,
+ ppaf->ch_offset, ppaf->ch_len,
+ ppaf->lun_offset, ppaf->lun_len,
+ ppaf->blk_offset, ppaf->blk_len,
+ ppaf->pg_offset, ppaf->pg_len,
+ ppaf->pln_offset, ppaf->pln_len,
+ ppaf->sect_offset, ppaf->sect_len);
sz += snprintf(page + sz, PAGE_SIZE - sz,
- "d:blk:%d/%d,pg:%d/%d,lun:%d/%d,ch:%d/%d,pl:%d/%d,sec:%d/%d\n",
- geo->ppaf.blk_offset, geo->ppaf.blk_len,
- geo->ppaf.pg_offset, geo->ppaf.pg_len,
- geo->ppaf.lun_offset, geo->ppaf.lun_len,
- geo->ppaf.ch_offset, geo->ppaf.ch_len,
- geo->ppaf.pln_offset, geo->ppaf.pln_len,
- geo->ppaf.sect_offset, geo->ppaf.sect_len);
+ "device:ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
+ geo_ppaf->ch_offset, geo_ppaf->ch_len,
+ geo_ppaf->lun_offset, geo_ppaf->lun_len,
+ geo_ppaf->blk_offset, geo_ppaf->blk_len,
+ geo_ppaf->pg_offset, geo_ppaf->pg_len,
+ geo_ppaf->pln_offset, geo_ppaf->pln_len,
+ geo_ppaf->sect_offset, geo_ppaf->sect_len);
return sz;
}
@@ -288,7 +293,7 @@ static ssize_t pblk_sysfs_lines_info(struct pblk *pblk, char *page)
"blk_line:%d, sec_line:%d, sec_blk:%d\n",
lm->blk_per_line,
lm->sec_per_line,
- geo->sec_per_chk);
+ geo->c.clba);
return sz;
}
diff --git a/drivers/lightnvm/pblk-write.c b/drivers/lightnvm/pblk-write.c
index aae86ed60b98..c49b27539d5a 100644
--- a/drivers/lightnvm/pblk-write.c
+++ b/drivers/lightnvm/pblk-write.c
@@ -333,7 +333,7 @@ int pblk_submit_meta_io(struct pblk *pblk, struct pblk_line *meta_line)
m_ctx = nvm_rq_to_pdu(rqd);
m_ctx->private = meta_line;
- rq_len = rq_ppas * geo->sec_size;
+ rq_len = rq_ppas * geo->c.csecs;
data = ((void *)emeta->buf) + emeta->mem;
bio = pblk_bio_map_addr(pblk, data, rq_ppas, rq_len,
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index 282dfc8780e8..67ffb53608f7 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -551,21 +551,6 @@ struct pblk_line_meta {
unsigned int meta_distance; /* Distance between data and metadata */
};
-struct pblk_addr_format {
- u64 ch_mask;
- u64 lun_mask;
- u64 pln_mask;
- u64 blk_mask;
- u64 pg_mask;
- u64 sec_mask;
- u8 ch_offset;
- u8 lun_offset;
- u8 pln_offset;
- u8 blk_offset;
- u8 pg_offset;
- u8 sec_offset;
-};
-
enum {
PBLK_STATE_RUNNING = 0,
PBLK_STATE_STOPPING = 1,
@@ -585,8 +570,8 @@ struct pblk {
struct pblk_line_mgmt l_mg; /* Line management */
struct pblk_line_meta lm; /* Line metadata */
+ struct nvm_addr_format ppaf;
int ppaf_bitsize;
- struct pblk_addr_format ppaf;
struct pblk_rb rwb;
@@ -941,14 +926,12 @@ static inline int pblk_line_vsc(struct pblk_line *line)
return le32_to_cpu(*line->vsc);
}
-#define NVM_MEM_PAGE_WRITE (8)
-
static inline int pblk_pad_distance(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
- return NVM_MEM_PAGE_WRITE * geo->all_luns * geo->sec_per_pl;
+ return geo->c.mw_cunits * geo->all_luns * geo->c.ws_opt;
}
static inline int pblk_ppa_to_line(struct ppa_addr p)
@@ -964,15 +947,17 @@ static inline int pblk_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
u64 line_id)
{
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->ppaf;
struct ppa_addr ppa;
ppa.ppa = 0;
ppa.g.blk = line_id;
- ppa.g.pg = (paddr & pblk->ppaf.pg_mask) >> pblk->ppaf.pg_offset;
- ppa.g.lun = (paddr & pblk->ppaf.lun_mask) >> pblk->ppaf.lun_offset;
- ppa.g.ch = (paddr & pblk->ppaf.ch_mask) >> pblk->ppaf.ch_offset;
- ppa.g.pl = (paddr & pblk->ppaf.pln_mask) >> pblk->ppaf.pln_offset;
- ppa.g.sec = (paddr & pblk->ppaf.sec_mask) >> pblk->ppaf.sec_offset;
+ ppa.g.pg = (paddr & ppaf->pg_mask) >> ppaf->pg_offset;
+ ppa.g.lun = (paddr & ppaf->lun_mask) >> ppaf->lun_offset;
+ ppa.g.ch = (paddr & ppaf->ch_mask) >> ppaf->ch_offset;
+ ppa.g.pl = (paddr & ppaf->pln_mask) >> ppaf->pln_offset;
+ ppa.g.sec = (paddr & ppaf->sec_mask) >> ppaf->sect_offset;
return ppa;
}
@@ -980,13 +965,15 @@ static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
static inline u64 pblk_dev_ppa_to_line_addr(struct pblk *pblk,
struct ppa_addr p)
{
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->ppaf;
u64 paddr;
- paddr = (u64)p.g.pg << pblk->ppaf.pg_offset;
- paddr |= (u64)p.g.lun << pblk->ppaf.lun_offset;
- paddr |= (u64)p.g.ch << pblk->ppaf.ch_offset;
- paddr |= (u64)p.g.pl << pblk->ppaf.pln_offset;
- paddr |= (u64)p.g.sec << pblk->ppaf.sec_offset;
+ paddr = (u64)p.g.ch << ppaf->ch_offset;
+ paddr |= (u64)p.g.lun << ppaf->lun_offset;
+ paddr |= (u64)p.g.pg << ppaf->pg_offset;
+ paddr |= (u64)p.g.pl << ppaf->pln_offset;
+ paddr |= (u64)p.g.sec << ppaf->sect_offset;
return paddr;
}
@@ -1003,18 +990,15 @@ static inline struct ppa_addr pblk_ppa32_to_ppa64(struct pblk *pblk, u32 ppa32)
ppa64.c.line = ppa32 & ((~0U) >> 1);
ppa64.c.is_cached = 1;
} else {
- ppa64.g.blk = (ppa32 & pblk->ppaf.blk_mask) >>
- pblk->ppaf.blk_offset;
- ppa64.g.pg = (ppa32 & pblk->ppaf.pg_mask) >>
- pblk->ppaf.pg_offset;
- ppa64.g.lun = (ppa32 & pblk->ppaf.lun_mask) >>
- pblk->ppaf.lun_offset;
- ppa64.g.ch = (ppa32 & pblk->ppaf.ch_mask) >>
- pblk->ppaf.ch_offset;
- ppa64.g.pl = (ppa32 & pblk->ppaf.pln_mask) >>
- pblk->ppaf.pln_offset;
- ppa64.g.sec = (ppa32 & pblk->ppaf.sec_mask) >>
- pblk->ppaf.sec_offset;
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->ppaf;
+
+ ppa64.g.ch = (ppa32 & ppaf->ch_mask) >> ppaf->ch_offset;
+ ppa64.g.lun = (ppa32 & ppaf->lun_mask) >> ppaf->lun_offset;
+ ppa64.g.blk = (ppa32 & ppaf->blk_mask) >> ppaf->blk_offset;
+ ppa64.g.pg = (ppa32 & ppaf->pg_mask) >> ppaf->pg_offset;
+ ppa64.g.pl = (ppa32 & ppaf->pln_mask) >> ppaf->pln_offset;
+ ppa64.g.sec = (ppa32 & ppaf->sec_mask) >> ppaf->sect_offset;
}
return ppa64;
@@ -1030,12 +1014,15 @@ static inline u32 pblk_ppa64_to_ppa32(struct pblk *pblk, struct ppa_addr ppa64)
ppa32 |= ppa64.c.line;
ppa32 |= 1U << 31;
} else {
- ppa32 |= ppa64.g.blk << pblk->ppaf.blk_offset;
- ppa32 |= ppa64.g.pg << pblk->ppaf.pg_offset;
- ppa32 |= ppa64.g.lun << pblk->ppaf.lun_offset;
- ppa32 |= ppa64.g.ch << pblk->ppaf.ch_offset;
- ppa32 |= ppa64.g.pl << pblk->ppaf.pln_offset;
- ppa32 |= ppa64.g.sec << pblk->ppaf.sec_offset;
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->ppaf;
+
+ ppa32 |= ppa64.g.ch << ppaf->ch_offset;
+ ppa32 |= ppa64.g.lun << ppaf->lun_offset;
+ ppa32 |= ppa64.g.blk << ppaf->blk_offset;
+ ppa32 |= ppa64.g.pg << ppaf->pg_offset;
+ ppa32 |= ppa64.g.pl << ppaf->pln_offset;
+ ppa32 |= ppa64.g.sec << ppaf->sect_offset;
}
return ppa32;
@@ -1153,7 +1140,7 @@ static inline int pblk_set_progr_mode(struct pblk *pblk, int type)
struct nvm_geo *geo = &dev->geo;
int flags;
- flags = geo->plane_mode >> 1;
+ flags = geo->c.pln_mode >> 1;
if (type == PBLK_WRITE)
flags |= NVM_IO_SCRAMBLE_ENABLE;
@@ -1174,7 +1161,7 @@ static inline int pblk_set_read_mode(struct pblk *pblk, int type)
flags = NVM_IO_SUSPEND | NVM_IO_SCRAMBLE_ENABLE;
if (type == PBLK_READ_SEQUENTIAL)
- flags |= geo->plane_mode >> 1;
+ flags |= geo->c.pln_mode >> 1;
return flags;
}
@@ -1229,10 +1216,10 @@ static inline int pblk_boundary_ppa_checks(struct nvm_tgt_dev *tgt_dev,
if (!ppa->c.is_cached &&
ppa->g.ch < geo->nr_chnls &&
ppa->g.lun < geo->nr_luns &&
- ppa->g.pl < geo->nr_planes &&
- ppa->g.blk < geo->nr_chks &&
- ppa->g.pg < geo->ws_per_chk &&
- ppa->g.sec < geo->sec_per_pg)
+ ppa->g.pl < geo->c.num_pln &&
+ ppa->g.blk < geo->c.num_chk &&
+ ppa->g.pg < geo->c.num_pg &&
+ ppa->g.sec < geo->c.ws_min)
continue;
print_ppa(ppa, "boundary", i);
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 839c0b96466a..8f541331b1a9 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -152,8 +152,8 @@ struct nvme_nvm_id12_addrf {
__u8 blk_len;
__u8 pg_offset;
__u8 pg_len;
- __u8 sect_offset;
- __u8 sect_len;
+ __u8 sec_offset;
+ __u8 sec_len;
__u8 res[4];
} __packed;
@@ -254,106 +254,164 @@ static inline void _nvme_nvm_check_size(void)
BUILD_BUG_ON(sizeof(struct nvme_nvm_id20) != NVME_IDENTIFY_DATA_SIZE);
}
-static int init_grp(struct nvm_id *nvm_id, struct nvme_nvm_id12 *id12)
+static void nvme_nvm_set_addr_12(struct nvm_addr_format_12 *dst,
+ struct nvme_nvm_id12_addrf *src)
+{
+ dst->ch_len = src->ch_len;
+ dst->lun_len = src->lun_len;
+ dst->blk_len = src->blk_len;
+ dst->pg_len = src->pg_len;
+ dst->pln_len = src->pln_len;
+ dst->sect_len = src->sec_len;
+
+ dst->ch_offset = src->ch_offset;
+ dst->lun_offset = src->lun_offset;
+ dst->blk_offset = src->blk_offset;
+ dst->pg_offset = src->pg_offset;
+ dst->pln_offset = src->pln_offset;
+ dst->sect_offset = src->sec_offset;
+
+ dst->ch_mask = ((1ULL << dst->ch_len) - 1) << dst->ch_offset;
+ dst->lun_mask = ((1ULL << dst->lun_len) - 1) << dst->lun_offset;
+ dst->blk_mask = ((1ULL << dst->blk_len) - 1) << dst->blk_offset;
+ dst->pg_mask = ((1ULL << dst->pg_len) - 1) << dst->pg_offset;
+ dst->pln_mask = ((1ULL << dst->pln_len) - 1) << dst->pln_offset;
+ dst->sec_mask = ((1ULL << dst->sect_len) - 1) << dst->sect_offset;
+}
+
+static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
+ struct nvm_geo *dev_geo)
{
struct nvme_nvm_id12_grp *src;
int sec_per_pg, sec_per_pl, pg_per_blk;
- if (id12->cgrps != 1)
+ if (id->cgrps != 1)
return -EINVAL;
- src = &id12->grp;
+ src = &id->grp;
- nvm_id->mtype = src->mtype;
- nvm_id->fmtype = src->fmtype;
+ if (src->mtype != 0) {
+ pr_err("nvm: memory type not supported\n");
+ return -EINVAL;
+ }
+
+ dev_geo->ver_id = id->ver_id;
+
+ dev_geo->nr_chnls = src->num_ch;
+ dev_geo->nr_luns = src->num_lun;
+ dev_geo->all_luns = dev_geo->nr_chnls * dev_geo->nr_luns;
- nvm_id->num_ch = src->num_ch;
- nvm_id->num_lun = src->num_lun;
+ dev_geo->c.num_chk = le16_to_cpu(src->num_chk);
- nvm_id->num_chk = le16_to_cpu(src->num_chk);
- nvm_id->csecs = le16_to_cpu(src->csecs);
- nvm_id->sos = le16_to_cpu(src->sos);
+ dev_geo->c.csecs = le16_to_cpu(src->csecs);
+ dev_geo->c.sos = le16_to_cpu(src->sos);
pg_per_blk = le16_to_cpu(src->num_pg);
- sec_per_pg = le16_to_cpu(src->fpg_sz) / nvm_id->csecs;
+ sec_per_pg = le16_to_cpu(src->fpg_sz) / dev_geo->c.csecs;
sec_per_pl = sec_per_pg * src->num_pln;
- nvm_id->clba = sec_per_pl * pg_per_blk;
- nvm_id->ws_per_chk = pg_per_blk;
-
- nvm_id->mpos = le32_to_cpu(src->mpos);
- nvm_id->cpar = le16_to_cpu(src->cpar);
- nvm_id->mccap = le32_to_cpu(src->mccap);
-
- nvm_id->ws_opt = nvm_id->ws_min = sec_per_pg;
- nvm_id->ws_seq = NVM_IO_SNGL_ACCESS;
-
- if (nvm_id->mpos & 0x020202) {
- nvm_id->ws_seq = NVM_IO_DUAL_ACCESS;
- nvm_id->ws_opt <<= 1;
- } else if (nvm_id->mpos & 0x040404) {
- nvm_id->ws_seq = NVM_IO_QUAD_ACCESS;
- nvm_id->ws_opt <<= 2;
+ dev_geo->c.clba = sec_per_pl * pg_per_blk;
+
+ dev_geo->all_chunks = dev_geo->all_luns * dev_geo->c.num_chk;
+ dev_geo->total_secs = dev_geo->c.clba * dev_geo->all_chunks;
+
+ dev_geo->c.ws_min = sec_per_pg;
+ dev_geo->c.ws_opt = sec_per_pg;
+ dev_geo->c.mw_cunits = 8; /* default to MLC safe values */
+
+ dev_geo->c.mccap = le32_to_cpu(src->mccap);
+
+ dev_geo->c.trdt = le32_to_cpu(src->trdt);
+ dev_geo->c.trdm = le32_to_cpu(src->trdm);
+ dev_geo->c.tprt = le32_to_cpu(src->tprt);
+ dev_geo->c.tprm = le32_to_cpu(src->tprm);
+ dev_geo->c.tbet = le32_to_cpu(src->tbet);
+ dev_geo->c.tbem = le32_to_cpu(src->tbem);
+
+ /* 1.2 compatibility */
+ dev_geo->c.vmnt = id->vmnt;
+ dev_geo->c.cap = le32_to_cpu(id->cap);
+ dev_geo->c.dom = le32_to_cpu(id->dom);
+
+ dev_geo->c.mtype = src->mtype;
+ dev_geo->c.fmtype = src->fmtype;
+
+ dev_geo->c.cpar = le16_to_cpu(src->cpar);
+ dev_geo->c.mpos = le32_to_cpu(src->mpos);
+
+ dev_geo->c.pln_mode = NVM_PLANE_SINGLE;
+
+ if (dev_geo->c.mpos & 0x020202) {
+ dev_geo->c.pln_mode = NVM_PLANE_DOUBLE;
+ dev_geo->c.ws_opt <<= 1;
+ } else if (dev_geo->c.mpos & 0x040404) {
+ dev_geo->c.pln_mode = NVM_PLANE_QUAD;
+ dev_geo->c.ws_opt <<= 2;
}
- nvm_id->trdt = le32_to_cpu(src->trdt);
- nvm_id->trdm = le32_to_cpu(src->trdm);
- nvm_id->tprt = le32_to_cpu(src->tprt);
- nvm_id->tprm = le32_to_cpu(src->tprm);
- nvm_id->tbet = le32_to_cpu(src->tbet);
- nvm_id->tbem = le32_to_cpu(src->tbem);
-
- /* 1.2 compatibility */
- nvm_id->num_pln = src->num_pln;
- nvm_id->num_pg = le16_to_cpu(src->num_pg);
- nvm_id->fpg_sz = le16_to_cpu(src->fpg_sz);
+ dev_geo->c.num_pln = src->num_pln;
+ dev_geo->c.num_pg = le16_to_cpu(src->num_pg);
+ dev_geo->c.fpg_sz = le16_to_cpu(src->fpg_sz);
+
+ nvme_nvm_set_addr_12((struct nvm_addr_format_12 *)&dev_geo->c.addrf,
+ &id->ppaf);
return 0;
}
-static int nvme_nvm_setup_12(struct nvm_dev *nvmdev, struct nvm_id *nvm_id,
- struct nvme_nvm_id12 *id)
+static void nvme_nvm_set_addr_20(struct nvm_addr_format *dst,
+ struct nvme_nvm_id20_addrf *src)
{
- nvm_id->ver_id = id->ver_id;
- nvm_id->vmnt = id->vmnt;
- nvm_id->cap = le32_to_cpu(id->cap);
- nvm_id->dom = le32_to_cpu(id->dom);
- memcpy(&nvm_id->ppaf, &id->ppaf,
- sizeof(struct nvm_addr_format));
-
- return init_grp(nvm_id, id);
+ dst->ch_len = src->grp_len;
+ dst->lun_len = src->pu_len;
+ dst->chk_len = src->chk_len;
+ dst->sec_len = src->lba_len;
+
+ dst->sec_offset = 0;
+ dst->chk_offset = dst->sec_len;
+ dst->lun_offset = dst->chk_offset + dst->chk_len;
+ dst->ch_offset = dst->lun_offset + dst->lun_len;
+
+ dst->ch_mask = ((1ULL << dst->ch_len) - 1) << dst->ch_offset;
+ dst->lun_mask = ((1ULL << dst->lun_len) - 1) << dst->lun_offset;
+ dst->chk_mask = ((1ULL << dst->chk_len) - 1) << dst->chk_offset;
+ dst->sec_mask = ((1ULL << dst->sec_len) - 1) << dst->sec_offset;
}
-static int nvme_nvm_setup_20(struct nvm_dev *nvmdev, struct nvm_id *nvm_id,
- struct nvme_nvm_id20 *id)
+static int nvme_nvm_setup_20(struct nvme_nvm_id20 *id,
+ struct nvm_geo *dev_geo)
{
- nvm_id->ver_id = id->mjr;
+ dev_geo->ver_id = id->mjr;
+
+ dev_geo->nr_chnls = le16_to_cpu(id->num_grp);
+ dev_geo->nr_luns = le16_to_cpu(id->num_pu);
+ dev_geo->all_luns = dev_geo->nr_chnls * dev_geo->nr_luns;
+
+ dev_geo->c.num_chk = le32_to_cpu(id->num_chk);
+ dev_geo->c.clba = le32_to_cpu(id->clba);
- nvm_id->num_ch = le16_to_cpu(id->num_grp);
- nvm_id->num_lun = le16_to_cpu(id->num_pu);
- nvm_id->num_chk = le32_to_cpu(id->num_chk);
- nvm_id->clba = le32_to_cpu(id->clba);
+ dev_geo->all_chunks = dev_geo->all_luns * dev_geo->c.num_chk;
+ dev_geo->total_secs = dev_geo->c.clba * dev_geo->all_chunks;
- nvm_id->ws_min = le32_to_cpu(id->ws_min);
- nvm_id->ws_opt = le32_to_cpu(id->ws_opt);
- nvm_id->mw_cunits = le32_to_cpu(id->mw_cunits);
+ dev_geo->c.csecs = -1; /* Set by nvme identify */
+ dev_geo->c.sos = -1; /* Set bu nvme identify */
- nvm_id->trdt = le32_to_cpu(id->trdt);
- nvm_id->trdm = le32_to_cpu(id->trdm);
- nvm_id->tprt = le32_to_cpu(id->twrt);
- nvm_id->tprm = le32_to_cpu(id->twrm);
- nvm_id->tbet = le32_to_cpu(id->tcrst);
- nvm_id->tbem = le32_to_cpu(id->tcrsm);
+ dev_geo->c.ws_min = le32_to_cpu(id->ws_min);
+ dev_geo->c.ws_opt = le32_to_cpu(id->ws_opt);
+ dev_geo->c.mw_cunits = le32_to_cpu(id->mw_cunits);
- /* calculated values */
- nvm_id->ws_per_chk = nvm_id->clba / nvm_id->ws_min;
+ dev_geo->c.trdt = le32_to_cpu(id->trdt);
+ dev_geo->c.trdm = le32_to_cpu(id->trdm);
+ dev_geo->c.tprt = le32_to_cpu(id->twrt);
+ dev_geo->c.tprm = le32_to_cpu(id->twrm);
+ dev_geo->c.tbet = le32_to_cpu(id->tcrst);
+ dev_geo->c.tbem = le32_to_cpu(id->tcrsm);
- /* 1.2 compatibility */
- nvm_id->ws_seq = NVM_IO_SNGL_ACCESS;
+ nvme_nvm_set_addr_20(&dev_geo->c.addrf, &id->lbaf);
return 0;
}
-static int nvme_nvm_identity(struct nvm_dev *nvmdev, struct nvm_id *nvm_id)
+static int nvme_nvm_identity(struct nvm_dev *nvmdev)
{
struct nvme_ns *ns = nvmdev->q->queuedata;
struct nvme_nvm_id12 *id;
@@ -380,18 +438,18 @@ static int nvme_nvm_identity(struct nvm_dev *nvmdev, struct nvm_id *nvm_id)
*/
switch (id->ver_id) {
case 1:
- ret = nvme_nvm_setup_12(nvmdev, nvm_id, id);
+ ret = nvme_nvm_setup_12(id, &nvmdev->geo);
break;
case 2:
- ret = nvme_nvm_setup_20(nvmdev, nvm_id,
- (struct nvme_nvm_id20 *)id);
+ ret = nvme_nvm_setup_20((struct nvme_nvm_id20 *)id,
+ &nvmdev->geo);
break;
default:
- dev_err(ns->ctrl->device,
- "OCSSD revision not supported (%d)\n",
- nvm_id->ver_id);
+ dev_err(ns->ctrl->device, "OCSSD revision not supported (%d)\n",
+ id->ver_id);
ret = -EINVAL;
}
+
out:
kfree(id);
return ret;
@@ -401,12 +459,12 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
u8 *blks)
{
struct request_queue *q = nvmdev->q;
- struct nvm_geo *geo = &nvmdev->geo;
+ struct nvm_geo *dev_geo = &nvmdev->geo;
struct nvme_ns *ns = q->queuedata;
struct nvme_ctrl *ctrl = ns->ctrl;
struct nvme_nvm_command c = {};
struct nvme_nvm_bb_tbl *bb_tbl;
- int nr_blks = geo->nr_chks * geo->plane_mode;
+ int nr_blks = dev_geo->c.num_chk * dev_geo->c.num_pln;
int tblsz = sizeof(struct nvme_nvm_bb_tbl) + nr_blks;
int ret = 0;
@@ -447,7 +505,7 @@ static int nvme_nvm_get_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr ppa,
goto out;
}
- memcpy(blks, bb_tbl->blk, geo->nr_chks * geo->plane_mode);
+ memcpy(blks, bb_tbl->blk, dev_geo->c.num_chk * dev_geo->c.num_pln);
out:
kfree(bb_tbl);
return ret;
@@ -815,9 +873,10 @@ int nvme_nvm_ioctl(struct nvme_ns *ns, unsigned int cmd, unsigned long arg)
void nvme_nvm_update_nvm_info(struct nvme_ns *ns)
{
struct nvm_dev *ndev = ns->ndev;
+ struct nvm_geo *dev_geo = &ndev->geo;
- ndev->identity.csecs = ndev->geo.sec_size = 1 << ns->lba_shift;
- ndev->identity.sos = ndev->geo.oob_size = ns->ms;
+ dev_geo->c.csecs = 1 << ns->lba_shift;
+ dev_geo->c.sos = ns->ms;
}
int nvme_nvm_register(struct nvme_ns *ns, char *disk_name, int node)
@@ -850,23 +909,22 @@ static ssize_t nvm_dev_attr_show(struct device *dev,
{
struct nvme_ns *ns = nvme_get_ns_from_dev(dev);
struct nvm_dev *ndev = ns->ndev;
- struct nvm_id *id;
+ struct nvm_geo *dev_geo = &ndev->geo;
struct attribute *attr;
if (!ndev)
return 0;
- id = &ndev->identity;
attr = &dattr->attr;
if (strcmp(attr->name, "version") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->ver_id);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->ver_id);
} else if (strcmp(attr->name, "capabilities") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->cap);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.cap);
} else if (strcmp(attr->name, "read_typ") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->trdt);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.trdt);
} else if (strcmp(attr->name, "read_max") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->trdm);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.trdm);
} else {
return scnprintf(page,
PAGE_SIZE,
@@ -875,75 +933,79 @@ static ssize_t nvm_dev_attr_show(struct device *dev,
}
}
+static ssize_t nvm_dev_attr_show_ppaf(struct nvm_addr_format_12 *ppaf,
+ char *page)
+{
+ return scnprintf(page, PAGE_SIZE,
+ "0x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x\n",
+ ppaf->ch_offset, ppaf->ch_len,
+ ppaf->lun_offset, ppaf->lun_len,
+ ppaf->pln_offset, ppaf->pln_len,
+ ppaf->blk_offset, ppaf->blk_len,
+ ppaf->pg_offset, ppaf->pg_len,
+ ppaf->sect_offset, ppaf->sect_len);
+}
+
static ssize_t nvm_dev_attr_show_12(struct device *dev,
struct device_attribute *dattr, char *page)
{
struct nvme_ns *ns = nvme_get_ns_from_dev(dev);
struct nvm_dev *ndev = ns->ndev;
- struct nvm_id *id;
+ struct nvm_geo *dev_geo = &ndev->geo;
struct attribute *attr;
if (!ndev)
return 0;
- id = &ndev->identity;
attr = &dattr->attr;
if (strcmp(attr->name, "vendor_opcode") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->vmnt);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.vmnt);
} else if (strcmp(attr->name, "device_mode") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->dom);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.dom);
/* kept for compatibility */
} else if (strcmp(attr->name, "media_manager") == 0) {
return scnprintf(page, PAGE_SIZE, "%s\n", "gennvm");
} else if (strcmp(attr->name, "ppa_format") == 0) {
- return scnprintf(page, PAGE_SIZE,
- "0x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x\n",
- id->ppaf.ch_offset, id->ppaf.ch_len,
- id->ppaf.lun_offset, id->ppaf.lun_len,
- id->ppaf.pln_offset, id->ppaf.pln_len,
- id->ppaf.blk_offset, id->ppaf.blk_len,
- id->ppaf.pg_offset, id->ppaf.pg_len,
- id->ppaf.sect_offset, id->ppaf.sect_len);
+ return nvm_dev_attr_show_ppaf((void *)&dev_geo->c.addrf, page);
} else if (strcmp(attr->name, "media_type") == 0) { /* u8 */
- return scnprintf(page, PAGE_SIZE, "%u\n", id->mtype);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.mtype);
} else if (strcmp(attr->name, "flash_media_type") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->fmtype);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.fmtype);
} else if (strcmp(attr->name, "num_channels") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_ch);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_chnls);
} else if (strcmp(attr->name, "num_luns") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_lun);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_luns);
} else if (strcmp(attr->name, "num_planes") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_pln);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.num_pln);
} else if (strcmp(attr->name, "num_blocks") == 0) { /* u16 */
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_chk);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.num_chk);
} else if (strcmp(attr->name, "num_pages") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_pg);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.num_pg);
} else if (strcmp(attr->name, "page_size") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->fpg_sz);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.fpg_sz);
} else if (strcmp(attr->name, "hw_sector_size") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->csecs);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.csecs);
} else if (strcmp(attr->name, "oob_sector_size") == 0) {/* u32 */
- return scnprintf(page, PAGE_SIZE, "%u\n", id->sos);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.sos);
} else if (strcmp(attr->name, "prog_typ") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tprt);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tprt);
} else if (strcmp(attr->name, "prog_max") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tprm);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tprm);
} else if (strcmp(attr->name, "erase_typ") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tbet);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tbet);
} else if (strcmp(attr->name, "erase_max") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tbem);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tbem);
} else if (strcmp(attr->name, "multiplane_modes") == 0) {
- return scnprintf(page, PAGE_SIZE, "0x%08x\n", id->mpos);
+ return scnprintf(page, PAGE_SIZE, "0x%08x\n", dev_geo->c.mpos);
} else if (strcmp(attr->name, "media_capabilities") == 0) {
- return scnprintf(page, PAGE_SIZE, "0x%08x\n", id->mccap);
+ return scnprintf(page, PAGE_SIZE, "0x%08x\n", dev_geo->c.mccap);
} else if (strcmp(attr->name, "max_phys_secs") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", NVM_MAX_VLBA);
} else {
- return scnprintf(page,
- PAGE_SIZE,
- "Unhandled attr(%s) in `nvm_dev_attr_show_12`\n",
- attr->name);
+ return scnprintf(page, PAGE_SIZE,
+ "Unhandled attr(%s) in `nvm_dev_attr_show_12`\n",
+ attr->name);
}
}
@@ -952,42 +1014,40 @@ static ssize_t nvm_dev_attr_show_20(struct device *dev,
{
struct nvme_ns *ns = nvme_get_ns_from_dev(dev);
struct nvm_dev *ndev = ns->ndev;
- struct nvm_id *id;
+ struct nvm_geo *dev_geo = &ndev->geo;
struct attribute *attr;
if (!ndev)
return 0;
- id = &ndev->identity;
attr = &dattr->attr;
if (strcmp(attr->name, "groups") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_ch);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_chnls);
} else if (strcmp(attr->name, "punits") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_lun);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_luns);
} else if (strcmp(attr->name, "chunks") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->num_chk);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.num_chk);
} else if (strcmp(attr->name, "clba") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->clba);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.clba);
} else if (strcmp(attr->name, "ws_min") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->ws_min);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.ws_min);
} else if (strcmp(attr->name, "ws_opt") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->ws_opt);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.ws_opt);
} else if (strcmp(attr->name, "mw_cunits") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->mw_cunits);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.mw_cunits);
} else if (strcmp(attr->name, "write_typ") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tprt);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tprt);
} else if (strcmp(attr->name, "write_max") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tprm);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tprm);
} else if (strcmp(attr->name, "reset_typ") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tbet);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tbet);
} else if (strcmp(attr->name, "reset_max") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", id->tbem);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.tbem);
} else {
- return scnprintf(page,
- PAGE_SIZE,
- "Unhandled attr(%s) in `nvm_dev_attr_show_20`\n",
- attr->name);
+ return scnprintf(page, PAGE_SIZE,
+ "Unhandled attr(%s) in `nvm_dev_attr_show_20`\n",
+ attr->name);
}
}
@@ -1106,10 +1166,13 @@ static const struct attribute_group nvm_dev_attr_group_20 = {
int nvme_nvm_register_sysfs(struct nvme_ns *ns)
{
- if (!ns->ndev)
+ struct nvm_dev *ndev = ns->ndev;
+ struct nvm_geo *dev_geo = &ndev->geo;
+
+ if (!ndev)
return -EINVAL;
- switch (ns->ndev->identity.ver_id) {
+ switch (dev_geo->ver_id) {
case 1:
return sysfs_create_group(&disk_to_dev(ns->disk)->kobj,
&nvm_dev_attr_group_12);
@@ -1123,7 +1186,10 @@ int nvme_nvm_register_sysfs(struct nvme_ns *ns)
void nvme_nvm_unregister_sysfs(struct nvme_ns *ns)
{
- switch (ns->ndev->identity.ver_id) {
+ struct nvm_dev *ndev = ns->ndev;
+ struct nvm_geo *dev_geo = &ndev->geo;
+
+ switch (dev_geo->ver_id) {
case 1:
sysfs_remove_group(&disk_to_dev(ns->disk)->kobj,
&nvm_dev_attr_group_12);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index e55b10573c99..4537c0140386 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -50,7 +50,7 @@ struct nvm_id;
struct nvm_dev;
struct nvm_tgt_dev;
-typedef int (nvm_id_fn)(struct nvm_dev *, struct nvm_id *);
+typedef int (nvm_id_fn)(struct nvm_dev *);
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 ppa_addr *, int, int);
typedef int (nvm_submit_io_fn)(struct nvm_dev *, struct nvm_rq *);
@@ -152,62 +152,48 @@ struct nvm_id_lp_tbl {
struct nvm_id_lp_mlc mlc;
};
-struct nvm_addr_format {
- u8 ch_offset;
+struct nvm_addr_format_12 {
u8 ch_len;
- u8 lun_offset;
u8 lun_len;
- u8 pln_offset;
+ u8 blk_len;
+ u8 pg_len;
u8 pln_len;
+ u8 sect_len;
+
+ u8 ch_offset;
+ u8 lun_offset;
u8 blk_offset;
- u8 blk_len;
u8 pg_offset;
- u8 pg_len;
+ u8 pln_offset;
u8 sect_offset;
- u8 sect_len;
-};
-
-struct nvm_id {
- u8 ver_id;
- u8 vmnt;
- u32 cap;
- u32 dom;
-
- struct nvm_addr_format ppaf;
-
- u8 num_ch;
- u8 num_lun;
- u16 num_chk;
- u16 clba;
- u16 csecs;
- u16 sos;
-
- u32 ws_min;
- u32 ws_opt;
- u32 mw_cunits;
- u32 trdt;
- u32 trdm;
- u32 tprt;
- u32 tprm;
- u32 tbet;
- u32 tbem;
- u32 mpos;
- u32 mccap;
- u16 cpar;
-
- /* calculated values */
- u16 ws_seq;
- u16 ws_per_chk;
-
- /* 1.2 compatibility */
- u8 mtype;
- u8 fmtype;
+ u64 ch_mask;
+ u64 lun_mask;
+ u64 blk_mask;
+ u64 pg_mask;
+ u64 pln_mask;
+ u64 sec_mask;
+};
- u8 num_pln;
- u16 num_pg;
- u16 fpg_sz;
-} __packed;
+struct nvm_addr_format {
+ u8 ch_len;
+ u8 lun_len;
+ u8 chk_len;
+ u8 sec_len;
+ u8 rsv_len[2];
+
+ u8 ch_offset;
+ u8 lun_offset;
+ u8 chk_offset;
+ u8 sec_offset;
+ u8 rsv_off[2];
+
+ u64 ch_mask;
+ u64 lun_mask;
+ u64 chk_mask;
+ u64 sec_mask;
+ u64 rsv_mask[2];
+};
struct nvm_target {
struct list_head list;
@@ -274,36 +260,69 @@ enum {
NVM_BLK_ST_BAD = 0x8, /* Bad block */
};
+/* Device common geometry */
+struct nvm_common_geo {
+ /* chunk geometry */
+ u32 num_chk; /* chunks per lun */
+ u32 clba; /* sectors per chunk */
+ u16 csecs; /* sector size */
+ u16 sos; /* out-of-band area size */
-/* Device generic information */
+ /* device write constrains */
+ u32 ws_min; /* minimum write size */
+ u32 ws_opt; /* optimal write size */
+ u32 mw_cunits; /* distance required for successful read */
+
+ /* device capabilities */
+ u32 mccap;
+
+ /* device timings */
+ u32 trdt; /* Avg. Tread (ns) */
+ u32 trdm; /* Max Tread (ns) */
+ u32 tprt; /* Avg. Tprog (ns) */
+ u32 tprm; /* Max Tprog (ns) */
+ u32 tbet; /* Avg. Terase (ns) */
+ u32 tbem; /* Max Terase (ns) */
+
+ /* generic address format */
+ struct nvm_addr_format addrf;
+
+ /* 1.2 compatibility */
+ u8 vmnt;
+ u32 cap;
+ u32 dom;
+
+ u8 mtype;
+ u8 fmtype;
+
+ u16 cpar;
+ u32 mpos;
+
+ u8 num_pln;
+ u8 pln_mode;
+ u16 num_pg;
+ u16 fpg_sz;
+};
+
+/* Instance geometry */
struct nvm_geo {
- /* generic geometry */
+ /* device reported version */
+ u8 ver_id;
+
+ /* instance specific geometry */
int nr_chnls;
- int all_luns; /* across channels */
- int nr_luns; /* per channel */
- int nr_chks; /* per lun */
+ int nr_luns; /* per channel */
- int sec_size;
- int oob_size;
- int mccap;
+ int op; /* over-provision in instance */
- int sec_per_chk;
- int sec_per_lun;
+ /* common geometry */
+ struct nvm_common_geo c;
- int ws_min;
- int ws_opt;
- int ws_seq;
- int ws_per_chk;
+ /* calculated values */
+ int all_luns; /* across channels */
+ int all_chunks; /* across channels */
- int op;
-
- struct nvm_addr_format ppaf;
-
- /* Legacy 1.2 specific geometry */
- int plane_mode; /* drive device in single, double or quad mode */
- int nr_planes;
- int sec_per_pg; /* only sectors for a single page */
- int sec_per_pl; /* all sectors across planes */
+ sector_t total_secs; /* across channels */
};
/* sub-device structure */
@@ -314,9 +333,6 @@ struct nvm_tgt_dev {
/* Base ppas for target LUNs */
struct ppa_addr *luns;
- sector_t total_secs;
-
- struct nvm_id identity;
struct request_queue *q;
struct nvm_dev *parent;
@@ -331,13 +347,9 @@ struct nvm_dev {
/* Device information */
struct nvm_geo geo;
- unsigned long total_secs;
-
unsigned long *lun_map;
void *dma_pool;
- struct nvm_id identity;
-
/* Backend device */
struct request_queue *q;
char name[DISK_NAME_LEN];
@@ -357,14 +369,16 @@ static inline struct ppa_addr generic_to_dev_addr(struct nvm_tgt_dev *tgt_dev,
struct ppa_addr r)
{
struct nvm_geo *geo = &tgt_dev->geo;
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&geo->c.addrf;
struct ppa_addr l;
- l.ppa = ((u64)r.g.blk) << geo->ppaf.blk_offset;
- l.ppa |= ((u64)r.g.pg) << geo->ppaf.pg_offset;
- l.ppa |= ((u64)r.g.sec) << geo->ppaf.sect_offset;
- l.ppa |= ((u64)r.g.pl) << geo->ppaf.pln_offset;
- l.ppa |= ((u64)r.g.lun) << geo->ppaf.lun_offset;
- l.ppa |= ((u64)r.g.ch) << geo->ppaf.ch_offset;
+ l.ppa = ((u64)r.g.ch) << ppaf->ch_offset;
+ l.ppa |= ((u64)r.g.lun) << ppaf->lun_offset;
+ l.ppa |= ((u64)r.g.blk) << ppaf->blk_offset;
+ l.ppa |= ((u64)r.g.pg) << ppaf->pg_offset;
+ l.ppa |= ((u64)r.g.pl) << ppaf->pln_offset;
+ l.ppa |= ((u64)r.g.sec) << ppaf->sect_offset;
return l;
}
@@ -373,24 +387,18 @@ static inline struct ppa_addr dev_to_generic_addr(struct nvm_tgt_dev *tgt_dev,
struct ppa_addr r)
{
struct nvm_geo *geo = &tgt_dev->geo;
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&geo->c.addrf;
struct ppa_addr l;
l.ppa = 0;
- /*
- * (r.ppa << X offset) & X len bitmask. X eq. blk, pg, etc.
- */
- l.g.blk = (r.ppa >> geo->ppaf.blk_offset) &
- (((1 << geo->ppaf.blk_len) - 1));
- l.g.pg |= (r.ppa >> geo->ppaf.pg_offset) &
- (((1 << geo->ppaf.pg_len) - 1));
- l.g.sec |= (r.ppa >> geo->ppaf.sect_offset) &
- (((1 << geo->ppaf.sect_len) - 1));
- l.g.pl |= (r.ppa >> geo->ppaf.pln_offset) &
- (((1 << geo->ppaf.pln_len) - 1));
- l.g.lun |= (r.ppa >> geo->ppaf.lun_offset) &
- (((1 << geo->ppaf.lun_len) - 1));
- l.g.ch |= (r.ppa >> geo->ppaf.ch_offset) &
- (((1 << geo->ppaf.ch_len) - 1));
+
+ l.g.ch = (r.ppa & ppaf->ch_mask) >> ppaf->ch_offset;
+ l.g.lun = (r.ppa & ppaf->lun_mask) >> ppaf->lun_offset;
+ l.g.blk = (r.ppa & ppaf->blk_mask) >> ppaf->blk_offset;
+ l.g.pg = (r.ppa & ppaf->pg_mask) >> ppaf->pg_offset;
+ l.g.pl = (r.ppa & ppaf->pln_mask) >> ppaf->pln_offset;
+ l.g.sec = (r.ppa & ppaf->sec_mask) >> ppaf->sect_offset;
return l;
}
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 02/19] lightnvm: add controller capabilities to 2.0
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
2018-02-26 13:17 ` [PATCH 01/19] lightnvm: simplify geometry structure Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 03/19] lightnvm: add minor version to generic geometry Javier González
` (17 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Assign missing mccap value on 2.0 path
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/nvme/host/lightnvm.c | 4 +++-
include/linux/lightnvm.h | 8 +++++---
2 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 8f541331b1a9..10392a664b50 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -318,7 +318,7 @@ static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
dev_geo->c.ws_opt = sec_per_pg;
dev_geo->c.mw_cunits = 8; /* default to MLC safe values */
- dev_geo->c.mccap = le32_to_cpu(src->mccap);
+ dev_geo->c.cap = le32_to_cpu(src->mccap);
dev_geo->c.trdt = le32_to_cpu(src->trdt);
dev_geo->c.trdm = le32_to_cpu(src->trdm);
@@ -399,6 +399,8 @@ static int nvme_nvm_setup_20(struct nvme_nvm_id20 *id,
dev_geo->c.ws_opt = le32_to_cpu(id->ws_opt);
dev_geo->c.mw_cunits = le32_to_cpu(id->mw_cunits);
+ dev_geo->c.cap = le32_to_cpu(id->mccap);
+
dev_geo->c.trdt = le32_to_cpu(id->trdt);
dev_geo->c.trdm = le32_to_cpu(id->trdm);
dev_geo->c.tprt = le32_to_cpu(id->twrt);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 4537c0140386..42704c4ca9af 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -273,8 +273,10 @@ struct nvm_common_geo {
u32 ws_opt; /* optimal write size */
u32 mw_cunits; /* distance required for successful read */
- /* device capabilities */
- u32 mccap;
+ /* device capabilities. Note that this represents capabilities in 1.2
+ * and media and controller capabilities in 2.0
+ */
+ u32 cap;
/* device timings */
u32 trdt; /* Avg. Tread (ns) */
@@ -289,7 +291,7 @@ struct nvm_common_geo {
/* 1.2 compatibility */
u8 vmnt;
- u32 cap;
+ u32 mccap;
u32 dom;
u8 mtype;
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 03/19] lightnvm: add minor version to generic geometry
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
2018-02-26 13:17 ` [PATCH 01/19] lightnvm: simplify geometry structure Javier González
2018-02-26 13:17 ` [PATCH 02/19] lightnvm: add controller capabilities to 2.0 Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 04/19] lightnvm: add shorten OCSSD version in geo Javier González
` (16 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Separate the version between major and minor on the generic geometry and
represent it through sysfs in the 2.0 path. The 1.2 path only shows the major
version to preserve the existing user space interface.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/core.c | 4 ++--
drivers/nvme/host/lightnvm.c | 25 ++++++++++++++++++++-----
include/linux/lightnvm.h | 3 ++-
3 files changed, 24 insertions(+), 8 deletions(-)
diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 43e3d6bb5be6..96f4e62d383b 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -897,8 +897,8 @@ static int nvm_init(struct nvm_dev *dev)
goto err;
}
- pr_debug("nvm: ver:%u nvm_vendor:%x\n",
- dev_geo->ver_id,
+ pr_debug("nvm: ver:%u.%u nvm_vendor:%x\n",
+ dev_geo->major_ver_id, dev_geo->minor_ver_id,
dev_geo->c.vmnt);
ret = nvm_core_init(dev);
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 10392a664b50..8befb60eeacb 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -295,7 +295,9 @@ static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
return -EINVAL;
}
- dev_geo->ver_id = id->ver_id;
+ /* 1.2 spec. only reports a single version id - unfold */
+ dev_geo->major_ver_id = id->ver_id;
+ dev_geo->minor_ver_id = 2;
dev_geo->nr_chnls = src->num_ch;
dev_geo->nr_luns = src->num_lun;
@@ -380,7 +382,14 @@ static void nvme_nvm_set_addr_20(struct nvm_addr_format *dst,
static int nvme_nvm_setup_20(struct nvme_nvm_id20 *id,
struct nvm_geo *dev_geo)
{
- dev_geo->ver_id = id->mjr;
+ dev_geo->major_ver_id = id->mjr;
+ dev_geo->minor_ver_id = id->mnr;
+
+ if (!(dev_geo->major_ver_id == 2 && dev_geo->minor_ver_id == 0)) {
+ pr_err("nvm: OCSSD version not supported (v%d.%d)\n",
+ dev_geo->major_ver_id, dev_geo->minor_ver_id);
+ return -EINVAL;
+ }
dev_geo->nr_chnls = le16_to_cpu(id->num_grp);
dev_geo->nr_luns = le16_to_cpu(id->num_pu);
@@ -920,7 +929,13 @@ static ssize_t nvm_dev_attr_show(struct device *dev,
attr = &dattr->attr;
if (strcmp(attr->name, "version") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->ver_id);
+ if (dev_geo->major_ver_id == 1)
+ return scnprintf(page, PAGE_SIZE, "%u\n",
+ dev_geo->major_ver_id);
+ else
+ return scnprintf(page, PAGE_SIZE, "%u.%u\n",
+ dev_geo->major_ver_id,
+ dev_geo->minor_ver_id);
} else if (strcmp(attr->name, "capabilities") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.cap);
} else if (strcmp(attr->name, "read_typ") == 0) {
@@ -1174,7 +1189,7 @@ int nvme_nvm_register_sysfs(struct nvme_ns *ns)
if (!ndev)
return -EINVAL;
- switch (dev_geo->ver_id) {
+ switch (dev_geo->major_ver_id) {
case 1:
return sysfs_create_group(&disk_to_dev(ns->disk)->kobj,
&nvm_dev_attr_group_12);
@@ -1191,7 +1206,7 @@ void nvme_nvm_unregister_sysfs(struct nvme_ns *ns)
struct nvm_dev *ndev = ns->ndev;
struct nvm_geo *dev_geo = &ndev->geo;
- switch (dev_geo->ver_id) {
+ switch (dev_geo->major_ver_id) {
case 1:
sysfs_remove_group(&disk_to_dev(ns->disk)->kobj,
&nvm_dev_attr_group_12);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 42704c4ca9af..ccaf8a30202d 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -309,7 +309,8 @@ struct nvm_common_geo {
/* Instance geometry */
struct nvm_geo {
/* device reported version */
- u8 ver_id;
+ u8 major_ver_id;
+ u8 minor_ver_id;
/* instance specific geometry */
int nr_chnls;
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 04/19] lightnvm: add shorten OCSSD version in geo
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (2 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 03/19] lightnvm: add minor version to generic geometry Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 05/19] lightnvm: complete geo structure with maxoc* Javier González
` (15 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Create a shorten version to use in the generic geometry.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/nvme/host/lightnvm.c | 6 ++++++
include/linux/lightnvm.h | 8 ++++++++
2 files changed, 14 insertions(+)
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 8befb60eeacb..a48950a37b8d 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -299,6 +299,9 @@ static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
dev_geo->major_ver_id = id->ver_id;
dev_geo->minor_ver_id = 2;
+ /* Set compacted version for upper layers */
+ dev_geo->c.version = NVM_OCSSD_SPEC_12;
+
dev_geo->nr_chnls = src->num_ch;
dev_geo->nr_luns = src->num_lun;
dev_geo->all_luns = dev_geo->nr_chnls * dev_geo->nr_luns;
@@ -385,6 +388,9 @@ static int nvme_nvm_setup_20(struct nvme_nvm_id20 *id,
dev_geo->major_ver_id = id->mjr;
dev_geo->minor_ver_id = id->mnr;
+ /* Set compacted version for upper layers */
+ dev_geo->c.version = NVM_OCSSD_SPEC_20;
+
if (!(dev_geo->major_ver_id == 2 && dev_geo->minor_ver_id == 0)) {
pr_err("nvm: OCSSD version not supported (v%d.%d)\n",
dev_geo->major_ver_id, dev_geo->minor_ver_id);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index ccaf8a30202d..57d00644a1c6 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -23,6 +23,11 @@ enum {
#define NVM_LUN_BITS (8)
#define NVM_CH_BITS (7)
+enum {
+ NVM_OCSSD_SPEC_12 = 12,
+ NVM_OCSSD_SPEC_20 = 20,
+};
+
struct ppa_addr {
/* Generic structure for all addresses */
union {
@@ -262,6 +267,9 @@ enum {
/* Device common geometry */
struct nvm_common_geo {
+ /* kernel short version */
+ u8 version;
+
/* chunk geometry */
u32 num_chk; /* chunks per lun */
u32 clba; /* sectors per chunk */
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 05/19] lightnvm: complete geo structure with maxoc*
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (3 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 04/19] lightnvm: add shorten OCSSD version in geo Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 06/19] lightnvm: pblk: check for supported version Javier González
` (14 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Complete the generic geometry structure with the maxoc and maxocpu
felds, present in the 2.0 spec.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/nvme/host/lightnvm.c | 9 +++++++++
include/linux/lightnvm.h | 2 ++
2 files changed, 11 insertions(+)
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index a48950a37b8d..a8f35e8b7953 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -323,6 +323,13 @@ static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
dev_geo->c.ws_opt = sec_per_pg;
dev_geo->c.mw_cunits = 8; /* default to MLC safe values */
+ /* Do not impose values for maximum number of open blocks as it is
+ * unspecified in 1.2. Users of 1.2 must be aware of this and eventually
+ * specify these values through a quirk if restrictions apply.
+ */
+ dev_geo->c.maxoc = dev_geo->all_luns * dev_geo->c.num_chk;
+ dev_geo->c.maxocpu = dev_geo->c.num_chk;
+
dev_geo->c.cap = le32_to_cpu(src->mccap);
dev_geo->c.trdt = le32_to_cpu(src->trdt);
@@ -413,6 +420,8 @@ static int nvme_nvm_setup_20(struct nvme_nvm_id20 *id,
dev_geo->c.ws_min = le32_to_cpu(id->ws_min);
dev_geo->c.ws_opt = le32_to_cpu(id->ws_opt);
dev_geo->c.mw_cunits = le32_to_cpu(id->mw_cunits);
+ dev_geo->c.maxoc = le32_to_cpu(id->maxoc);
+ dev_geo->c.maxocpu = le32_to_cpu(id->maxocpu);
dev_geo->c.cap = le32_to_cpu(id->mccap);
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 57d00644a1c6..e8e6bfd4d0d4 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -280,6 +280,8 @@ struct nvm_common_geo {
u32 ws_min; /* minimum write size */
u32 ws_opt; /* optimal write size */
u32 mw_cunits; /* distance required for successful read */
+ u32 maxoc; /* maximum open chunks */
+ u32 maxocpu; /* maximum open chunks per parallel unit */
/* device capabilities. Note that this represents capabilities in 1.2
* and media and controller capabilities in 2.0
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 06/19] lightnvm: pblk: check for supported version
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (4 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 05/19] lightnvm: complete geo structure with maxoc* Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 07/19] lightnvm: complete 2.0 values in sysfs Javier González
` (13 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
At this point, only 1.2 spec is supported, thus check for it. Also,
since device-side L2P is only supported in the 1.2 spec, make sure to
only check its value under 1.2.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-init.c | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index 95ecb0ec736b..6e0567971cf6 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -990,9 +990,15 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
struct pblk *pblk;
int ret;
- if (dev->geo.c.dom & NVM_RSP_L2P) {
+ if (geo->c.version != NVM_OCSSD_SPEC_12) {
+ pr_err("pblk: OCSSD version not supported (%u)\n",
+ geo->c.version);
+ return ERR_PTR(-EINVAL);
+ }
+
+ if (geo->c.version == NVM_OCSSD_SPEC_12 && geo->c.dom & NVM_RSP_L2P) {
pr_err("pblk: host-side L2P table not supported. (%x)\n",
- dev->geo.c.dom);
+ geo->c.dom);
return ERR_PTR(-EINVAL);
}
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 07/19] lightnvm: complete 2.0 values in sysfs
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (5 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 06/19] lightnvm: pblk: check for supported version Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 08/19] lightnvm: rename number of channels and luns Javier González
` (12 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Add missing geometry values to sysfs. Namely, maxoc and maxocpu.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/nvme/host/lightnvm.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index a8f35e8b7953..4ea553c91756 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -1066,6 +1066,10 @@ static ssize_t nvm_dev_attr_show_20(struct device *dev,
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.ws_min);
} else if (strcmp(attr->name, "ws_opt") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.ws_opt);
+ } else if (strcmp(attr->name, "maxoc") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.maxoc);
+ } else if (strcmp(attr->name, "maxocpu") == 0) {
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.maxocpu);
} else if (strcmp(attr->name, "mw_cunits") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.mw_cunits);
} else if (strcmp(attr->name, "write_typ") == 0) {
@@ -1163,6 +1167,8 @@ static NVM_DEV_ATTR_20_RO(chunks);
static NVM_DEV_ATTR_20_RO(clba);
static NVM_DEV_ATTR_20_RO(ws_min);
static NVM_DEV_ATTR_20_RO(ws_opt);
+static NVM_DEV_ATTR_20_RO(maxoc);
+static NVM_DEV_ATTR_20_RO(maxocpu);
static NVM_DEV_ATTR_20_RO(mw_cunits);
static NVM_DEV_ATTR_20_RO(write_typ);
static NVM_DEV_ATTR_20_RO(write_max);
@@ -1179,6 +1185,8 @@ static struct attribute *nvm_dev_attrs_20[] = {
&dev_attr_clba.attr,
&dev_attr_ws_min.attr,
&dev_attr_ws_opt.attr,
+ &dev_attr_maxoc.attr,
+ &dev_attr_maxocpu.attr,
&dev_attr_mw_cunits.attr,
&dev_attr_read_typ.attr,
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 08/19] lightnvm: rename number of channels and luns
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (6 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 07/19] lightnvm: complete 2.0 values in sysfs Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 09/19] lightnvm: rename sect_* to sec_* Javier González
` (11 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Normalize nomenclature for naming number of channels and number of luns
in order to improve readability. Use num_ch and num_lun.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/core.c | 69 ++++++++++++++++++++++----------------------
drivers/lightnvm/pblk-core.c | 4 +--
drivers/lightnvm/pblk-init.c | 16 +++++-----
drivers/lightnvm/pblk.h | 6 ++--
drivers/nvme/host/lightnvm.c | 20 ++++++-------
include/linux/lightnvm.h | 4 +--
6 files changed, 59 insertions(+), 60 deletions(-)
diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 96f4e62d383b..45bed54488f9 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -36,13 +36,13 @@ static DECLARE_RWSEM(nvm_lock);
/* Map between virtual and physical channel and lun */
struct nvm_ch_map {
int ch_off;
- int nr_luns;
+ int num_lun;
int *lun_offs;
};
struct nvm_dev_map {
struct nvm_ch_map *chnls;
- int nr_chnls;
+ int num_ch;
};
static struct nvm_target *nvm_find_target(struct nvm_dev *dev, const char *name)
@@ -115,15 +115,15 @@ static void nvm_remove_tgt_dev(struct nvm_tgt_dev *tgt_dev, int clear)
struct nvm_dev_map *dev_map = tgt_dev->map;
int i, j;
- for (i = 0; i < dev_map->nr_chnls; i++) {
+ for (i = 0; i < dev_map->num_ch; i++) {
struct nvm_ch_map *ch_map = &dev_map->chnls[i];
int *lun_offs = ch_map->lun_offs;
int ch = i + ch_map->ch_off;
if (clear) {
- for (j = 0; j < ch_map->nr_luns; j++) {
+ for (j = 0; j < ch_map->num_lun; j++) {
int lun = j + lun_offs[j];
- int lunid = (ch * dev_geo->nr_luns) + lun;
+ int lunid = (ch * dev_geo->num_lun) + lun;
WARN_ON(!test_and_clear_bit(lunid,
dev->lun_map));
@@ -149,47 +149,46 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
struct nvm_dev_map *dev_rmap = dev->rmap;
struct nvm_dev_map *dev_map;
struct ppa_addr *luns;
- int nr_luns = lun_end - lun_begin + 1;
- int luns_left = nr_luns;
- int nr_chnls = nr_luns / dev_geo->nr_luns;
- int nr_chnls_mod = nr_luns % dev_geo->nr_luns;
- int bch = lun_begin / dev_geo->nr_luns;
- int blun = lun_begin % dev_geo->nr_luns;
+ int num_lun = lun_end - lun_begin + 1;
+ int luns_left = num_lun;
+ int num_ch = num_lun / dev_geo->num_lun;
+ int num_ch_mod = num_lun % dev_geo->num_lun;
+ int bch = lun_begin / dev_geo->num_lun;
+ int blun = lun_begin % dev_geo->num_lun;
int lunid = 0;
int lun_balanced = 1;
- int sec_per_lun, prev_nr_luns;
+ int sec_per_lun, prev_num_lun;
int i, j;
- nr_chnls = (nr_chnls_mod == 0) ? nr_chnls : nr_chnls + 1;
+ num_ch = (num_ch_mod == 0) ? num_ch : num_ch + 1;
dev_map = kmalloc(sizeof(struct nvm_dev_map), GFP_KERNEL);
if (!dev_map)
goto err_dev;
- dev_map->chnls = kcalloc(nr_chnls, sizeof(struct nvm_ch_map),
- GFP_KERNEL);
+ dev_map->chnls = kcalloc(num_ch, sizeof(struct nvm_ch_map), GFP_KERNEL);
if (!dev_map->chnls)
goto err_chnls;
- luns = kcalloc(nr_luns, sizeof(struct ppa_addr), GFP_KERNEL);
+ luns = kcalloc(num_lun, sizeof(struct ppa_addr), GFP_KERNEL);
if (!luns)
goto err_luns;
- prev_nr_luns = (luns_left > dev_geo->nr_luns) ?
- dev_geo->nr_luns : luns_left;
- for (i = 0; i < nr_chnls; i++) {
+ prev_num_lun = (luns_left > dev_geo->num_lun) ?
+ dev_geo->num_lun : luns_left;
+ for (i = 0; i < num_ch; i++) {
struct nvm_ch_map *ch_rmap = &dev_rmap->chnls[i + bch];
int *lun_roffs = ch_rmap->lun_offs;
struct nvm_ch_map *ch_map = &dev_map->chnls[i];
int *lun_offs;
- int luns_in_chnl = (luns_left > dev_geo->nr_luns) ?
- dev_geo->nr_luns : luns_left;
+ int luns_in_chnl = (luns_left > dev_geo->num_lun) ?
+ dev_geo->num_lun : luns_left;
- if (lun_balanced && prev_nr_luns != luns_in_chnl)
+ if (lun_balanced && prev_num_lun != luns_in_chnl)
lun_balanced = 0;
ch_map->ch_off = ch_rmap->ch_off = bch;
- ch_map->nr_luns = luns_in_chnl;
+ ch_map->num_lun = luns_in_chnl;
lun_offs = kcalloc(luns_in_chnl, sizeof(int), GFP_KERNEL);
if (!lun_offs)
@@ -211,22 +210,22 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
luns_left -= luns_in_chnl;
}
- dev_map->nr_chnls = nr_chnls;
+ dev_map->num_ch = num_ch;
tgt_dev = kmalloc(sizeof(struct nvm_tgt_dev), GFP_KERNEL);
if (!tgt_dev)
goto err_ch;
/* Target device only owns a portion of the physical device */
- tgt_dev->geo.nr_chnls = nr_chnls;
- tgt_dev->geo.nr_luns = (lun_balanced) ? prev_nr_luns : -1;
- tgt_dev->geo.all_luns = nr_luns;
- tgt_dev->geo.all_chunks = nr_luns * dev_geo->c.num_chk;
+ tgt_dev->geo.num_ch = num_ch;
+ tgt_dev->geo.num_lun = (lun_balanced) ? prev_num_lun : -1;
+ tgt_dev->geo.all_luns = num_lun;
+ tgt_dev->geo.all_chunks = num_lun * dev_geo->c.num_chk;
tgt_dev->geo.op = op;
sec_per_lun = dev_geo->c.clba * dev_geo->c.num_chk;
- tgt_dev->geo.total_secs = nr_luns * sec_per_lun;
+ tgt_dev->geo.total_secs = num_lun * sec_per_lun;
tgt_dev->geo.c = dev_geo->c;
@@ -511,20 +510,20 @@ static int nvm_register_map(struct nvm_dev *dev)
if (!rmap)
goto err_rmap;
- rmap->chnls = kcalloc(dev_geo->nr_chnls, sizeof(struct nvm_ch_map),
+ rmap->chnls = kcalloc(dev_geo->num_ch, sizeof(struct nvm_ch_map),
GFP_KERNEL);
if (!rmap->chnls)
goto err_chnls;
- for (i = 0; i < dev_geo->nr_chnls; i++) {
+ for (i = 0; i < dev_geo->num_ch; i++) {
struct nvm_ch_map *ch_rmap;
int *lun_roffs;
- int luns_in_chnl = dev_geo->nr_luns;
+ int luns_in_chnl = dev_geo->num_lun;
ch_rmap = &rmap->chnls[i];
ch_rmap->ch_off = -1;
- ch_rmap->nr_luns = luns_in_chnl;
+ ch_rmap->num_lun = luns_in_chnl;
lun_roffs = kcalloc(luns_in_chnl, sizeof(int), GFP_KERNEL);
if (!lun_roffs)
@@ -554,7 +553,7 @@ static void nvm_unregister_map(struct nvm_dev *dev)
struct nvm_dev_map *rmap = dev->rmap;
int i;
- for (i = 0; i < dev_geo->nr_chnls; i++)
+ for (i = 0; i < dev_geo->num_ch; i++)
kfree(rmap->chnls[i].lun_offs);
kfree(rmap->chnls);
@@ -910,7 +909,7 @@ static int nvm_init(struct nvm_dev *dev)
pr_info("nvm: registered %s [%u/%u/%u/%u/%u]\n",
dev->name, dev_geo->c.ws_min, dev_geo->c.ws_opt,
dev_geo->c.num_chk, dev_geo->all_luns,
- dev_geo->nr_chnls);
+ dev_geo->num_ch);
return 0;
err:
pr_err("nvm: failed to initialize nvm\n");
diff --git a/drivers/lightnvm/pblk-core.c b/drivers/lightnvm/pblk-core.c
index 519af8b9eab7..7c726003a8d2 100644
--- a/drivers/lightnvm/pblk-core.c
+++ b/drivers/lightnvm/pblk-core.c
@@ -1752,10 +1752,10 @@ void pblk_up_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
struct pblk_lun *rlun;
- int nr_luns = geo->all_luns;
+ int num_lun = geo->all_luns;
int bit = -1;
- while ((bit = find_next_bit(lun_bitmap, nr_luns, bit + 1)) < nr_luns) {
+ while ((bit = find_next_bit(lun_bitmap, num_lun, bit + 1)) < num_lun) {
rlun = &pblk->luns[bit];
up(&rlun->wr_sem);
}
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index 6e0567971cf6..7e3297be517d 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -162,15 +162,15 @@ static int pblk_set_addrf_12(struct nvm_geo *geo,
int power_len;
/* Re-calculate channel and lun format to adapt to configuration */
- power_len = get_count_order(geo->nr_chnls);
- if (1 << power_len != geo->nr_chnls) {
+ power_len = get_count_order(geo->num_ch);
+ if (1 << power_len != geo->num_ch) {
pr_err("pblk: supports only power-of-two channel config.\n");
return -EINVAL;
}
dst->ch_len = power_len;
- power_len = get_count_order(geo->nr_luns);
- if (1 << power_len != geo->nr_luns) {
+ power_len = get_count_order(geo->num_lun);
+ if (1 << power_len != geo->num_lun) {
pr_err("pblk: supports only power-of-two LUN config.\n");
return -EINVAL;
}
@@ -498,7 +498,7 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
int i;
/* TODO: Implement unbalanced LUN support */
- if (geo->nr_luns < 0) {
+ if (geo->num_lun < 0) {
pr_err("pblk: unbalanced LUN config.\n");
return -EINVAL;
}
@@ -510,9 +510,9 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
for (i = 0; i < geo->all_luns; i++) {
/* Stripe across channels */
- int ch = i % geo->nr_chnls;
- int lun_raw = i / geo->nr_chnls;
- int lunid = lun_raw + ch * geo->nr_luns;
+ int ch = i % geo->num_ch;
+ int lun_raw = i / geo->num_ch;
+ int lunid = lun_raw + ch * geo->num_lun;
rlun = &pblk->luns[i];
rlun->bppa = luns[lunid];
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index 67ffb53608f7..1f32284b0aec 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -941,7 +941,7 @@ static inline int pblk_ppa_to_line(struct ppa_addr p)
static inline int pblk_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
{
- return p.g.lun * geo->nr_chnls + p.g.ch;
+ return p.g.lun * geo->num_ch + p.g.ch;
}
static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
@@ -1214,8 +1214,8 @@ static inline int pblk_boundary_ppa_checks(struct nvm_tgt_dev *tgt_dev,
ppa = &ppas[i];
if (!ppa->c.is_cached &&
- ppa->g.ch < geo->nr_chnls &&
- ppa->g.lun < geo->nr_luns &&
+ ppa->g.ch < geo->num_ch &&
+ ppa->g.lun < geo->num_lun &&
ppa->g.pl < geo->c.num_pln &&
ppa->g.blk < geo->c.num_chk &&
ppa->g.pg < geo->c.num_pg &&
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 4ea553c91756..f3f524dc1389 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -302,9 +302,9 @@ static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
/* Set compacted version for upper layers */
dev_geo->c.version = NVM_OCSSD_SPEC_12;
- dev_geo->nr_chnls = src->num_ch;
- dev_geo->nr_luns = src->num_lun;
- dev_geo->all_luns = dev_geo->nr_chnls * dev_geo->nr_luns;
+ dev_geo->num_ch = src->num_ch;
+ dev_geo->num_lun = src->num_lun;
+ dev_geo->all_luns = dev_geo->num_ch * dev_geo->num_lun;
dev_geo->c.num_chk = le16_to_cpu(src->num_chk);
@@ -404,9 +404,9 @@ static int nvme_nvm_setup_20(struct nvme_nvm_id20 *id,
return -EINVAL;
}
- dev_geo->nr_chnls = le16_to_cpu(id->num_grp);
- dev_geo->nr_luns = le16_to_cpu(id->num_pu);
- dev_geo->all_luns = dev_geo->nr_chnls * dev_geo->nr_luns;
+ dev_geo->num_ch = le16_to_cpu(id->num_grp);
+ dev_geo->num_lun = le16_to_cpu(id->num_pu);
+ dev_geo->all_luns = dev_geo->num_ch * dev_geo->num_lun;
dev_geo->c.num_chk = le32_to_cpu(id->num_chk);
dev_geo->c.clba = le32_to_cpu(id->clba);
@@ -1005,9 +1005,9 @@ static ssize_t nvm_dev_attr_show_12(struct device *dev,
} else if (strcmp(attr->name, "flash_media_type") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.fmtype);
} else if (strcmp(attr->name, "num_channels") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_chnls);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->num_ch);
} else if (strcmp(attr->name, "num_luns") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_luns);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->num_lun);
} else if (strcmp(attr->name, "num_planes") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.num_pln);
} else if (strcmp(attr->name, "num_blocks") == 0) { /* u16 */
@@ -1055,9 +1055,9 @@ static ssize_t nvm_dev_attr_show_20(struct device *dev,
attr = &dattr->attr;
if (strcmp(attr->name, "groups") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_chnls);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->num_ch);
} else if (strcmp(attr->name, "punits") == 0) {
- return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->nr_luns);
+ return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->num_lun);
} else if (strcmp(attr->name, "chunks") == 0) {
return scnprintf(page, PAGE_SIZE, "%u\n", dev_geo->c.num_chk);
} else if (strcmp(attr->name, "clba") == 0) {
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index e8e6bfd4d0d4..797594d6519f 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -323,8 +323,8 @@ struct nvm_geo {
u8 minor_ver_id;
/* instance specific geometry */
- int nr_chnls;
- int nr_luns; /* per channel */
+ int num_ch;
+ int num_lun; /* per channel */
int op; /* over-provision in instance */
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 09/19] lightnvm: rename sect_* to sec_*
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (7 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 08/19] lightnvm: rename number of channels and luns Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 10/19] lightnvm: add support for 2.0 address format Javier González
` (10 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Rename abbreviations for sector from sect_* to sec_* as most of the code
uses this format and it is confusing when using the different
structures.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-init.c | 8 ++++----
drivers/lightnvm/pblk-sysfs.c | 4 ++--
drivers/lightnvm/pblk.h | 8 ++++----
drivers/nvme/host/lightnvm.c | 8 ++++----
include/linux/lightnvm.h | 8 ++++----
5 files changed, 18 insertions(+), 18 deletions(-)
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index 7e3297be517d..30aec863f136 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -179,16 +179,16 @@ static int pblk_set_addrf_12(struct nvm_geo *geo,
dst->blk_len = src->blk_len;
dst->pg_len = src->pg_len;
dst->pln_len = src->pln_len;
- dst->sect_len = src->sect_len;
+ dst->sec_len = src->sec_len;
- dst->sect_offset = 0;
- dst->pln_offset = dst->sect_len;
+ dst->sec_offset = 0;
+ dst->pln_offset = dst->sec_len;
dst->ch_offset = dst->pln_offset + dst->pln_len;
dst->lun_offset = dst->ch_offset + dst->ch_len;
dst->pg_offset = dst->lun_offset + dst->lun_len;
dst->blk_offset = dst->pg_offset + dst->pg_len;
- dst->sec_mask = ((1ULL << dst->sect_len) - 1) << dst->sect_offset;
+ dst->sec_mask = ((1ULL << dst->sec_len) - 1) << dst->sec_offset;
dst->pln_mask = ((1ULL << dst->pln_len) - 1) << dst->pln_offset;
dst->ch_mask = ((1ULL << dst->ch_len) - 1) << dst->ch_offset;
dst->lun_mask = ((1ULL << dst->lun_len) - 1) << dst->lun_offset;
diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
index 5eb21a279361..81288aa9162a 100644
--- a/drivers/lightnvm/pblk-sysfs.c
+++ b/drivers/lightnvm/pblk-sysfs.c
@@ -128,7 +128,7 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
ppaf->blk_offset, ppaf->blk_len,
ppaf->pg_offset, ppaf->pg_len,
ppaf->pln_offset, ppaf->pln_len,
- ppaf->sect_offset, ppaf->sect_len);
+ ppaf->sec_offset, ppaf->sec_len);
sz += snprintf(page + sz, PAGE_SIZE - sz,
"device:ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
@@ -137,7 +137,7 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
geo_ppaf->blk_offset, geo_ppaf->blk_len,
geo_ppaf->pg_offset, geo_ppaf->pg_len,
geo_ppaf->pln_offset, geo_ppaf->pln_len,
- geo_ppaf->sect_offset, geo_ppaf->sect_len);
+ geo_ppaf->sec_offset, geo_ppaf->sec_len);
return sz;
}
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index 1f32284b0aec..4f7a365436f1 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -957,7 +957,7 @@ static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
ppa.g.lun = (paddr & ppaf->lun_mask) >> ppaf->lun_offset;
ppa.g.ch = (paddr & ppaf->ch_mask) >> ppaf->ch_offset;
ppa.g.pl = (paddr & ppaf->pln_mask) >> ppaf->pln_offset;
- ppa.g.sec = (paddr & ppaf->sec_mask) >> ppaf->sect_offset;
+ ppa.g.sec = (paddr & ppaf->sec_mask) >> ppaf->sec_offset;
return ppa;
}
@@ -973,7 +973,7 @@ static inline u64 pblk_dev_ppa_to_line_addr(struct pblk *pblk,
paddr |= (u64)p.g.lun << ppaf->lun_offset;
paddr |= (u64)p.g.pg << ppaf->pg_offset;
paddr |= (u64)p.g.pl << ppaf->pln_offset;
- paddr |= (u64)p.g.sec << ppaf->sect_offset;
+ paddr |= (u64)p.g.sec << ppaf->sec_offset;
return paddr;
}
@@ -998,7 +998,7 @@ static inline struct ppa_addr pblk_ppa32_to_ppa64(struct pblk *pblk, u32 ppa32)
ppa64.g.blk = (ppa32 & ppaf->blk_mask) >> ppaf->blk_offset;
ppa64.g.pg = (ppa32 & ppaf->pg_mask) >> ppaf->pg_offset;
ppa64.g.pl = (ppa32 & ppaf->pln_mask) >> ppaf->pln_offset;
- ppa64.g.sec = (ppa32 & ppaf->sec_mask) >> ppaf->sect_offset;
+ ppa64.g.sec = (ppa32 & ppaf->sec_mask) >> ppaf->sec_offset;
}
return ppa64;
@@ -1022,7 +1022,7 @@ static inline u32 pblk_ppa64_to_ppa32(struct pblk *pblk, struct ppa_addr ppa64)
ppa32 |= ppa64.g.blk << ppaf->blk_offset;
ppa32 |= ppa64.g.pg << ppaf->pg_offset;
ppa32 |= ppa64.g.pl << ppaf->pln_offset;
- ppa32 |= ppa64.g.sec << ppaf->sect_offset;
+ ppa32 |= ppa64.g.sec << ppaf->sec_offset;
}
return ppa32;
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index f3f524dc1389..377e6525e53f 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -262,21 +262,21 @@ static void nvme_nvm_set_addr_12(struct nvm_addr_format_12 *dst,
dst->blk_len = src->blk_len;
dst->pg_len = src->pg_len;
dst->pln_len = src->pln_len;
- dst->sect_len = src->sec_len;
+ dst->sec_len = src->sec_len;
dst->ch_offset = src->ch_offset;
dst->lun_offset = src->lun_offset;
dst->blk_offset = src->blk_offset;
dst->pg_offset = src->pg_offset;
dst->pln_offset = src->pln_offset;
- dst->sect_offset = src->sec_offset;
+ dst->sec_offset = src->sec_offset;
dst->ch_mask = ((1ULL << dst->ch_len) - 1) << dst->ch_offset;
dst->lun_mask = ((1ULL << dst->lun_len) - 1) << dst->lun_offset;
dst->blk_mask = ((1ULL << dst->blk_len) - 1) << dst->blk_offset;
dst->pg_mask = ((1ULL << dst->pg_len) - 1) << dst->pg_offset;
dst->pln_mask = ((1ULL << dst->pln_len) - 1) << dst->pln_offset;
- dst->sec_mask = ((1ULL << dst->sect_len) - 1) << dst->sect_offset;
+ dst->sec_mask = ((1ULL << dst->sec_len) - 1) << dst->sec_offset;
}
static int nvme_nvm_setup_12(struct nvme_nvm_id12 *id,
@@ -975,7 +975,7 @@ static ssize_t nvm_dev_attr_show_ppaf(struct nvm_addr_format_12 *ppaf,
ppaf->pln_offset, ppaf->pln_len,
ppaf->blk_offset, ppaf->blk_len,
ppaf->pg_offset, ppaf->pg_len,
- ppaf->sect_offset, ppaf->sect_len);
+ ppaf->sec_offset, ppaf->sec_len);
}
static ssize_t nvm_dev_attr_show_12(struct device *dev,
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 797594d6519f..18bb9e9774d6 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -163,14 +163,14 @@ struct nvm_addr_format_12 {
u8 blk_len;
u8 pg_len;
u8 pln_len;
- u8 sect_len;
+ u8 sec_len;
u8 ch_offset;
u8 lun_offset;
u8 blk_offset;
u8 pg_offset;
u8 pln_offset;
- u8 sect_offset;
+ u8 sec_offset;
u64 ch_mask;
u64 lun_mask;
@@ -391,7 +391,7 @@ static inline struct ppa_addr generic_to_dev_addr(struct nvm_tgt_dev *tgt_dev,
l.ppa |= ((u64)r.g.blk) << ppaf->blk_offset;
l.ppa |= ((u64)r.g.pg) << ppaf->pg_offset;
l.ppa |= ((u64)r.g.pl) << ppaf->pln_offset;
- l.ppa |= ((u64)r.g.sec) << ppaf->sect_offset;
+ l.ppa |= ((u64)r.g.sec) << ppaf->sec_offset;
return l;
}
@@ -411,7 +411,7 @@ static inline struct ppa_addr dev_to_generic_addr(struct nvm_tgt_dev *tgt_dev,
l.g.blk = (r.ppa & ppaf->blk_mask) >> ppaf->blk_offset;
l.g.pg = (r.ppa & ppaf->pg_mask) >> ppaf->pg_offset;
l.g.pl = (r.ppa & ppaf->pln_mask) >> ppaf->pln_offset;
- l.g.sec = (r.ppa & ppaf->sec_mask) >> ppaf->sect_offset;
+ l.g.sec = (r.ppa & ppaf->sec_mask) >> ppaf->sec_offset;
return l;
}
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 10/19] lightnvm: add support for 2.0 address format
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (8 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 09/19] lightnvm: rename sect_* to sec_* Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 11/19] lightnvm: pblk: rename ppaf* to addrf* Javier González
` (9 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Add support for 2.0 address format. Also, align address bits for 1.2 and
2.0 to be able to operate on channel and luns without requiring a format
conversion. Use a generic address format for this purpose.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/core.c | 20 ++++-----
include/linux/lightnvm.h | 105 ++++++++++++++++++++++++++++++++++-------------
2 files changed, 86 insertions(+), 39 deletions(-)
diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 45bed54488f9..92ea6e39dd64 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -196,8 +196,8 @@ static struct nvm_tgt_dev *nvm_create_tgt_dev(struct nvm_dev *dev,
for (j = 0; j < luns_in_chnl; j++) {
luns[lunid].ppa = 0;
- luns[lunid].g.ch = i;
- luns[lunid++].g.lun = j;
+ luns[lunid].a.ch = i;
+ luns[lunid++].a.lun = j;
lun_offs[j] = blun;
lun_roffs[j + blun] = blun;
@@ -563,22 +563,22 @@ static void nvm_unregister_map(struct nvm_dev *dev)
static void nvm_map_to_dev(struct nvm_tgt_dev *tgt_dev, struct ppa_addr *p)
{
struct nvm_dev_map *dev_map = tgt_dev->map;
- struct nvm_ch_map *ch_map = &dev_map->chnls[p->g.ch];
- int lun_off = ch_map->lun_offs[p->g.lun];
+ struct nvm_ch_map *ch_map = &dev_map->chnls[p->a.ch];
+ int lun_off = ch_map->lun_offs[p->a.lun];
- p->g.ch += ch_map->ch_off;
- p->g.lun += lun_off;
+ p->a.ch += ch_map->ch_off;
+ p->a.lun += lun_off;
}
static void nvm_map_to_tgt(struct nvm_tgt_dev *tgt_dev, struct ppa_addr *p)
{
struct nvm_dev *dev = tgt_dev->parent;
struct nvm_dev_map *dev_rmap = dev->rmap;
- struct nvm_ch_map *ch_rmap = &dev_rmap->chnls[p->g.ch];
- int lun_roff = ch_rmap->lun_offs[p->g.lun];
+ struct nvm_ch_map *ch_rmap = &dev_rmap->chnls[p->a.ch];
+ int lun_roff = ch_rmap->lun_offs[p->a.lun];
- p->g.ch -= ch_rmap->ch_off;
- p->g.lun -= lun_roff;
+ p->a.ch -= ch_rmap->ch_off;
+ p->a.lun -= lun_roff;
}
static void nvm_ppa_tgt_to_dev(struct nvm_tgt_dev *tgt_dev,
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 18bb9e9774d6..8cda1125c34a 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -16,12 +16,21 @@ enum {
NVM_IOTYPE_GC = 1,
};
-#define NVM_BLK_BITS (16)
-#define NVM_PG_BITS (16)
-#define NVM_SEC_BITS (8)
-#define NVM_PL_BITS (8)
-#define NVM_LUN_BITS (8)
-#define NVM_CH_BITS (7)
+/* common format */
+#define NVM_GEN_CH_BITS (8)
+#define NVM_GEN_LUN_BITS (8)
+#define NVM_GEN_BLK_BITS (16)
+#define NVM_GEN_RESERVED (32)
+
+/* 1.2 format */
+#define NVM_12_PG_BITS (16)
+#define NVM_12_PL_BITS (4)
+#define NVM_12_SEC_BITS (4)
+#define NVM_12_RESERVED (8)
+
+/* 2.0 format */
+#define NVM_20_SEC_BITS (24)
+#define NVM_20_RESERVED (8)
enum {
NVM_OCSSD_SPEC_12 = 12,
@@ -31,16 +40,34 @@ enum {
struct ppa_addr {
/* Generic structure for all addresses */
union {
+ /* generic device format */
struct {
- u64 blk : NVM_BLK_BITS;
- u64 pg : NVM_PG_BITS;
- u64 sec : NVM_SEC_BITS;
- u64 pl : NVM_PL_BITS;
- u64 lun : NVM_LUN_BITS;
- u64 ch : NVM_CH_BITS;
- u64 reserved : 1;
+ u64 ch : NVM_GEN_CH_BITS;
+ u64 lun : NVM_GEN_LUN_BITS;
+ u64 blk : NVM_GEN_BLK_BITS;
+ u64 reserved : NVM_GEN_RESERVED;
+ } a;
+
+ /* 1.2 device format */
+ struct {
+ u64 ch : NVM_GEN_CH_BITS;
+ u64 lun : NVM_GEN_LUN_BITS;
+ u64 blk : NVM_GEN_BLK_BITS;
+ u64 pg : NVM_12_PG_BITS;
+ u64 pl : NVM_12_PL_BITS;
+ u64 sec : NVM_12_SEC_BITS;
+ u64 reserved : NVM_12_RESERVED;
} g;
+ /* 2.0 device format */
+ struct {
+ u64 grp : NVM_GEN_CH_BITS;
+ u64 pu : NVM_GEN_LUN_BITS;
+ u64 chk : NVM_GEN_BLK_BITS;
+ u64 sec : NVM_20_SEC_BITS;
+ u64 reserved : NVM_20_RESERVED;
+ } m;
+
struct {
u64 line : 63;
u64 is_cached : 1;
@@ -382,16 +409,26 @@ static inline struct ppa_addr generic_to_dev_addr(struct nvm_tgt_dev *tgt_dev,
struct ppa_addr r)
{
struct nvm_geo *geo = &tgt_dev->geo;
- struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&geo->c.addrf;
struct ppa_addr l;
- l.ppa = ((u64)r.g.ch) << ppaf->ch_offset;
- l.ppa |= ((u64)r.g.lun) << ppaf->lun_offset;
- l.ppa |= ((u64)r.g.blk) << ppaf->blk_offset;
- l.ppa |= ((u64)r.g.pg) << ppaf->pg_offset;
- l.ppa |= ((u64)r.g.pl) << ppaf->pln_offset;
- l.ppa |= ((u64)r.g.sec) << ppaf->sec_offset;
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&geo->c.addrf;
+
+ l.ppa = ((u64)r.g.ch) << ppaf->ch_offset;
+ l.ppa |= ((u64)r.g.lun) << ppaf->lun_offset;
+ l.ppa |= ((u64)r.g.blk) << ppaf->blk_offset;
+ l.ppa |= ((u64)r.g.pg) << ppaf->pg_offset;
+ l.ppa |= ((u64)r.g.pl) << ppaf->pln_offset;
+ l.ppa |= ((u64)r.g.sec) << ppaf->sec_offset;
+ } else {
+ struct nvm_addr_format *lbaf = &geo->c.addrf;
+
+ l.ppa = ((u64)r.m.grp) << lbaf->ch_offset;
+ l.ppa |= ((u64)r.m.pu) << lbaf->lun_offset;
+ l.ppa |= ((u64)r.m.chk) << lbaf->chk_offset;
+ l.ppa |= ((u64)r.m.sec) << lbaf->sec_offset;
+ }
return l;
}
@@ -400,18 +437,28 @@ static inline struct ppa_addr dev_to_generic_addr(struct nvm_tgt_dev *tgt_dev,
struct ppa_addr r)
{
struct nvm_geo *geo = &tgt_dev->geo;
- struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&geo->c.addrf;
struct ppa_addr l;
l.ppa = 0;
- l.g.ch = (r.ppa & ppaf->ch_mask) >> ppaf->ch_offset;
- l.g.lun = (r.ppa & ppaf->lun_mask) >> ppaf->lun_offset;
- l.g.blk = (r.ppa & ppaf->blk_mask) >> ppaf->blk_offset;
- l.g.pg = (r.ppa & ppaf->pg_mask) >> ppaf->pg_offset;
- l.g.pl = (r.ppa & ppaf->pln_mask) >> ppaf->pln_offset;
- l.g.sec = (r.ppa & ppaf->sec_mask) >> ppaf->sec_offset;
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&geo->c.addrf;
+
+ l.g.ch = (r.ppa & ppaf->ch_mask) >> ppaf->ch_offset;
+ l.g.lun = (r.ppa & ppaf->lun_mask) >> ppaf->lun_offset;
+ l.g.blk = (r.ppa & ppaf->blk_mask) >> ppaf->blk_offset;
+ l.g.pg = (r.ppa & ppaf->pg_mask) >> ppaf->pg_offset;
+ l.g.pl = (r.ppa & ppaf->pln_mask) >> ppaf->pln_offset;
+ l.g.sec = (r.ppa & ppaf->sec_mask) >> ppaf->sec_offset;
+ } else {
+ struct nvm_addr_format *lbaf = &geo->c.addrf;
+
+ l.m.grp = (r.ppa & lbaf->ch_mask) >> lbaf->ch_offset;
+ l.m.pu = (r.ppa & lbaf->lun_mask) >> lbaf->lun_offset;
+ l.m.chk = (r.ppa & lbaf->chk_mask) >> lbaf->chk_offset;
+ l.m.sec = (r.ppa & lbaf->sec_mask) >> lbaf->sec_offset;
+ }
return l;
}
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 11/19] lightnvm: pblk: rename ppaf* to addrf*
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (9 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 10/19] lightnvm: add support for 2.0 address format Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 12/19] lightnvn: pblk: use generic address format Javier González
` (8 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
In preparation for 2.0 support in pblk, rename variables referring to
the address format to addrf and reserve ppaf for the 1.2 path.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-init.c | 8 ++++----
drivers/lightnvm/pblk-sysfs.c | 4 ++--
drivers/lightnvm/pblk.h | 16 ++++++++--------
3 files changed, 14 insertions(+), 14 deletions(-)
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index 30aec863f136..ec39800eea42 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -80,7 +80,7 @@ static size_t pblk_trans_map_size(struct pblk *pblk)
{
int entry_size = 8;
- if (pblk->ppaf_bitsize < 32)
+ if (pblk->addrf_len < 32)
entry_size = 4;
return entry_size * pblk->rl.nr_secs;
@@ -198,7 +198,7 @@ static int pblk_set_addrf_12(struct nvm_geo *geo,
return dst->blk_offset + src->blk_len;
}
-static int pblk_set_ppaf(struct pblk *pblk)
+static int pblk_set_addrf(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
@@ -210,7 +210,7 @@ static int pblk_set_ppaf(struct pblk *pblk)
return -EINVAL;
}
- pblk->ppaf_bitsize = pblk_set_addrf_12(geo, (void *)&pblk->ppaf);
+ pblk->addrf_len = pblk_set_addrf_12(geo, (void *)&pblk->addrf);
return 0;
}
@@ -319,7 +319,7 @@ static int pblk_core_init(struct pblk *pblk)
if (!pblk->r_end_wq)
goto free_bb_wq;
- if (pblk_set_ppaf(pblk))
+ if (pblk_set_addrf(pblk))
goto free_r_end_wq;
if (pblk_rwb_init(pblk))
diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
index 81288aa9162a..d3b50741b691 100644
--- a/drivers/lightnvm/pblk-sysfs.c
+++ b/drivers/lightnvm/pblk-sysfs.c
@@ -117,12 +117,12 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
struct nvm_addr_format_12 *geo_ppaf;
ssize_t sz = 0;
- ppaf = (struct nvm_addr_format_12 *)&pblk->ppaf;
+ ppaf = (struct nvm_addr_format_12 *)&pblk->addrf;
geo_ppaf = (struct nvm_addr_format_12 *)&geo->c.addrf;
sz = snprintf(page, PAGE_SIZE,
"pblk:(s:%d)ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
- pblk->ppaf_bitsize,
+ pblk->addrf_len,
ppaf->ch_offset, ppaf->ch_len,
ppaf->lun_offset, ppaf->lun_len,
ppaf->blk_offset, ppaf->blk_len,
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index 4f7a365436f1..46b29a492f74 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -570,8 +570,8 @@ struct pblk {
struct pblk_line_mgmt l_mg; /* Line management */
struct pblk_line_meta lm; /* Line metadata */
- struct nvm_addr_format ppaf;
- int ppaf_bitsize;
+ struct nvm_addr_format addrf;
+ int addrf_len;
struct pblk_rb rwb;
@@ -948,7 +948,7 @@ static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
u64 line_id)
{
struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&pblk->ppaf;
+ (struct nvm_addr_format_12 *)&pblk->addrf;
struct ppa_addr ppa;
ppa.ppa = 0;
@@ -966,7 +966,7 @@ static inline u64 pblk_dev_ppa_to_line_addr(struct pblk *pblk,
struct ppa_addr p)
{
struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&pblk->ppaf;
+ (struct nvm_addr_format_12 *)&pblk->addrf;
u64 paddr;
paddr = (u64)p.g.ch << ppaf->ch_offset;
@@ -991,7 +991,7 @@ static inline struct ppa_addr pblk_ppa32_to_ppa64(struct pblk *pblk, u32 ppa32)
ppa64.c.is_cached = 1;
} else {
struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&pblk->ppaf;
+ (struct nvm_addr_format_12 *)&pblk->addrf;
ppa64.g.ch = (ppa32 & ppaf->ch_mask) >> ppaf->ch_offset;
ppa64.g.lun = (ppa32 & ppaf->lun_mask) >> ppaf->lun_offset;
@@ -1015,7 +1015,7 @@ static inline u32 pblk_ppa64_to_ppa32(struct pblk *pblk, struct ppa_addr ppa64)
ppa32 |= 1U << 31;
} else {
struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&pblk->ppaf;
+ (struct nvm_addr_format_12 *)&pblk->addrf;
ppa32 |= ppa64.g.ch << ppaf->ch_offset;
ppa32 |= ppa64.g.lun << ppaf->lun_offset;
@@ -1033,7 +1033,7 @@ static inline struct ppa_addr pblk_trans_map_get(struct pblk *pblk,
{
struct ppa_addr ppa;
- if (pblk->ppaf_bitsize < 32) {
+ if (pblk->addrf_len < 32) {
u32 *map = (u32 *)pblk->trans_map;
ppa = pblk_ppa32_to_ppa64(pblk, map[lba]);
@@ -1049,7 +1049,7 @@ static inline struct ppa_addr pblk_trans_map_get(struct pblk *pblk,
static inline void pblk_trans_map_set(struct pblk *pblk, sector_t lba,
struct ppa_addr ppa)
{
- if (pblk->ppaf_bitsize < 32) {
+ if (pblk->addrf_len < 32) {
u32 *map = (u32 *)pblk->trans_map;
map[lba] = pblk_ppa64_to_ppa32(pblk, ppa);
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 12/19] lightnvn: pblk: use generic address format
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (10 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 11/19] lightnvm: pblk: rename ppaf* to addrf* Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 13/19] lightnvm: make address conversions depend on generic device Javier González
` (7 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Use the generic address format on common address manipulations.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-core.c | 10 +++++-----
drivers/lightnvm/pblk-map.c | 4 ++--
drivers/lightnvm/pblk-sysfs.c | 4 ++--
drivers/lightnvm/pblk.h | 4 ++--
4 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/drivers/lightnvm/pblk-core.c b/drivers/lightnvm/pblk-core.c
index 7c726003a8d2..1d7c6313f3d9 100644
--- a/drivers/lightnvm/pblk-core.c
+++ b/drivers/lightnvm/pblk-core.c
@@ -885,7 +885,7 @@ int pblk_line_erase(struct pblk *pblk, struct pblk_line *line)
}
ppa = pblk->luns[bit].bppa; /* set ch and lun */
- ppa.g.blk = line->id;
+ ppa.a.blk = line->id;
atomic_dec(&line->left_eblks);
WARN_ON(test_and_set_bit(bit, line->erase_bitmap));
@@ -1693,8 +1693,8 @@ static void __pblk_down_page(struct pblk *pblk, struct ppa_addr *ppa_list,
int i;
for (i = 1; i < nr_ppas; i++)
- WARN_ON(ppa_list[0].g.lun != ppa_list[i].g.lun ||
- ppa_list[0].g.ch != ppa_list[i].g.ch);
+ WARN_ON(ppa_list[0].a.lun != ppa_list[i].a.lun ||
+ ppa_list[0].a.ch != ppa_list[i].a.ch);
#endif
ret = down_timeout(&rlun->wr_sem, msecs_to_jiffies(30000));
@@ -1738,8 +1738,8 @@ void pblk_up_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas)
int i;
for (i = 1; i < nr_ppas; i++)
- WARN_ON(ppa_list[0].g.lun != ppa_list[i].g.lun ||
- ppa_list[0].g.ch != ppa_list[i].g.ch);
+ WARN_ON(ppa_list[0].a.lun != ppa_list[i].a.lun ||
+ ppa_list[0].a.ch != ppa_list[i].a.ch);
#endif
rlun = &pblk->luns[pos];
diff --git a/drivers/lightnvm/pblk-map.c b/drivers/lightnvm/pblk-map.c
index 04e08d76ea5f..20dbaa89c9df 100644
--- a/drivers/lightnvm/pblk-map.c
+++ b/drivers/lightnvm/pblk-map.c
@@ -127,7 +127,7 @@ void pblk_map_erase_rq(struct pblk *pblk, struct nvm_rq *rqd,
atomic_dec(&e_line->left_eblks);
*erase_ppa = rqd->ppa_list[i];
- erase_ppa->g.blk = e_line->id;
+ erase_ppa->a.blk = e_line->id;
spin_unlock(&e_line->lock);
@@ -168,6 +168,6 @@ void pblk_map_erase_rq(struct pblk *pblk, struct nvm_rq *rqd,
set_bit(bit, e_line->erase_bitmap);
atomic_dec(&e_line->left_eblks);
*erase_ppa = pblk->luns[bit].bppa; /* set ch and lun */
- erase_ppa->g.blk = e_line->id;
+ erase_ppa->a.blk = e_line->id;
}
}
diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
index d3b50741b691..ccfb3abd2773 100644
--- a/drivers/lightnvm/pblk-sysfs.c
+++ b/drivers/lightnvm/pblk-sysfs.c
@@ -39,8 +39,8 @@ static ssize_t pblk_sysfs_luns_show(struct pblk *pblk, char *page)
sz += snprintf(page + sz, PAGE_SIZE - sz,
"pblk: pos:%d, ch:%d, lun:%d - %d\n",
i,
- rlun->bppa.g.ch,
- rlun->bppa.g.lun,
+ rlun->bppa.a.ch,
+ rlun->bppa.a.lun,
active);
}
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index 46b29a492f74..6e1fcd1a538a 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -936,12 +936,12 @@ static inline int pblk_pad_distance(struct pblk *pblk)
static inline int pblk_ppa_to_line(struct ppa_addr p)
{
- return p.g.blk;
+ return p.a.blk;
}
static inline int pblk_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
{
- return p.g.lun * geo->num_ch + p.g.ch;
+ return p.a.lun * geo->num_ch + p.a.ch;
}
static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 13/19] lightnvm: make address conversions depend on generic device
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (11 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 12/19] lightnvn: pblk: use generic address format Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 14/19] nvme: make nvme_get_log_ext available Javier González
` (6 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
On address conversions, use the generic device, instead of the target
device. This allows to use conversions outside of the target's realm.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/core.c | 4 ++--
include/linux/lightnvm.h | 20 ++++++++++----------
2 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index 92ea6e39dd64..fae5e8c6fcd0 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -588,7 +588,7 @@ static void nvm_ppa_tgt_to_dev(struct nvm_tgt_dev *tgt_dev,
for (i = 0; i < nr_ppas; i++) {
nvm_map_to_dev(tgt_dev, &ppa_list[i]);
- ppa_list[i] = generic_to_dev_addr(tgt_dev, ppa_list[i]);
+ ppa_list[i] = generic_to_dev_addr(tgt_dev->parent, ppa_list[i]);
}
}
@@ -598,7 +598,7 @@ static void nvm_ppa_dev_to_tgt(struct nvm_tgt_dev *tgt_dev,
int i;
for (i = 0; i < nr_ppas; i++) {
- ppa_list[i] = dev_to_generic_addr(tgt_dev, ppa_list[i]);
+ ppa_list[i] = dev_to_generic_addr(tgt_dev->parent, ppa_list[i]);
nvm_map_to_tgt(tgt_dev, &ppa_list[i]);
}
}
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 8cda1125c34a..2b909592568a 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -405,15 +405,15 @@ struct nvm_dev {
struct list_head targets;
};
-static inline struct ppa_addr generic_to_dev_addr(struct nvm_tgt_dev *tgt_dev,
+static inline struct ppa_addr generic_to_dev_addr(struct nvm_dev *dev,
struct ppa_addr r)
{
- struct nvm_geo *geo = &tgt_dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
struct ppa_addr l;
- if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ if (dev_geo->c.version == NVM_OCSSD_SPEC_12) {
struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&geo->c.addrf;
+ (struct nvm_addr_format_12 *)&dev_geo->c.addrf;
l.ppa = ((u64)r.g.ch) << ppaf->ch_offset;
l.ppa |= ((u64)r.g.lun) << ppaf->lun_offset;
@@ -422,7 +422,7 @@ static inline struct ppa_addr generic_to_dev_addr(struct nvm_tgt_dev *tgt_dev,
l.ppa |= ((u64)r.g.pl) << ppaf->pln_offset;
l.ppa |= ((u64)r.g.sec) << ppaf->sec_offset;
} else {
- struct nvm_addr_format *lbaf = &geo->c.addrf;
+ struct nvm_addr_format *lbaf = &dev_geo->c.addrf;
l.ppa = ((u64)r.m.grp) << lbaf->ch_offset;
l.ppa |= ((u64)r.m.pu) << lbaf->lun_offset;
@@ -433,17 +433,17 @@ static inline struct ppa_addr generic_to_dev_addr(struct nvm_tgt_dev *tgt_dev,
return l;
}
-static inline struct ppa_addr dev_to_generic_addr(struct nvm_tgt_dev *tgt_dev,
+static inline struct ppa_addr dev_to_generic_addr(struct nvm_dev *dev,
struct ppa_addr r)
{
- struct nvm_geo *geo = &tgt_dev->geo;
+ struct nvm_geo *dev_geo = &dev->geo;
struct ppa_addr l;
l.ppa = 0;
- if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ if (dev_geo->c.version == NVM_OCSSD_SPEC_12) {
struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&geo->c.addrf;
+ (struct nvm_addr_format_12 *)&dev_geo->c.addrf;
l.g.ch = (r.ppa & ppaf->ch_mask) >> ppaf->ch_offset;
l.g.lun = (r.ppa & ppaf->lun_mask) >> ppaf->lun_offset;
@@ -452,7 +452,7 @@ static inline struct ppa_addr dev_to_generic_addr(struct nvm_tgt_dev *tgt_dev,
l.g.pl = (r.ppa & ppaf->pln_mask) >> ppaf->pln_offset;
l.g.sec = (r.ppa & ppaf->sec_mask) >> ppaf->sec_offset;
} else {
- struct nvm_addr_format *lbaf = &geo->c.addrf;
+ struct nvm_addr_format *lbaf = &dev_geo->c.addrf;
l.m.grp = (r.ppa & lbaf->ch_mask) >> lbaf->ch_offset;
l.m.pu = (r.ppa & lbaf->lun_mask) >> lbaf->lun_offset;
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 14/19] nvme: make nvme_get_log_ext available
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (12 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 13/19] lightnvm: make address conversions depend on generic device Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 15/19] lightnvm: implement get log report chunk helpers Javier González
` (5 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Make nvme_get_log_ext available outside of core.c.
This is in preparation for using it in lightnvm.c
Signed-off-by: Javier González <javier@cnexlabs.>
---
drivers/nvme/host/nvme.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index 1ca08f4993ba..505f797f8c6c 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -396,6 +396,9 @@ int nvme_reset_ctrl(struct nvme_ctrl *ctrl);
int nvme_delete_ctrl(struct nvme_ctrl *ctrl);
int nvme_delete_ctrl_sync(struct nvme_ctrl *ctrl);
+int nvme_get_log_ext(struct nvme_ctrl *ctrl, struct nvme_ns *ns,
+ u8 log_page, void *log, size_t size, size_t offset);
+
extern const struct attribute_group nvme_ns_id_attr_group;
extern const struct block_device_operations nvme_ns_head_ops;
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 15/19] lightnvm: implement get log report chunk helpers
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (13 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 14/19] nvme: make nvme_get_log_ext available Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 16/19] lightnvm: define chunk states Javier González
` (4 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb
Cc: linux-block, linux-kernel, linux-nvme, Javier González,
Javier González
From: Javier González <javier@javigon.com>
The 2.0 spec provides a report chunk log page that can be retrieved
using the stangard nvme get log page. This replaces the dedicated
get/put bad block table in 1.2.
This patch implements the helper functions to allow targets retrieve the
chunk metadata using get log page
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/core.c | 11 +++++++
drivers/nvme/host/lightnvm.c | 74 ++++++++++++++++++++++++++++++++++++++++++++
include/linux/lightnvm.h | 24 ++++++++++++++
3 files changed, 109 insertions(+)
diff --git a/drivers/lightnvm/core.c b/drivers/lightnvm/core.c
index fae5e8c6fcd0..0f4114cb48ee 100644
--- a/drivers/lightnvm/core.c
+++ b/drivers/lightnvm/core.c
@@ -719,6 +719,17 @@ static void nvm_free_rqd_ppalist(struct nvm_tgt_dev *tgt_dev,
nvm_dev_dma_free(tgt_dev->parent, rqd->ppa_list, rqd->dma_ppa_list);
}
+int nvm_get_chunk_meta(struct nvm_tgt_dev *tgt_dev, struct nvm_chk_meta *meta,
+ struct ppa_addr ppa, int nchks)
+{
+ struct nvm_dev *dev = tgt_dev->parent;
+
+ nvm_ppa_tgt_to_dev(tgt_dev, &ppa, 1);
+
+ return dev->ops->get_chk_meta(tgt_dev->parent, meta,
+ (sector_t)ppa.ppa, nchks);
+}
+EXPORT_SYMBOL(nvm_get_chunk_meta);
int nvm_set_tgt_bb_tbl(struct nvm_tgt_dev *tgt_dev, struct ppa_addr *ppas,
int nr_ppas, int type)
diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index 377e6525e53f..7c4af132d072 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -35,6 +35,10 @@ enum nvme_nvm_admin_opcode {
nvme_nvm_admin_set_bb_tbl = 0xf1,
};
+enum nvme_nvm_log_page {
+ NVME_NVM_LOG_REPORT_CHUNK = 0xca,
+};
+
struct nvme_nvm_ph_rw {
__u8 opcode;
__u8 flags;
@@ -236,6 +240,16 @@ struct nvme_nvm_id20 {
__u8 vs[1024];
};
+struct nvme_nvm_chk_meta {
+ __u8 state;
+ __u8 type;
+ __u8 wi;
+ __u8 rsvd[5];
+ __le64 slba;
+ __le64 cnlb;
+ __le64 wp;
+};
+
/*
* Check we didn't inadvertently grow the command struct
*/
@@ -252,6 +266,9 @@ static inline void _nvme_nvm_check_size(void)
BUILD_BUG_ON(sizeof(struct nvme_nvm_bb_tbl) != 64);
BUILD_BUG_ON(sizeof(struct nvme_nvm_id20_addrf) != 8);
BUILD_BUG_ON(sizeof(struct nvme_nvm_id20) != NVME_IDENTIFY_DATA_SIZE);
+ BUILD_BUG_ON(sizeof(struct nvme_nvm_chk_meta) != 32);
+ BUILD_BUG_ON(sizeof(struct nvme_nvm_chk_meta) !=
+ sizeof(struct nvm_chk_meta));
}
static void nvme_nvm_set_addr_12(struct nvm_addr_format_12 *dst,
@@ -558,6 +575,61 @@ static int nvme_nvm_set_bb_tbl(struct nvm_dev *nvmdev, struct ppa_addr *ppas,
return ret;
}
+/*
+ * Expect the lba in device format
+ */
+static int nvme_nvm_get_chk_meta(struct nvm_dev *ndev,
+ struct nvm_chk_meta *meta,
+ sector_t slba, int nchks)
+{
+ struct nvm_geo *dev_geo = &ndev->geo;
+ struct nvme_ns *ns = ndev->q->queuedata;
+ struct nvme_ctrl *ctrl = ns->ctrl;
+ struct nvme_nvm_chk_meta *dev_meta = (struct nvme_nvm_chk_meta *)meta;
+ struct ppa_addr ppa;
+ size_t left = nchks * sizeof(struct nvme_nvm_chk_meta);
+ size_t log_pos, offset, len;
+ int ret, i;
+
+ /* Normalize lba address space to obtain log offset */
+ ppa.ppa = slba;
+ ppa = dev_to_generic_addr(ndev, ppa);
+
+ log_pos = ppa.m.chk;
+ log_pos += ppa.m.pu * dev_geo->c.num_chk;
+ log_pos += ppa.m.grp * dev_geo->num_lun * dev_geo->c.num_chk;
+
+ offset = log_pos * sizeof(struct nvme_nvm_chk_meta);
+
+ while (left) {
+ len = min_t(unsigned int, left, ctrl->max_hw_sectors << 9);
+
+ ret = nvme_get_log_ext(ctrl, ns, NVME_NVM_LOG_REPORT_CHUNK,
+ dev_meta, len, offset);
+ if (ret) {
+ dev_err(ctrl->device, "Get REPORT CHUNK log error\n");
+ break;
+ }
+
+ for (i = 0; i < len; i += sizeof(struct nvme_nvm_chk_meta)) {
+ meta->state = dev_meta->state;
+ meta->type = dev_meta->type;
+ meta->wi = dev_meta->wi;
+ meta->slba = le64_to_cpu(dev_meta->slba);
+ meta->cnlb = le64_to_cpu(dev_meta->cnlb);
+ meta->wp = le64_to_cpu(dev_meta->wp);
+
+ meta++;
+ dev_meta++;
+ }
+
+ offset += len;
+ left -= len;
+ }
+
+ return ret;
+}
+
static inline void nvme_nvm_rqtocmd(struct nvm_rq *rqd, struct nvme_ns *ns,
struct nvme_nvm_command *c)
{
@@ -689,6 +761,8 @@ static struct nvm_dev_ops nvme_nvm_dev_ops = {
.get_bb_tbl = nvme_nvm_get_bb_tbl,
.set_bb_tbl = nvme_nvm_set_bb_tbl,
+ .get_chk_meta = nvme_nvm_get_chk_meta,
+
.submit_io = nvme_nvm_submit_io,
.submit_io_sync = nvme_nvm_submit_io_sync,
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 2b909592568a..776ffdc2e137 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -81,10 +81,13 @@ struct nvm_rq;
struct nvm_id;
struct nvm_dev;
struct nvm_tgt_dev;
+struct nvm_chk_meta;
typedef int (nvm_id_fn)(struct nvm_dev *);
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 ppa_addr *, int, int);
+typedef int (nvm_get_chk_meta_fn)(struct nvm_dev *, struct nvm_chk_meta *,
+ sector_t, int);
typedef int (nvm_submit_io_fn)(struct nvm_dev *, struct nvm_rq *);
typedef int (nvm_submit_io_sync_fn)(struct nvm_dev *, struct nvm_rq *);
typedef void *(nvm_create_dma_pool_fn)(struct nvm_dev *, char *);
@@ -98,6 +101,8 @@ struct nvm_dev_ops {
nvm_op_bb_tbl_fn *get_bb_tbl;
nvm_op_set_bb_fn *set_bb_tbl;
+ nvm_get_chk_meta_fn *get_chk_meta;
+
nvm_submit_io_fn *submit_io;
nvm_submit_io_sync_fn *submit_io_sync;
@@ -227,6 +232,20 @@ struct nvm_addr_format {
u64 rsv_mask[2];
};
+/*
+ * Note: The structure size is linked to nvme_nvm_chk_meta such that the same
+ * buffer can be used when converting from little endian to cpu addressing.
+ */
+struct nvm_chk_meta {
+ u8 state;
+ u8 type;
+ u8 wi;
+ u8 rsvd[5];
+ u64 slba;
+ u64 cnlb;
+ u64 wp;
+};
+
struct nvm_target {
struct list_head list;
struct nvm_tgt_dev *dev;
@@ -502,6 +521,11 @@ extern struct nvm_dev *nvm_alloc_dev(int);
extern int nvm_register(struct nvm_dev *);
extern void nvm_unregister(struct nvm_dev *);
+
+extern int nvm_get_chunk_meta(struct nvm_tgt_dev *tgt_dev,
+ struct nvm_chk_meta *meta, struct ppa_addr ppa,
+ int nchks);
+
extern int nvm_set_tgt_bb_tbl(struct nvm_tgt_dev *, struct ppa_addr *,
int, int);
extern int nvm_submit_io(struct nvm_tgt_dev *, struct nvm_rq *);
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 16/19] lightnvm: define chunk states
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (14 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 15/19] lightnvm: implement get log report chunk helpers Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 17/19] lightnvm: pblk: implement get log report chunk Javier González
` (3 subsequent siblings)
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Define chunk states as given in the 2.0 spec. Also, add an extra chunk
state that signals that the chunk is in use by the host. This allows for
the chunk metadata to be "owned" by a target when active, thus
completing the chunk state machine from the host perspective and
facilitating sanity checks. This state is transparent for the device.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
include/linux/lightnvm.h | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/include/linux/lightnvm.h b/include/linux/lightnvm.h
index 776ffdc2e137..74830d5dbd6e 100644
--- a/include/linux/lightnvm.h
+++ b/include/linux/lightnvm.h
@@ -232,6 +232,20 @@ struct nvm_addr_format {
u64 rsv_mask[2];
};
+enum {
+ /* Chunk states */
+ NVM_CHK_ST_FREE = 1 << 0,
+ NVM_CHK_ST_CLOSED = 1 << 1,
+ NVM_CHK_ST_OPEN = 1 << 2,
+ NVM_CHK_ST_OFFLINE = 1 << 3,
+ NVM_CHK_ST_HOST_USE = 1 << 7,
+
+ /* Chunk types */
+ NVM_CHK_TP_W_SEQ = 1 << 0,
+ NVM_CHK_TP_W_RAN = 1 << 1,
+ NVM_CHK_TP_SZ_SPEC = 1 << 4,
+};
+
/*
* Note: The structure size is linked to nvme_nvm_chk_meta such that the same
* buffer can be used when converting from little endian to cpu addressing.
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 17/19] lightnvm: pblk: implement get log report chunk
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (15 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 16/19] lightnvm: define chunk states Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 19:04 ` Matias Bjørling
2018-02-26 13:17 ` [PATCH 18/19] lightnvm: pblk: refactor init/exit sequences Javier González
` (2 subsequent siblings)
19 siblings, 1 reply; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb
Cc: linux-block, linux-kernel, linux-nvme, Javier González,
Javier González
From: Javier González <javier@javigon.com>
In preparation of pblk supporting 2.0, implement the get log report
chunk in pblk.
This patch only replicates de bad block functionality as the rest of the
metadata requires new pblk functionality (e.g., wear-index to implement
wear-leveling). This functionality will come in future patches.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-core.c | 121 +++++++++++++++++++----
drivers/lightnvm/pblk-init.c | 218 ++++++++++++++++++++++++++++++------------
drivers/lightnvm/pblk-sysfs.c | 67 +++++++++++++
drivers/lightnvm/pblk.h | 20 ++++
4 files changed, 346 insertions(+), 80 deletions(-)
diff --git a/drivers/lightnvm/pblk-core.c b/drivers/lightnvm/pblk-core.c
index 1d7c6313f3d9..ce6a7cfdba66 100644
--- a/drivers/lightnvm/pblk-core.c
+++ b/drivers/lightnvm/pblk-core.c
@@ -44,11 +44,12 @@ static void pblk_line_mark_bb(struct work_struct *work)
}
static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
- struct ppa_addr *ppa)
+ struct ppa_addr ppa_addr)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
- int pos = pblk_ppa_to_pos(geo, *ppa);
+ struct ppa_addr *ppa;
+ int pos = pblk_ppa_to_pos(geo, ppa_addr);
pr_debug("pblk: erase failed: line:%d, pos:%d\n", line->id, pos);
atomic_long_inc(&pblk->erase_failed);
@@ -58,6 +59,15 @@ static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
pr_err("pblk: attempted to erase bb: line:%d, pos:%d\n",
line->id, pos);
+ /* Not necessary to mark bad blocks on 2.0 spec. */
+ if (geo->c.version == NVM_OCSSD_SPEC_20)
+ return;
+
+ ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
+ if (!ppa)
+ return;
+
+ *ppa = ppa_addr;
pblk_gen_run_ws(pblk, NULL, ppa, pblk_line_mark_bb,
GFP_ATOMIC, pblk->bb_wq);
}
@@ -69,16 +79,8 @@ static void __pblk_end_io_erase(struct pblk *pblk, struct nvm_rq *rqd)
line = &pblk->lines[pblk_ppa_to_line(rqd->ppa_addr)];
atomic_dec(&line->left_seblks);
- if (rqd->error) {
- struct ppa_addr *ppa;
-
- ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
- if (!ppa)
- return;
-
- *ppa = rqd->ppa_addr;
- pblk_mark_bb(pblk, line, ppa);
- }
+ if (rqd->error)
+ pblk_mark_bb(pblk, line, rqd->ppa_addr);
atomic_dec(&pblk->inflight_io);
}
@@ -92,6 +94,50 @@ static void pblk_end_io_erase(struct nvm_rq *rqd)
mempool_free(rqd, pblk->e_rq_pool);
}
+/*
+ * Get information for all chunks from the device.
+ *
+ * The caller is responsible for freeing the returned structure
+ */
+struct nvm_chk_meta *pblk_chunk_get_info(struct pblk *pblk)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ struct nvm_chk_meta *meta;
+ struct ppa_addr ppa;
+ unsigned long len;
+ int ret;
+
+ ppa.ppa = 0;
+
+ len = geo->all_chunks * sizeof(*meta);
+ meta = kzalloc(len, GFP_KERNEL);
+ if (!meta)
+ return ERR_PTR(-ENOMEM);
+
+ ret = nvm_get_chunk_meta(dev, meta, ppa, geo->all_chunks);
+ if (ret) {
+ pr_err("pblk: could not get chunk metadata (%d)\n", ret);
+ kfree(meta);
+ return ERR_PTR(-EIO);
+ }
+
+ return meta;
+}
+
+struct nvm_chk_meta *pblk_chunk_get_off(struct pblk *pblk,
+ struct nvm_chk_meta *meta,
+ struct ppa_addr ppa)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ int ch_off = ppa.m.grp * geo->c.num_chk * geo->num_lun;
+ int lun_off = ppa.m.pu * geo->c.num_chk;
+ int chk_off = ppa.m.chk;
+
+ return meta + ch_off + lun_off + chk_off;
+}
+
void __pblk_map_invalidate(struct pblk *pblk, struct pblk_line *line,
u64 paddr)
{
@@ -1094,10 +1140,38 @@ static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
return 1;
}
+static int pblk_prepare_new_line(struct pblk *pblk, struct pblk_line *line)
+{
+ struct pblk_line_meta *lm = &pblk->lm;
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ int blk_to_erase = atomic_read(&line->blk_in_line);
+ int i;
+
+ for (i = 0; i < lm->blk_per_line; i++) {
+ int state = line->chks[i].state;
+ struct pblk_lun *rlun = &pblk->luns[i];
+
+ /* Free chunks should not be erased */
+ if (state & NVM_CHK_ST_FREE) {
+ set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
+ line->erase_bitmap);
+ blk_to_erase--;
+ line->chks[i].state = NVM_CHK_ST_HOST_USE;
+ }
+
+ WARN_ONCE(state & NVM_CHK_ST_OPEN,
+ "pblk: open chunk in new line: %d\n",
+ line->id);
+ }
+
+ return blk_to_erase;
+}
+
static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
{
struct pblk_line_meta *lm = &pblk->lm;
- int blk_in_line = atomic_read(&line->blk_in_line);
+ int blk_to_erase;
line->map_bitmap = kzalloc(lm->sec_bitmap_len, GFP_ATOMIC);
if (!line->map_bitmap)
@@ -1110,7 +1184,21 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
return -ENOMEM;
}
+ /* Bad blocks do not need to be erased */
+ bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
+
spin_lock(&line->lock);
+
+ /* If we have not written to this line, we need to mark up free chunks
+ * as already erased
+ */
+ if (line->state == PBLK_LINESTATE_NEW) {
+ blk_to_erase = pblk_prepare_new_line(pblk, line);
+ line->state = PBLK_LINESTATE_FREE;
+ } else {
+ blk_to_erase = atomic_read(&line->blk_in_line);
+ }
+
if (line->state != PBLK_LINESTATE_FREE) {
kfree(line->map_bitmap);
kfree(line->invalid_bitmap);
@@ -1122,15 +1210,12 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
line->state = PBLK_LINESTATE_OPEN;
- atomic_set(&line->left_eblks, blk_in_line);
- atomic_set(&line->left_seblks, blk_in_line);
+ atomic_set(&line->left_eblks, blk_to_erase);
+ atomic_set(&line->left_seblks, blk_to_erase);
line->meta_distance = lm->meta_distance;
spin_unlock(&line->lock);
- /* Bad blocks do not need to be erased */
- bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
-
kref_init(&line->ref);
return 0;
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index ec39800eea42..cf4f49d48aed 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -401,6 +401,7 @@ static void pblk_line_meta_free(struct pblk_line *line)
{
kfree(line->blk_bitmap);
kfree(line->erase_bitmap);
+ kfree(line->chks);
}
static void pblk_lines_free(struct pblk *pblk)
@@ -440,54 +441,44 @@ static int pblk_bb_get_tbl(struct nvm_tgt_dev *dev, struct pblk_lun *rlun,
return 0;
}
-static void *pblk_bb_get_log(struct pblk *pblk)
+static void *pblk_bb_get_meta(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
- u8 *log;
+ u8 *meta;
int i, nr_blks, blk_per_lun;
int ret;
blk_per_lun = geo->c.num_chk * geo->c.pln_mode;
nr_blks = blk_per_lun * geo->all_luns;
- log = kmalloc(nr_blks, GFP_KERNEL);
- if (!log)
+ meta = kmalloc(nr_blks, GFP_KERNEL);
+ if (!meta)
return ERR_PTR(-ENOMEM);
for (i = 0; i < geo->all_luns; i++) {
struct pblk_lun *rlun = &pblk->luns[i];
- u8 *log_pos = log + i * blk_per_lun;
+ u8 *meta_pos = meta + i * blk_per_lun;
- ret = pblk_bb_get_tbl(dev, rlun, log_pos, blk_per_lun);
+ ret = pblk_bb_get_tbl(dev, rlun, meta_pos, blk_per_lun);
if (ret) {
- kfree(log);
+ kfree(meta);
return ERR_PTR(-EIO);
}
}
- return log;
+ return meta;
}
-static int pblk_bb_line(struct pblk *pblk, struct pblk_line *line,
- u8 *bb_log, int blk_per_line)
+static void *pblk_chunk_get_meta(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
- int i, bb_cnt = 0;
- for (i = 0; i < blk_per_line; i++) {
- struct pblk_lun *rlun = &pblk->luns[i];
- u8 *lun_bb_log = bb_log + i * blk_per_line;
-
- if (lun_bb_log[line->id] == NVM_BLK_T_FREE)
- continue;
-
- set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
- bb_cnt++;
- }
-
- return bb_cnt;
+ if (geo->c.version == NVM_OCSSD_SPEC_12)
+ return pblk_bb_get_meta(pblk);
+ else
+ return pblk_chunk_get_info(pblk);
}
static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
@@ -516,6 +507,7 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
rlun = &pblk->luns[i];
rlun->bppa = luns[lunid];
+ rlun->chunk_bppa = luns[i];
sema_init(&rlun->wr_sem, 1);
}
@@ -695,8 +687,125 @@ static int pblk_lines_alloc_metadata(struct pblk *pblk)
return -ENOMEM;
}
-static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
- void *chunk_log, long *nr_bad_blks)
+static int pblk_setup_line_meta_12(struct pblk *pblk, struct pblk_line *line,
+ void *chunk_meta)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ struct pblk_line_meta *lm = &pblk->lm;
+ int i, chk_per_lun, nr_bad_chks = 0;
+
+ chk_per_lun = geo->c.num_chk * geo->c.pln_mode;
+
+ for (i = 0; i < lm->blk_per_line; i++) {
+ struct pblk_chunk *chunk = &line->chks[i];
+ struct pblk_lun *rlun = &pblk->luns[i];
+ u8 *lun_bb_meta = chunk_meta + i * chk_per_lun;
+
+ /*
+ * In 1.2 spec. chunk state is not persisted by the device. Thus
+ * some of the values are reset each time pblk is instantiated.
+ */
+ if (lun_bb_meta[line->id] == NVM_BLK_T_FREE)
+ chunk->state = NVM_CHK_ST_HOST_USE;
+ else
+ chunk->state = NVM_CHK_ST_OFFLINE;
+
+ chunk->type = NVM_CHK_TP_W_SEQ;
+ chunk->wi = 0;
+ chunk->slba = -1;
+ chunk->cnlb = geo->c.clba;
+ chunk->wp = 0;
+
+ if (!(chunk->state & NVM_CHK_ST_OFFLINE))
+ continue;
+
+ set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
+ nr_bad_chks++;
+ }
+
+ return nr_bad_chks;
+}
+
+static int pblk_setup_line_meta_20(struct pblk *pblk, struct pblk_line *line,
+ struct nvm_chk_meta *meta)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ struct pblk_line_meta *lm = &pblk->lm;
+ int i, nr_bad_chks = 0;
+
+ for (i = 0; i < lm->blk_per_line; i++) {
+ struct pblk_chunk *chunk = &line->chks[i];
+ struct pblk_lun *rlun = &pblk->luns[i];
+ struct nvm_chk_meta *chunk_meta;
+ struct ppa_addr ppa;
+
+ ppa = rlun->chunk_bppa;
+ ppa.m.chk = line->id;
+ chunk_meta = pblk_chunk_get_off(pblk, meta, ppa);
+
+ chunk->state = chunk_meta->state;
+ chunk->type = chunk_meta->type;
+ chunk->wi = chunk_meta->wi;
+ chunk->slba = chunk_meta->slba;
+ chunk->cnlb = chunk_meta->cnlb;
+ chunk->wp = chunk_meta->wp;
+
+ if (!(chunk->state & NVM_CHK_ST_OFFLINE))
+ continue;
+
+ if (chunk->type & NVM_CHK_TP_SZ_SPEC) {
+ WARN_ONCE(1, "pblk: custom-sized chunks unsupported\n");
+ continue;
+ }
+
+ set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
+ line->blk_bitmap);
+ nr_bad_chks++;
+ }
+
+ return nr_bad_chks;
+}
+
+static long pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
+ void *chunk_meta, int line_id)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ struct pblk_line_mgmt *l_mg = &pblk->l_mg;
+ struct pblk_line_meta *lm = &pblk->lm;
+ long nr_bad_chks, chk_in_line;
+
+ line->pblk = pblk;
+ line->id = line_id;
+ line->type = PBLK_LINETYPE_FREE;
+ line->state = PBLK_LINESTATE_NEW;
+ line->gc_group = PBLK_LINEGC_NONE;
+ line->vsc = &l_mg->vsc_list[line_id];
+ spin_lock_init(&line->lock);
+
+ if (geo->c.version == NVM_OCSSD_SPEC_12)
+ nr_bad_chks = pblk_setup_line_meta_12(pblk, line, chunk_meta);
+ else
+ nr_bad_chks = pblk_setup_line_meta_20(pblk, line, chunk_meta);
+
+ chk_in_line = lm->blk_per_line - nr_bad_chks;
+ if (nr_bad_chks < 0 || nr_bad_chks > lm->blk_per_line ||
+ chk_in_line < lm->min_blk_line) {
+ line->state = PBLK_LINESTATE_BAD;
+ list_add_tail(&line->list, &l_mg->bad_list);
+ return 0;
+ }
+
+ atomic_set(&line->blk_in_line, chk_in_line);
+ list_add_tail(&line->list, &l_mg->free_list);
+ l_mg->nr_free_lines++;
+
+ return chk_in_line;
+}
+
+static int pblk_alloc_line_meta(struct pblk *pblk, struct pblk_line *line)
{
struct pblk_line_meta *lm = &pblk->lm;
@@ -710,7 +819,13 @@ static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
return -ENOMEM;
}
- *nr_bad_blks = pblk_bb_line(pblk, line, chunk_log, lm->blk_per_line);
+ line->chks = kmalloc(lm->blk_per_line * sizeof(struct pblk_chunk),
+ GFP_KERNEL);
+ if (!line->chks) {
+ kfree(line->erase_bitmap);
+ kfree(line->blk_bitmap);
+ return -ENOMEM;
+ }
return 0;
}
@@ -722,9 +837,9 @@ static int pblk_lines_init(struct pblk *pblk)
struct pblk_line_mgmt *l_mg = &pblk->l_mg;
struct pblk_line_meta *lm = &pblk->lm;
struct pblk_line *line;
- void *chunk_log;
+ void *chunk_meta;
unsigned int smeta_len, emeta_len;
- long nr_bad_blks = 0, nr_free_blks = 0;
+ long nr_free_chks = 0;
int bb_distance, max_write_ppas;
int i, ret;
@@ -743,6 +858,7 @@ static int pblk_lines_init(struct pblk *pblk)
l_mg->log_line = l_mg->data_line = NULL;
l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
l_mg->nr_free_lines = 0;
+ atomic_set(&l_mg->sysfs_line_state, -1);
bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
lm->sec_per_line = geo->c.clba * geo->all_luns;
@@ -841,53 +957,31 @@ static int pblk_lines_init(struct pblk *pblk)
goto fail_free_bb_aux;
}
- chunk_log = pblk_bb_get_log(pblk);
- if (IS_ERR(chunk_log)) {
- pr_err("pblk: could not get bad block log (%lu)\n",
- PTR_ERR(chunk_log));
- ret = PTR_ERR(chunk_log);
+ chunk_meta = pblk_chunk_get_meta(pblk);
+ if (IS_ERR(chunk_meta)) {
+ pr_err("pblk: could not get chunk log (%lu)\n",
+ PTR_ERR(chunk_meta));
+ ret = PTR_ERR(chunk_meta);
goto fail_free_lines;
}
for (i = 0; i < l_mg->nr_lines; i++) {
- int chk_in_line;
-
line = &pblk->lines[i];
- line->pblk = pblk;
- line->id = i;
- line->type = PBLK_LINETYPE_FREE;
- line->state = PBLK_LINESTATE_FREE;
- line->gc_group = PBLK_LINEGC_NONE;
- line->vsc = &l_mg->vsc_list[i];
- spin_lock_init(&line->lock);
-
- ret = pblk_setup_line_meta(pblk, line, chunk_log, &nr_bad_blks);
+ ret = pblk_alloc_line_meta(pblk, line);
if (ret)
- goto fail_free_chunk_log;
+ goto fail_free_chunk_meta;
- chk_in_line = lm->blk_per_line - nr_bad_blks;
- if (nr_bad_blks < 0 || nr_bad_blks > lm->blk_per_line ||
- chk_in_line < lm->min_blk_line) {
- line->state = PBLK_LINESTATE_BAD;
- list_add_tail(&line->list, &l_mg->bad_list);
- continue;
- }
-
- nr_free_blks += chk_in_line;
- atomic_set(&line->blk_in_line, chk_in_line);
-
- l_mg->nr_free_lines++;
- list_add_tail(&line->list, &l_mg->free_list);
+ nr_free_chks += pblk_setup_line_meta(pblk, line, chunk_meta, i);
}
- pblk_set_provision(pblk, nr_free_blks);
+ pblk_set_provision(pblk, nr_free_chks);
- kfree(chunk_log);
+ kfree(chunk_meta);
return 0;
-fail_free_chunk_log:
- kfree(chunk_log);
+fail_free_chunk_meta:
+ kfree(chunk_meta);
while (--i >= 0)
pblk_line_meta_free(&pblk->lines[i]);
fail_free_lines:
diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
index ccfb3abd2773..1ce5b956c622 100644
--- a/drivers/lightnvm/pblk-sysfs.c
+++ b/drivers/lightnvm/pblk-sysfs.c
@@ -142,6 +142,40 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
return sz;
}
+static ssize_t pblk_sysfs_line_state_show(struct pblk *pblk, char *page)
+{
+ struct pblk_line_meta *lm = &pblk->lm;
+ struct pblk_line_mgmt *l_mg = &pblk->l_mg;
+ struct pblk_line *line;
+ int line_id = atomic_read(&l_mg->sysfs_line_state);
+ ssize_t sz = 0;
+ int i;
+
+ if (line_id < 0 || line_id >= l_mg->nr_lines)
+ return 0;
+
+ sz = snprintf(page, PAGE_SIZE,
+ "line\tchunk\tstate\ttype\twear-index\tslba\t\tcnlb\twp\n");
+
+ line = &pblk->lines[line_id];
+
+ for (i = 0; i < lm->blk_per_line; i++) {
+ struct pblk_chunk *chunk = &line->chks[i];
+
+ sz += snprintf(page + sz, PAGE_SIZE - sz,
+ "%d\t%d\t%d\t%d\t%d\t\t%llu\t\t%llu\t%llu\n",
+ line->id, i,
+ chunk->state,
+ chunk->type,
+ chunk->wi,
+ chunk->slba,
+ chunk->cnlb,
+ chunk->wp);
+ }
+
+ return sz;
+}
+
static ssize_t pblk_sysfs_lines(struct pblk *pblk, char *page)
{
struct nvm_tgt_dev *dev = pblk->dev;
@@ -398,6 +432,29 @@ static ssize_t pblk_sysfs_stats_debug(struct pblk *pblk, char *page)
}
#endif
+
+static ssize_t pblk_sysfs_line_state_store(struct pblk *pblk, const char *page,
+ size_t len)
+{
+ struct pblk_line_mgmt *l_mg = &pblk->l_mg;
+ size_t c_len;
+ int line_id;
+
+ c_len = strcspn(page, "\n");
+ if (c_len >= len)
+ return -EINVAL;
+
+ if (kstrtouint(page, 0, &line_id))
+ return -EINVAL;
+
+ if (line_id < 0 || line_id >= l_mg->nr_lines)
+ return -EINVAL;
+
+ atomic_set(&l_mg->sysfs_line_state, line_id);
+
+ return len;
+}
+
static ssize_t pblk_sysfs_gc_force(struct pblk *pblk, const char *page,
size_t len)
{
@@ -529,6 +586,11 @@ static struct attribute sys_lines_info_attr = {
.mode = 0444,
};
+static struct attribute sys_line_state_attr = {
+ .name = "line_state",
+ .mode = 0644,
+};
+
static struct attribute sys_gc_force = {
.name = "gc_force",
.mode = 0200,
@@ -572,6 +634,7 @@ static struct attribute *pblk_attrs[] = {
&sys_stats_ppaf_attr,
&sys_lines_attr,
&sys_lines_info_attr,
+ &sys_line_state_attr,
&sys_write_amp_mileage,
&sys_write_amp_trip,
&sys_padding_dist,
@@ -602,6 +665,8 @@ static ssize_t pblk_sysfs_show(struct kobject *kobj, struct attribute *attr,
return pblk_sysfs_lines(pblk, buf);
else if (strcmp(attr->name, "lines_info") == 0)
return pblk_sysfs_lines_info(pblk, buf);
+ else if (strcmp(attr->name, "line_state") == 0)
+ return pblk_sysfs_line_state_show(pblk, buf);
else if (strcmp(attr->name, "max_sec_per_write") == 0)
return pblk_sysfs_get_sec_per_write(pblk, buf);
else if (strcmp(attr->name, "write_amp_mileage") == 0)
@@ -628,6 +693,8 @@ static ssize_t pblk_sysfs_store(struct kobject *kobj, struct attribute *attr,
return pblk_sysfs_set_sec_per_write(pblk, buf, len);
else if (strcmp(attr->name, "write_amp_trip") == 0)
return pblk_sysfs_set_write_amp_trip(pblk, buf, len);
+ else if (strcmp(attr->name, "line_state") == 0)
+ return pblk_sysfs_line_state_store(pblk, buf, len);
else if (strcmp(attr->name, "padding_dist") == 0)
return pblk_sysfs_set_padding_dist(pblk, buf, len);
return 0;
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index 6e1fcd1a538a..bc31c67b725f 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -201,6 +201,8 @@ struct pblk_rb {
struct pblk_lun {
struct ppa_addr bppa;
+ struct ppa_addr chunk_bppa;
+
struct semaphore wr_sem;
};
@@ -297,6 +299,7 @@ enum {
PBLK_LINETYPE_DATA = 2,
/* Line state */
+ PBLK_LINESTATE_NEW = 9,
PBLK_LINESTATE_FREE = 10,
PBLK_LINESTATE_OPEN = 11,
PBLK_LINESTATE_CLOSED = 12,
@@ -412,6 +415,15 @@ struct pblk_smeta {
struct line_smeta *buf; /* smeta buffer in persistent format */
};
+struct pblk_chunk {
+ int state;
+ int type;
+ int wi;
+ u64 slba;
+ u64 cnlb;
+ u64 wp;
+};
+
struct pblk_line {
struct pblk *pblk;
unsigned int id; /* Line number corresponds to the
@@ -426,6 +438,8 @@ struct pblk_line {
unsigned long *lun_bitmap; /* Bitmap for LUNs mapped in line */
+ struct pblk_chunk *chks; /* Chunks forming line */
+
struct pblk_smeta *smeta; /* Start metadata */
struct pblk_emeta *emeta; /* End medatada */
@@ -513,6 +527,8 @@ struct pblk_line_mgmt {
unsigned long d_seq_nr; /* Data line unique sequence number */
unsigned long l_seq_nr; /* Log line unique sequence number */
+ atomic_t sysfs_line_state; /* Line being monitored in sysfs */
+
spinlock_t free_lock;
spinlock_t close_lock;
spinlock_t gc_lock;
@@ -729,6 +745,10 @@ void pblk_set_sec_per_write(struct pblk *pblk, int sec_per_write);
int pblk_setup_w_rec_rq(struct pblk *pblk, struct nvm_rq *rqd,
struct pblk_c_ctx *c_ctx);
void pblk_discard(struct pblk *pblk, struct bio *bio);
+struct nvm_chk_meta *pblk_chunk_get_info(struct pblk *pblk);
+struct nvm_chk_meta *pblk_chunk_get_off(struct pblk *pblk,
+ struct nvm_chk_meta *lp,
+ struct ppa_addr ppa);
void pblk_log_write_err(struct pblk *pblk, struct nvm_rq *rqd);
void pblk_log_read_err(struct pblk *pblk, struct nvm_rq *rqd);
int pblk_submit_io(struct pblk *pblk, struct nvm_rq *rqd);
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 18/19] lightnvm: pblk: refactor init/exit sequences
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (16 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 17/19] lightnvm: pblk: implement get log report chunk Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 13:17 ` [PATCH 19/19] lightnvm: pblk: implement 2.0 support Javier González
2018-02-26 18:19 ` [PATCH V3 00/19] " Matias Bjørling
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Refactor init and exit sequences to improve readability. In the way, fix
bad free ordering on the init error path.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-init.c | 535 +++++++++++++++++++++----------------------
1 file changed, 267 insertions(+), 268 deletions(-)
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index cf4f49d48aed..7f88fb937440 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -103,7 +103,40 @@ static void pblk_l2p_free(struct pblk *pblk)
vfree(pblk->trans_map);
}
-static int pblk_l2p_init(struct pblk *pblk)
+static int pblk_l2p_recover(struct pblk *pblk, bool factory_init)
+{
+ struct pblk_line *line = NULL;
+
+ if (factory_init) {
+ pblk_setup_uuid(pblk);
+ } else {
+ line = pblk_recov_l2p(pblk);
+ if (IS_ERR(line)) {
+ pr_err("pblk: could not recover l2p table\n");
+ return -EFAULT;
+ }
+ }
+
+#ifdef CONFIG_NVM_DEBUG
+ pr_info("pblk init: L2P CRC: %x\n", pblk_l2p_crc(pblk));
+#endif
+
+ /* Free full lines directly as GC has not been started yet */
+ pblk_gc_free_full_lines(pblk);
+
+ if (!line) {
+ /* Configure next line for user data */
+ line = pblk_line_get_first_data(pblk);
+ if (!line) {
+ pr_err("pblk: line list corrupted\n");
+ return -EFAULT;
+ }
+ }
+
+ return 0;
+}
+
+static int pblk_l2p_init(struct pblk *pblk, bool factory_init)
{
sector_t i;
struct ppa_addr ppa;
@@ -119,7 +152,7 @@ static int pblk_l2p_init(struct pblk *pblk)
for (i = 0; i < pblk->rl.nr_secs; i++)
pblk_trans_map_set(pblk, i, ppa);
- return 0;
+ return pblk_l2p_recover(pblk, factory_init);
}
static void pblk_rwb_free(struct pblk *pblk)
@@ -268,86 +301,113 @@ static int pblk_core_init(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
+ int max_write_ppas;
+
+ atomic64_set(&pblk->user_wa, 0);
+ atomic64_set(&pblk->pad_wa, 0);
+ atomic64_set(&pblk->gc_wa, 0);
+ pblk->user_rst_wa = 0;
+ pblk->pad_rst_wa = 0;
+ pblk->gc_rst_wa = 0;
+
+ atomic_long_set(&pblk->nr_flush, 0);
+ pblk->nr_flush_rst = 0;
pblk->pgs_in_buffer = geo->c.mw_cunits * geo->c.ws_opt * geo->all_luns;
+ pblk->min_write_pgs = geo->c.ws_opt * (geo->c.csecs / PAGE_SIZE);
+ max_write_ppas = pblk->min_write_pgs * geo->all_luns;
+ pblk->max_write_pgs = (max_write_ppas < NVM_MAX_VLBA) ?
+ max_write_ppas : NVM_MAX_VLBA;
+ pblk_set_sec_per_write(pblk, pblk->min_write_pgs);
+
+ if (pblk->max_write_pgs > PBLK_MAX_REQ_ADDRS) {
+ pr_err("pblk: cannot support device max_phys_sect\n");
+ return -EINVAL;
+ }
+
+ pblk->pad_dist = kzalloc((pblk->min_write_pgs - 1) * sizeof(atomic64_t),
+ GFP_KERNEL);
+ if (!pblk->pad_dist)
+ return -ENOMEM;
+
if (pblk_init_global_caches(pblk))
- return -ENOMEM;
+ goto fail_free_pad_dist;
/* Internal bios can be at most the sectors signaled by the device. */
pblk->page_bio_pool = mempool_create_page_pool(NVM_MAX_VLBA, 0);
if (!pblk->page_bio_pool)
- goto free_global_caches;
+ goto fail_free_global_caches;
pblk->gen_ws_pool = mempool_create_slab_pool(PBLK_GEN_WS_POOL_SIZE,
pblk_ws_cache);
if (!pblk->gen_ws_pool)
- goto free_page_bio_pool;
+ goto fail_free_page_bio_pool;
pblk->rec_pool = mempool_create_slab_pool(geo->all_luns,
pblk_rec_cache);
if (!pblk->rec_pool)
- goto free_gen_ws_pool;
+ goto fail_free_gen_ws_pool;
pblk->r_rq_pool = mempool_create_slab_pool(geo->all_luns,
pblk_g_rq_cache);
if (!pblk->r_rq_pool)
- goto free_rec_pool;
+ goto fail_free_rec_pool;
pblk->e_rq_pool = mempool_create_slab_pool(geo->all_luns,
pblk_g_rq_cache);
if (!pblk->e_rq_pool)
- goto free_r_rq_pool;
+ goto fail_free_r_rq_pool;
pblk->w_rq_pool = mempool_create_slab_pool(geo->all_luns,
pblk_w_rq_cache);
if (!pblk->w_rq_pool)
- goto free_e_rq_pool;
+ goto fail_free_e_rq_pool;
pblk->close_wq = alloc_workqueue("pblk-close-wq",
WQ_MEM_RECLAIM | WQ_UNBOUND, PBLK_NR_CLOSE_JOBS);
if (!pblk->close_wq)
- goto free_w_rq_pool;
+ goto fail_free_w_rq_pool;
pblk->bb_wq = alloc_workqueue("pblk-bb-wq",
WQ_MEM_RECLAIM | WQ_UNBOUND, 0);
if (!pblk->bb_wq)
- goto free_close_wq;
+ goto fail_free_close_wq;
pblk->r_end_wq = alloc_workqueue("pblk-read-end-wq",
WQ_MEM_RECLAIM | WQ_UNBOUND, 0);
if (!pblk->r_end_wq)
- goto free_bb_wq;
+ goto fail_free_bb_wq;
if (pblk_set_addrf(pblk))
- goto free_r_end_wq;
-
- if (pblk_rwb_init(pblk))
- goto free_r_end_wq;
+ goto fail_free_r_end_wq;
INIT_LIST_HEAD(&pblk->compl_list);
+
return 0;
-free_r_end_wq:
+fail_free_r_end_wq:
destroy_workqueue(pblk->r_end_wq);
-free_bb_wq:
+fail_free_bb_wq:
destroy_workqueue(pblk->bb_wq);
-free_close_wq:
+fail_free_close_wq:
destroy_workqueue(pblk->close_wq);
-free_w_rq_pool:
+fail_free_w_rq_pool:
mempool_destroy(pblk->w_rq_pool);
-free_e_rq_pool:
+fail_free_e_rq_pool:
mempool_destroy(pblk->e_rq_pool);
-free_r_rq_pool:
+fail_free_r_rq_pool:
mempool_destroy(pblk->r_rq_pool);
-free_rec_pool:
+fail_free_rec_pool:
mempool_destroy(pblk->rec_pool);
-free_gen_ws_pool:
+fail_free_gen_ws_pool:
mempool_destroy(pblk->gen_ws_pool);
-free_page_bio_pool:
+fail_free_page_bio_pool:
mempool_destroy(pblk->page_bio_pool);
-free_global_caches:
+fail_free_global_caches:
pblk_free_global_caches(pblk);
+fail_free_pad_dist:
+ kfree(pblk->pad_dist);
return -ENOMEM;
}
@@ -369,14 +429,8 @@ static void pblk_core_free(struct pblk *pblk)
mempool_destroy(pblk->e_rq_pool);
mempool_destroy(pblk->w_rq_pool);
- pblk_rwb_free(pblk);
-
pblk_free_global_caches(pblk);
-}
-
-static void pblk_luns_free(struct pblk *pblk)
-{
- kfree(pblk->luns);
+ kfree(pblk->pad_dist);
}
static void pblk_line_mg_free(struct pblk *pblk)
@@ -393,8 +447,6 @@ static void pblk_line_mg_free(struct pblk *pblk)
pblk_mfree(l_mg->eline_meta[i]->buf, l_mg->emeta_alloc_type);
kfree(l_mg->eline_meta[i]);
}
-
- kfree(pblk->lines);
}
static void pblk_line_meta_free(struct pblk_line *line)
@@ -418,6 +470,11 @@ static void pblk_lines_free(struct pblk *pblk)
pblk_line_meta_free(line);
}
spin_unlock(&l_mg->free_lock);
+
+ pblk_line_mg_free(pblk);
+
+ kfree(pblk->luns);
+ kfree(pblk->lines);
}
static int pblk_bb_get_tbl(struct nvm_tgt_dev *dev, struct pblk_lun *rlun,
@@ -481,7 +538,7 @@ static void *pblk_chunk_get_meta(struct pblk *pblk)
return pblk_chunk_get_info(pblk);
}
-static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
+static int pblk_luns_init(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
@@ -494,11 +551,6 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
return -EINVAL;
}
- pblk->luns = kcalloc(geo->all_luns, sizeof(struct pblk_lun),
- GFP_KERNEL);
- if (!pblk->luns)
- return -ENOMEM;
-
for (i = 0; i < geo->all_luns; i++) {
/* Stripe across channels */
int ch = i % geo->num_ch;
@@ -506,8 +558,8 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
int lunid = lun_raw + ch * geo->num_lun;
rlun = &pblk->luns[i];
- rlun->bppa = luns[lunid];
- rlun->chunk_bppa = luns[i];
+ rlun->bppa = dev->luns[lunid];
+ rlun->chunk_bppa = dev->luns[i];
sema_init(&rlun->wr_sem, 1);
}
@@ -515,38 +567,6 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
return 0;
}
-static int pblk_lines_configure(struct pblk *pblk, int flags)
-{
- struct pblk_line *line = NULL;
- int ret = 0;
-
- if (!(flags & NVM_TARGET_FACTORY)) {
- line = pblk_recov_l2p(pblk);
- if (IS_ERR(line)) {
- pr_err("pblk: could not recover l2p table\n");
- ret = -EFAULT;
- }
- }
-
-#ifdef CONFIG_NVM_DEBUG
- pr_info("pblk init: L2P CRC: %x\n", pblk_l2p_crc(pblk));
-#endif
-
- /* Free full lines directly as GC has not been started yet */
- pblk_gc_free_full_lines(pblk);
-
- if (!line) {
- /* Configure next line for user data */
- line = pblk_line_get_first_data(pblk);
- if (!line) {
- pr_err("pblk: line list corrupted\n");
- ret = -EFAULT;
- }
- }
-
- return ret;
-}
-
/* See comment over struct line_emeta definition */
static unsigned int calc_emeta_len(struct pblk *pblk)
{
@@ -612,81 +632,6 @@ static void pblk_set_provision(struct pblk *pblk, long nr_free_blks)
atomic_set(&pblk->rl.free_user_blocks, nr_free_blks);
}
-static int pblk_lines_alloc_metadata(struct pblk *pblk)
-{
- struct pblk_line_mgmt *l_mg = &pblk->l_mg;
- struct pblk_line_meta *lm = &pblk->lm;
- int i;
-
- /* smeta is always small enough to fit on a kmalloc memory allocation,
- * emeta depends on the number of LUNs allocated to the pblk instance
- */
- for (i = 0; i < PBLK_DATA_LINES; i++) {
- l_mg->sline_meta[i] = kmalloc(lm->smeta_len, GFP_KERNEL);
- if (!l_mg->sline_meta[i])
- goto fail_free_smeta;
- }
-
- /* emeta allocates three different buffers for managing metadata with
- * in-memory and in-media layouts
- */
- for (i = 0; i < PBLK_DATA_LINES; i++) {
- struct pblk_emeta *emeta;
-
- emeta = kmalloc(sizeof(struct pblk_emeta), GFP_KERNEL);
- if (!emeta)
- goto fail_free_emeta;
-
- if (lm->emeta_len[0] > KMALLOC_MAX_CACHE_SIZE) {
- l_mg->emeta_alloc_type = PBLK_VMALLOC_META;
-
- emeta->buf = vmalloc(lm->emeta_len[0]);
- if (!emeta->buf) {
- kfree(emeta);
- goto fail_free_emeta;
- }
-
- emeta->nr_entries = lm->emeta_sec[0];
- l_mg->eline_meta[i] = emeta;
- } else {
- l_mg->emeta_alloc_type = PBLK_KMALLOC_META;
-
- emeta->buf = kmalloc(lm->emeta_len[0], GFP_KERNEL);
- if (!emeta->buf) {
- kfree(emeta);
- goto fail_free_emeta;
- }
-
- emeta->nr_entries = lm->emeta_sec[0];
- l_mg->eline_meta[i] = emeta;
- }
- }
-
- l_mg->vsc_list = kcalloc(l_mg->nr_lines, sizeof(__le32), GFP_KERNEL);
- if (!l_mg->vsc_list)
- goto fail_free_emeta;
-
- for (i = 0; i < l_mg->nr_lines; i++)
- l_mg->vsc_list[i] = cpu_to_le32(EMPTY_ENTRY);
-
- return 0;
-
-fail_free_emeta:
- while (--i >= 0) {
- if (l_mg->emeta_alloc_type == PBLK_VMALLOC_META)
- vfree(l_mg->eline_meta[i]->buf);
- else
- kfree(l_mg->eline_meta[i]->buf);
- kfree(l_mg->eline_meta[i]);
- }
-
-fail_free_smeta:
- for (i = 0; i < PBLK_DATA_LINES; i++)
- kfree(l_mg->sline_meta[i]);
-
- return -ENOMEM;
-}
-
static int pblk_setup_line_meta_12(struct pblk *pblk, struct pblk_line *line,
void *chunk_meta)
{
@@ -830,29 +775,13 @@ static int pblk_alloc_line_meta(struct pblk *pblk, struct pblk_line *line)
return 0;
}
-static int pblk_lines_init(struct pblk *pblk)
+static int pblk_line_mg_init(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
struct pblk_line_mgmt *l_mg = &pblk->l_mg;
struct pblk_line_meta *lm = &pblk->lm;
- struct pblk_line *line;
- void *chunk_meta;
- unsigned int smeta_len, emeta_len;
- long nr_free_chks = 0;
- int bb_distance, max_write_ppas;
- int i, ret;
-
- pblk->min_write_pgs = geo->c.ws_opt * (geo->c.csecs / PAGE_SIZE);
- max_write_ppas = pblk->min_write_pgs * geo->all_luns;
- pblk->max_write_pgs = min_t(int, max_write_ppas, NVM_MAX_VLBA);
- pblk_set_sec_per_write(pblk, pblk->min_write_pgs);
-
- if (pblk->max_write_pgs > PBLK_MAX_REQ_ADDRS) {
- pr_err("pblk: vector list too big(%u > %u)\n",
- pblk->max_write_pgs, PBLK_MAX_REQ_ADDRS);
- return -EINVAL;
- }
+ int i, bb_distance;
l_mg->nr_lines = geo->c.num_chk;
l_mg->log_line = l_mg->data_line = NULL;
@@ -861,6 +790,119 @@ static int pblk_lines_init(struct pblk *pblk)
atomic_set(&l_mg->sysfs_line_state, -1);
bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
+ INIT_LIST_HEAD(&l_mg->free_list);
+ INIT_LIST_HEAD(&l_mg->corrupt_list);
+ INIT_LIST_HEAD(&l_mg->bad_list);
+ INIT_LIST_HEAD(&l_mg->gc_full_list);
+ INIT_LIST_HEAD(&l_mg->gc_high_list);
+ INIT_LIST_HEAD(&l_mg->gc_mid_list);
+ INIT_LIST_HEAD(&l_mg->gc_low_list);
+ INIT_LIST_HEAD(&l_mg->gc_empty_list);
+
+ INIT_LIST_HEAD(&l_mg->emeta_list);
+
+ l_mg->gc_lists[0] = &l_mg->gc_high_list;
+ l_mg->gc_lists[1] = &l_mg->gc_mid_list;
+ l_mg->gc_lists[2] = &l_mg->gc_low_list;
+
+ spin_lock_init(&l_mg->free_lock);
+ spin_lock_init(&l_mg->close_lock);
+ spin_lock_init(&l_mg->gc_lock);
+
+ l_mg->vsc_list = kcalloc(l_mg->nr_lines, sizeof(__le32), GFP_KERNEL);
+ if (!l_mg->vsc_list)
+ goto fail;
+
+ l_mg->bb_template = kzalloc(lm->sec_bitmap_len, GFP_KERNEL);
+ if (!l_mg->bb_template)
+ goto fail_free_vsc_list;
+
+ l_mg->bb_aux = kzalloc(lm->sec_bitmap_len, GFP_KERNEL);
+ if (!l_mg->bb_aux)
+ goto fail_free_bb_template;
+
+ /* smeta is always small enough to fit on a kmalloc memory allocation,
+ * emeta depends on the number of LUNs allocated to the pblk instance
+ */
+ for (i = 0; i < PBLK_DATA_LINES; i++) {
+ l_mg->sline_meta[i] = kmalloc(lm->smeta_len, GFP_KERNEL);
+ if (!l_mg->sline_meta[i])
+ goto fail_free_smeta;
+ }
+
+ /* emeta allocates three different buffers for managing metadata with
+ * in-memory and in-media layouts
+ */
+ for (i = 0; i < PBLK_DATA_LINES; i++) {
+ struct pblk_emeta *emeta;
+
+ emeta = kmalloc(sizeof(struct pblk_emeta), GFP_KERNEL);
+ if (!emeta)
+ goto fail_free_emeta;
+
+ if (lm->emeta_len[0] > KMALLOC_MAX_CACHE_SIZE) {
+ l_mg->emeta_alloc_type = PBLK_VMALLOC_META;
+
+ emeta->buf = vmalloc(lm->emeta_len[0]);
+ if (!emeta->buf) {
+ kfree(emeta);
+ goto fail_free_emeta;
+ }
+
+ emeta->nr_entries = lm->emeta_sec[0];
+ l_mg->eline_meta[i] = emeta;
+ } else {
+ l_mg->emeta_alloc_type = PBLK_KMALLOC_META;
+
+ emeta->buf = kmalloc(lm->emeta_len[0], GFP_KERNEL);
+ if (!emeta->buf) {
+ kfree(emeta);
+ goto fail_free_emeta;
+ }
+
+ emeta->nr_entries = lm->emeta_sec[0];
+ l_mg->eline_meta[i] = emeta;
+ }
+ }
+
+ for (i = 0; i < l_mg->nr_lines; i++)
+ l_mg->vsc_list[i] = cpu_to_le32(EMPTY_ENTRY);
+
+ bb_distance = (geo->all_luns) * geo->c.ws_opt;
+ for (i = 0; i < lm->sec_per_line; i += bb_distance)
+ bitmap_set(l_mg->bb_template, i, geo->c.ws_opt);
+
+ return 0;
+
+fail_free_emeta:
+ while (--i >= 0) {
+ if (l_mg->emeta_alloc_type == PBLK_VMALLOC_META)
+ vfree(l_mg->eline_meta[i]->buf);
+ else
+ kfree(l_mg->eline_meta[i]->buf);
+ kfree(l_mg->eline_meta[i]);
+ }
+fail_free_smeta:
+ kfree(l_mg->bb_aux);
+
+ for (i = 0; i < PBLK_DATA_LINES; i++)
+ kfree(l_mg->sline_meta[i]);
+fail_free_bb_template:
+ kfree(l_mg->bb_template);
+fail_free_vsc_list:
+ kfree(l_mg->vsc_list);
+fail:
+ return -ENOMEM;
+}
+
+static int pblk_line_meta_init(struct pblk *pblk)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ struct pblk_line_meta *lm = &pblk->lm;
+ unsigned int smeta_len, emeta_len;
+ int i;
+
lm->sec_per_line = geo->c.clba * geo->all_luns;
lm->blk_per_line = geo->all_luns;
lm->blk_bitmap_len = BITS_TO_LONGS(geo->all_luns) * sizeof(long);
@@ -911,58 +953,49 @@ static int pblk_lines_init(struct pblk *pblk)
return -EINVAL;
}
- ret = pblk_lines_alloc_metadata(pblk);
+ return 0;
+}
+
+static int pblk_lines_init(struct pblk *pblk)
+{
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+ struct pblk_line_mgmt *l_mg = &pblk->l_mg;
+ struct pblk_line *line;
+ void *chunk_meta;
+ long nr_free_chks = 0;
+ int i, ret;
+
+ ret = pblk_line_meta_init(pblk);
if (ret)
return ret;
- l_mg->bb_template = kzalloc(lm->sec_bitmap_len, GFP_KERNEL);
- if (!l_mg->bb_template) {
- ret = -ENOMEM;
- goto fail_free_meta;
- }
-
- l_mg->bb_aux = kzalloc(lm->sec_bitmap_len, GFP_KERNEL);
- if (!l_mg->bb_aux) {
- ret = -ENOMEM;
- goto fail_free_bb_template;
- }
-
- bb_distance = (geo->all_luns) * geo->c.ws_opt;
- for (i = 0; i < lm->sec_per_line; i += bb_distance)
- bitmap_set(l_mg->bb_template, i, geo->c.ws_opt);
-
- INIT_LIST_HEAD(&l_mg->free_list);
- INIT_LIST_HEAD(&l_mg->corrupt_list);
- INIT_LIST_HEAD(&l_mg->bad_list);
- INIT_LIST_HEAD(&l_mg->gc_full_list);
- INIT_LIST_HEAD(&l_mg->gc_high_list);
- INIT_LIST_HEAD(&l_mg->gc_mid_list);
- INIT_LIST_HEAD(&l_mg->gc_low_list);
- INIT_LIST_HEAD(&l_mg->gc_empty_list);
-
- INIT_LIST_HEAD(&l_mg->emeta_list);
-
- l_mg->gc_lists[0] = &l_mg->gc_high_list;
- l_mg->gc_lists[1] = &l_mg->gc_mid_list;
- l_mg->gc_lists[2] = &l_mg->gc_low_list;
-
- spin_lock_init(&l_mg->free_lock);
- spin_lock_init(&l_mg->close_lock);
- spin_lock_init(&l_mg->gc_lock);
+ ret = pblk_line_mg_init(pblk);
+ if (ret)
+ return ret;
- pblk->lines = kcalloc(l_mg->nr_lines, sizeof(struct pblk_line),
+ pblk->luns = kcalloc(geo->all_luns, sizeof(struct pblk_lun),
GFP_KERNEL);
- if (!pblk->lines) {
- ret = -ENOMEM;
- goto fail_free_bb_aux;
- }
+ if (!pblk->luns)
+ return -ENOMEM;
+
+ ret = pblk_luns_init(pblk);
+ if (ret)
+ goto fail_free_luns;
chunk_meta = pblk_chunk_get_meta(pblk);
if (IS_ERR(chunk_meta)) {
pr_err("pblk: could not get chunk log (%lu)\n",
PTR_ERR(chunk_meta));
ret = PTR_ERR(chunk_meta);
- goto fail_free_lines;
+ goto fail_free_meta;
+ }
+
+ pblk->lines = kcalloc(l_mg->nr_lines, sizeof(struct pblk_line),
+ GFP_KERNEL);
+ if (!pblk->lines) {
+ ret = -ENOMEM;
+ goto fail_free_chunk_meta;
}
for (i = 0; i < l_mg->nr_lines; i++) {
@@ -970,7 +1003,7 @@ static int pblk_lines_init(struct pblk *pblk)
ret = pblk_alloc_line_meta(pblk, line);
if (ret)
- goto fail_free_chunk_meta;
+ goto fail_free_lines;
nr_free_chks += pblk_setup_line_meta(pblk, line, chunk_meta, i);
}
@@ -980,18 +1013,16 @@ static int pblk_lines_init(struct pblk *pblk)
kfree(chunk_meta);
return 0;
-fail_free_chunk_meta:
- kfree(chunk_meta);
+fail_free_lines:
while (--i >= 0)
pblk_line_meta_free(&pblk->lines[i]);
-fail_free_lines:
kfree(pblk->lines);
-fail_free_bb_aux:
- kfree(l_mg->bb_aux);
-fail_free_bb_template:
- kfree(l_mg->bb_template);
+fail_free_chunk_meta:
+ kfree(chunk_meta);
fail_free_meta:
pblk_line_mg_free(pblk);
+fail_free_luns:
+ kfree(pblk->luns);
return ret;
}
@@ -1032,12 +1063,10 @@ static void pblk_writer_stop(struct pblk *pblk)
static void pblk_free(struct pblk *pblk)
{
- pblk_luns_free(pblk);
pblk_lines_free(pblk);
- kfree(pblk->pad_dist);
- pblk_line_mg_free(pblk);
- pblk_core_free(pblk);
pblk_l2p_free(pblk);
+ pblk_rwb_free(pblk);
+ pblk_core_free(pblk);
kfree(pblk);
}
@@ -1108,19 +1137,6 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
spin_lock_init(&pblk->trans_lock);
spin_lock_init(&pblk->lock);
- if (flags & NVM_TARGET_FACTORY)
- pblk_setup_uuid(pblk);
-
- atomic64_set(&pblk->user_wa, 0);
- atomic64_set(&pblk->pad_wa, 0);
- atomic64_set(&pblk->gc_wa, 0);
- pblk->user_rst_wa = 0;
- pblk->pad_rst_wa = 0;
- pblk->gc_rst_wa = 0;
-
- atomic_long_set(&pblk->nr_flush, 0);
- pblk->nr_flush_rst = 0;
-
#ifdef CONFIG_NVM_DEBUG
atomic_long_set(&pblk->inflight_writes, 0);
atomic_long_set(&pblk->padded_writes, 0);
@@ -1144,48 +1160,35 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
atomic_long_set(&pblk->write_failed, 0);
atomic_long_set(&pblk->erase_failed, 0);
- ret = pblk_luns_init(pblk, dev->luns);
- if (ret) {
- pr_err("pblk: could not initialize luns\n");
- goto fail;
- }
-
- ret = pblk_lines_init(pblk);
- if (ret) {
- pr_err("pblk: could not initialize lines\n");
- goto fail_free_luns;
- }
-
- pblk->pad_dist = kzalloc((pblk->min_write_pgs - 1) * sizeof(atomic64_t),
- GFP_KERNEL);
- if (!pblk->pad_dist) {
- ret = -ENOMEM;
- goto fail_free_line_meta;
- }
-
ret = pblk_core_init(pblk);
if (ret) {
pr_err("pblk: could not initialize core\n");
- goto fail_free_pad_dist;
+ goto fail;
}
- ret = pblk_l2p_init(pblk);
+ ret = pblk_lines_init(pblk);
if (ret) {
- pr_err("pblk: could not initialize maps\n");
+ pr_err("pblk: could not initialize lines\n");
goto fail_free_core;
}
- ret = pblk_lines_configure(pblk, flags);
+ ret = pblk_rwb_init(pblk);
if (ret) {
- pr_err("pblk: could not configure lines\n");
- goto fail_free_l2p;
+ pr_err("pblk: could not initialize write buffer\n");
+ goto fail_free_lines;
+ }
+
+ ret = pblk_l2p_init(pblk, flags & NVM_TARGET_FACTORY);
+ if (ret) {
+ pr_err("pblk: could not initialize maps\n");
+ goto fail_free_rwb;
}
ret = pblk_writer_init(pblk);
if (ret) {
if (ret != -EINTR)
pr_err("pblk: could not initialize write thread\n");
- goto fail_free_lines;
+ goto fail_free_l2p;
}
ret = pblk_gc_init(pblk);
@@ -1220,18 +1223,14 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
fail_stop_writer:
pblk_writer_stop(pblk);
-fail_free_lines:
- pblk_lines_free(pblk);
fail_free_l2p:
pblk_l2p_free(pblk);
+fail_free_rwb:
+ pblk_rwb_free(pblk);
+fail_free_lines:
+ pblk_lines_free(pblk);
fail_free_core:
pblk_core_free(pblk);
-fail_free_pad_dist:
- kfree(pblk->pad_dist);
-fail_free_line_meta:
- pblk_line_mg_free(pblk);
-fail_free_luns:
- pblk_luns_free(pblk);
fail:
kfree(pblk);
return ERR_PTR(ret);
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 19/19] lightnvm: pblk: implement 2.0 support
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (17 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 18/19] lightnvm: pblk: refactor init/exit sequences Javier González
@ 2018-02-26 13:17 ` Javier González
2018-02-26 18:19 ` [PATCH V3 00/19] " Matias Bjørling
19 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-26 13:17 UTC (permalink / raw)
To: mb; +Cc: linux-block, linux-kernel, linux-nvme, Javier González
Implement 2.0 support in pblk. This includes the address formatting and
mapping paths, as well as the sysfs entries for them.
Signed-off-by: Javier González <javier@cnexlabs.com>
---
drivers/lightnvm/pblk-init.c | 57 ++++++++++--
drivers/lightnvm/pblk-sysfs.c | 36 ++++++--
drivers/lightnvm/pblk.h | 198 ++++++++++++++++++++++++++++++++----------
3 files changed, 233 insertions(+), 58 deletions(-)
diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
index 7f88fb937440..fe04996cd3ba 100644
--- a/drivers/lightnvm/pblk-init.c
+++ b/drivers/lightnvm/pblk-init.c
@@ -231,20 +231,63 @@ static int pblk_set_addrf_12(struct nvm_geo *geo,
return dst->blk_offset + src->blk_len;
}
+static int pblk_set_addrf_20(struct nvm_geo *geo,
+ struct nvm_addr_format *adst,
+ struct pblk_addr_format *udst)
+{
+ struct nvm_addr_format *src = &geo->c.addrf;
+
+ adst->ch_len = get_count_order(geo->num_ch);
+ adst->lun_len = get_count_order(geo->num_lun);
+ adst->chk_len = src->chk_len;
+ adst->sec_len = src->sec_len;
+
+ adst->sec_offset = 0;
+ adst->ch_offset = adst->sec_len;
+ adst->lun_offset = adst->ch_offset + adst->ch_len;
+ adst->chk_offset = adst->lun_offset + adst->lun_len;
+
+ adst->sec_mask = ((1ULL << adst->sec_len) - 1) << adst->sec_offset;
+ adst->chk_mask = ((1ULL << adst->chk_len) - 1) << adst->chk_offset;
+ adst->lun_mask = ((1ULL << adst->lun_len) - 1) << adst->lun_offset;
+ adst->ch_mask = ((1ULL << adst->ch_len) - 1) << adst->ch_offset;
+
+ udst->sec_stripe = geo->c.ws_opt;
+ udst->ch_stripe = geo->num_ch;
+ udst->lun_stripe = geo->num_lun;
+
+ udst->sec_lun_stripe = udst->sec_stripe * udst->ch_stripe;
+ udst->sec_ws_stripe = udst->sec_lun_stripe * udst->lun_stripe;
+
+ return adst->chk_offset + adst->chk_len;
+}
+
static int pblk_set_addrf(struct pblk *pblk)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
int mod;
- div_u64_rem(geo->c.clba, pblk->min_write_pgs, &mod);
- if (mod) {
- pr_err("pblk: bad configuration of sectors/pages\n");
+ switch (geo->c.version) {
+ case NVM_OCSSD_SPEC_12:
+ div_u64_rem(geo->c.clba, pblk->min_write_pgs, &mod);
+ if (mod) {
+ pr_err("pblk: bad configuration of sectors/pages\n");
+ return -EINVAL;
+ }
+
+ pblk->addrf_len = pblk_set_addrf_12(geo, (void *)&pblk->addrf);
+ break;
+ case NVM_OCSSD_SPEC_20:
+ pblk->addrf_len = pblk_set_addrf_20(geo, (void *)&pblk->addrf,
+ &pblk->uaddrf);
+ break;
+ default:
+ pr_err("pblk: OCSSD revision not supported (%d)\n",
+ geo->c.version);
return -EINVAL;
}
- pblk->addrf_len = pblk_set_addrf_12(geo, (void *)&pblk->addrf);
-
return 0;
}
@@ -1113,7 +1156,9 @@ static void *pblk_init(struct nvm_tgt_dev *dev, struct gendisk *tdisk,
struct pblk *pblk;
int ret;
- if (geo->c.version != NVM_OCSSD_SPEC_12) {
+ /* pblk supports 1.2 and 2.0 versions */
+ if (!(geo->c.version == NVM_OCSSD_SPEC_12 ||
+ geo->c.version == NVM_OCSSD_SPEC_20)) {
pr_err("pblk: OCSSD version not supported (%u)\n",
geo->c.version);
return ERR_PTR(-EINVAL);
diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
index 1ce5b956c622..42ec00e8d393 100644
--- a/drivers/lightnvm/pblk-sysfs.c
+++ b/drivers/lightnvm/pblk-sysfs.c
@@ -113,15 +113,16 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
{
struct nvm_tgt_dev *dev = pblk->dev;
struct nvm_geo *geo = &dev->geo;
- struct nvm_addr_format_12 *ppaf;
- struct nvm_addr_format_12 *geo_ppaf;
ssize_t sz = 0;
- ppaf = (struct nvm_addr_format_12 *)&pblk->addrf;
- geo_ppaf = (struct nvm_addr_format_12 *)&geo->c.addrf;
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->addrf;
+ struct nvm_addr_format_12 *geo_ppaf =
+ (struct nvm_addr_format_12 *)&geo->c.addrf;
- sz = snprintf(page, PAGE_SIZE,
- "pblk:(s:%d)ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
+ sz = snprintf(page, PAGE_SIZE,
+ "pblk:(s:%d)ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
pblk->addrf_len,
ppaf->ch_offset, ppaf->ch_len,
ppaf->lun_offset, ppaf->lun_len,
@@ -130,14 +131,33 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
ppaf->pln_offset, ppaf->pln_len,
ppaf->sec_offset, ppaf->sec_len);
- sz += snprintf(page + sz, PAGE_SIZE - sz,
- "device:ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
+ sz += snprintf(page + sz, PAGE_SIZE - sz,
+ "device:ch:%d/%d,lun:%d/%d,blk:%d/%d,pg:%d/%d,pl:%d/%d,sec:%d/%d\n",
geo_ppaf->ch_offset, geo_ppaf->ch_len,
geo_ppaf->lun_offset, geo_ppaf->lun_len,
geo_ppaf->blk_offset, geo_ppaf->blk_len,
geo_ppaf->pg_offset, geo_ppaf->pg_len,
geo_ppaf->pln_offset, geo_ppaf->pln_len,
geo_ppaf->sec_offset, geo_ppaf->sec_len);
+ } else {
+ struct nvm_addr_format *ppaf = &pblk->addrf;
+ struct nvm_addr_format *geo_ppaf = &geo->c.addrf;
+
+ sz = snprintf(page, PAGE_SIZE,
+ "pblk:(s:%d)ch:%d/%d,lun:%d/%d,chk:%d/%d/sec:%d/%d\n",
+ pblk->addrf_len,
+ ppaf->ch_offset, ppaf->ch_len,
+ ppaf->lun_offset, ppaf->lun_len,
+ ppaf->chk_offset, ppaf->chk_len,
+ ppaf->sec_offset, ppaf->sec_len);
+
+ sz += snprintf(page + sz, PAGE_SIZE - sz,
+ "device:ch:%d/%d,lun:%d/%d,chk:%d/%d,sec:%d/%d\n",
+ geo_ppaf->ch_offset, geo_ppaf->ch_len,
+ geo_ppaf->lun_offset, geo_ppaf->lun_len,
+ geo_ppaf->chk_offset, geo_ppaf->chk_len,
+ geo_ppaf->sec_offset, geo_ppaf->sec_len);
+ }
return sz;
}
diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
index bc31c67b725f..d8f71a9371b1 100644
--- a/drivers/lightnvm/pblk.h
+++ b/drivers/lightnvm/pblk.h
@@ -574,6 +574,18 @@ enum {
PBLK_STATE_STOPPED = 3,
};
+/* Internal format to support not power-of-2 device formats (for now) */
+struct pblk_addr_format {
+ /* gen to dev */
+ int sec_stripe;
+ int ch_stripe;
+ int lun_stripe;
+
+ /* dev to gen */
+ int sec_lun_stripe;
+ int sec_ws_stripe;
+};
+
struct pblk {
struct nvm_tgt_dev *dev;
struct gendisk *disk;
@@ -586,7 +598,8 @@ struct pblk {
struct pblk_line_mgmt l_mg; /* Line management */
struct pblk_line_meta lm; /* Line metadata */
- struct nvm_addr_format addrf;
+ struct nvm_addr_format addrf; /* Aligned address format */
+ struct pblk_addr_format uaddrf; /* Unaligned address format */
int addrf_len;
struct pblk_rb rwb;
@@ -967,17 +980,43 @@ static inline int pblk_ppa_to_pos(struct nvm_geo *geo, struct ppa_addr p)
static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
u64 line_id)
{
- struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&pblk->addrf;
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
struct ppa_addr ppa;
- ppa.ppa = 0;
- ppa.g.blk = line_id;
- ppa.g.pg = (paddr & ppaf->pg_mask) >> ppaf->pg_offset;
- ppa.g.lun = (paddr & ppaf->lun_mask) >> ppaf->lun_offset;
- ppa.g.ch = (paddr & ppaf->ch_mask) >> ppaf->ch_offset;
- ppa.g.pl = (paddr & ppaf->pln_mask) >> ppaf->pln_offset;
- ppa.g.sec = (paddr & ppaf->sec_mask) >> ppaf->sec_offset;
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->addrf;
+
+ ppa.ppa = 0;
+ ppa.g.blk = line_id;
+ ppa.g.pg = (paddr & ppaf->pg_mask) >> ppaf->pg_offset;
+ ppa.g.lun = (paddr & ppaf->lun_mask) >> ppaf->lun_offset;
+ ppa.g.ch = (paddr & ppaf->ch_mask) >> ppaf->ch_offset;
+ ppa.g.pl = (paddr & ppaf->pln_mask) >> ppaf->pln_offset;
+ ppa.g.sec = (paddr & ppaf->sec_mask) >> ppaf->sec_offset;
+ } else {
+ struct pblk_addr_format *uaddrf = &pblk->uaddrf;
+ int secs, chnls, luns;
+
+ ppa.ppa = 0;
+
+ ppa.m.chk = line_id;
+
+ div_u64_rem(paddr, uaddrf->sec_stripe, &secs);
+ ppa.m.sec = secs;
+
+ sector_div(paddr, uaddrf->sec_stripe);
+ div_u64_rem(paddr, uaddrf->ch_stripe, &chnls);
+ ppa.m.grp = chnls;
+
+ sector_div(paddr, uaddrf->ch_stripe);
+ div_u64_rem(paddr, uaddrf->lun_stripe, &luns);
+ ppa.m.pu = luns;
+
+ sector_div(paddr, uaddrf->lun_stripe);
+ ppa.m.sec += uaddrf->sec_stripe * paddr;
+ }
return ppa;
}
@@ -985,15 +1024,32 @@ static inline struct ppa_addr addr_to_gen_ppa(struct pblk *pblk, u64 paddr,
static inline u64 pblk_dev_ppa_to_line_addr(struct pblk *pblk,
struct ppa_addr p)
{
- struct nvm_addr_format_12 *ppaf =
- (struct nvm_addr_format_12 *)&pblk->addrf;
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
u64 paddr;
- paddr = (u64)p.g.ch << ppaf->ch_offset;
- paddr |= (u64)p.g.lun << ppaf->lun_offset;
- paddr |= (u64)p.g.pg << ppaf->pg_offset;
- paddr |= (u64)p.g.pl << ppaf->pln_offset;
- paddr |= (u64)p.g.sec << ppaf->sec_offset;
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
+ (struct nvm_addr_format_12 *)&pblk->addrf;
+
+ paddr = (u64)p.g.ch << ppaf->ch_offset;
+ paddr |= (u64)p.g.lun << ppaf->lun_offset;
+ paddr |= (u64)p.g.pg << ppaf->pg_offset;
+ paddr |= (u64)p.g.pl << ppaf->pln_offset;
+ paddr |= (u64)p.g.sec << ppaf->sec_offset;
+ } else {
+ struct pblk_addr_format *uaddrf = &pblk->uaddrf;
+ u64 secs = (u64)p.m.sec;
+ int sec_stripe;
+
+ paddr = (u64)p.m.grp * uaddrf->sec_stripe;
+ paddr += (u64)p.m.pu * uaddrf->sec_lun_stripe;
+
+ div_u64_rem(secs, uaddrf->sec_stripe, &sec_stripe);
+ sector_div(secs, uaddrf->sec_stripe);
+ paddr += secs * uaddrf->sec_ws_stripe;
+ paddr += sec_stripe;
+ }
return paddr;
}
@@ -1010,15 +1066,37 @@ static inline struct ppa_addr pblk_ppa32_to_ppa64(struct pblk *pblk, u32 ppa32)
ppa64.c.line = ppa32 & ((~0U) >> 1);
ppa64.c.is_cached = 1;
} else {
- struct nvm_addr_format_12 *ppaf =
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
(struct nvm_addr_format_12 *)&pblk->addrf;
- ppa64.g.ch = (ppa32 & ppaf->ch_mask) >> ppaf->ch_offset;
- ppa64.g.lun = (ppa32 & ppaf->lun_mask) >> ppaf->lun_offset;
- ppa64.g.blk = (ppa32 & ppaf->blk_mask) >> ppaf->blk_offset;
- ppa64.g.pg = (ppa32 & ppaf->pg_mask) >> ppaf->pg_offset;
- ppa64.g.pl = (ppa32 & ppaf->pln_mask) >> ppaf->pln_offset;
- ppa64.g.sec = (ppa32 & ppaf->sec_mask) >> ppaf->sec_offset;
+ ppa64.g.ch = (ppa32 & ppaf->ch_mask) >>
+ ppaf->ch_offset;
+ ppa64.g.lun = (ppa32 & ppaf->lun_mask) >>
+ ppaf->lun_offset;
+ ppa64.g.blk = (ppa32 & ppaf->blk_mask) >>
+ ppaf->blk_offset;
+ ppa64.g.pg = (ppa32 & ppaf->pg_mask) >>
+ ppaf->pg_offset;
+ ppa64.g.pl = (ppa32 & ppaf->pln_mask) >>
+ ppaf->pln_offset;
+ ppa64.g.sec = (ppa32 & ppaf->sec_mask) >>
+ ppaf->sec_offset;
+ } else {
+ struct nvm_addr_format *lbaf = &pblk->addrf;
+
+ ppa64.m.grp = (ppa32 & lbaf->ch_mask) >>
+ lbaf->ch_offset;
+ ppa64.m.pu = (ppa32 & lbaf->lun_mask) >>
+ lbaf->lun_offset;
+ ppa64.m.chk = (ppa32 & lbaf->chk_mask) >>
+ lbaf->chk_offset;
+ ppa64.m.sec = (ppa32 & lbaf->sec_mask) >>
+ lbaf->sec_offset;
+ }
}
return ppa64;
@@ -1034,15 +1112,27 @@ static inline u32 pblk_ppa64_to_ppa32(struct pblk *pblk, struct ppa_addr ppa64)
ppa32 |= ppa64.c.line;
ppa32 |= 1U << 31;
} else {
- struct nvm_addr_format_12 *ppaf =
+ struct nvm_tgt_dev *dev = pblk->dev;
+ struct nvm_geo *geo = &dev->geo;
+
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ struct nvm_addr_format_12 *ppaf =
(struct nvm_addr_format_12 *)&pblk->addrf;
- ppa32 |= ppa64.g.ch << ppaf->ch_offset;
- ppa32 |= ppa64.g.lun << ppaf->lun_offset;
- ppa32 |= ppa64.g.blk << ppaf->blk_offset;
- ppa32 |= ppa64.g.pg << ppaf->pg_offset;
- ppa32 |= ppa64.g.pl << ppaf->pln_offset;
- ppa32 |= ppa64.g.sec << ppaf->sec_offset;
+ ppa32 |= ppa64.g.ch << ppaf->ch_offset;
+ ppa32 |= ppa64.g.lun << ppaf->lun_offset;
+ ppa32 |= ppa64.g.blk << ppaf->blk_offset;
+ ppa32 |= ppa64.g.pg << ppaf->pg_offset;
+ ppa32 |= ppa64.g.pl << ppaf->pln_offset;
+ ppa32 |= ppa64.g.sec << ppaf->sec_offset;
+ } else {
+ struct nvm_addr_format *lbaf = &pblk->addrf;
+
+ ppa32 |= ppa64.m.grp << lbaf->ch_offset;
+ ppa32 |= ppa64.m.pu << lbaf->lun_offset;
+ ppa32 |= ppa64.m.chk << lbaf->chk_offset;
+ ppa32 |= ppa64.m.sec << lbaf->sec_offset;
+ }
}
return ppa32;
@@ -1160,6 +1250,9 @@ static inline int pblk_set_progr_mode(struct pblk *pblk, int type)
struct nvm_geo *geo = &dev->geo;
int flags;
+ if (geo->c.version == NVM_OCSSD_SPEC_20)
+ return 0;
+
flags = geo->c.pln_mode >> 1;
if (type == PBLK_WRITE)
@@ -1179,6 +1272,9 @@ static inline int pblk_set_read_mode(struct pblk *pblk, int type)
struct nvm_geo *geo = &dev->geo;
int flags;
+ if (geo->c.version == NVM_OCSSD_SPEC_20)
+ return 0;
+
flags = NVM_IO_SUSPEND | NVM_IO_SCRAMBLE_ENABLE;
if (type == PBLK_READ_SEQUENTIAL)
flags |= geo->c.pln_mode >> 1;
@@ -1192,16 +1288,21 @@ static inline int pblk_io_aligned(struct pblk *pblk, int nr_secs)
}
#ifdef CONFIG_NVM_DEBUG
-static inline void print_ppa(struct ppa_addr *p, char *msg, int error)
+static inline void print_ppa(struct nvm_geo *geo, struct ppa_addr *p,
+ char *msg, int error)
{
if (p->c.is_cached) {
pr_err("ppa: (%s: %x) cache line: %llu\n",
msg, error, (u64)p->c.line);
- } else {
+ } else if (geo->c.version == NVM_OCSSD_SPEC_12) {
pr_err("ppa: (%s: %x):ch:%d,lun:%d,blk:%d,pg:%d,pl:%d,sec:%d\n",
msg, error,
p->g.ch, p->g.lun, p->g.blk,
p->g.pg, p->g.pl, p->g.sec);
+ } else {
+ pr_err("ppa: (%s: %x):ch:%d,lun:%d,chk:%d,sec:%d\n",
+ msg, error,
+ p->m.grp, p->m.pu, p->m.chk, p->m.sec);
}
}
@@ -1211,13 +1312,13 @@ static inline void pblk_print_failed_rqd(struct pblk *pblk, struct nvm_rq *rqd,
int bit = -1;
if (rqd->nr_ppas == 1) {
- print_ppa(&rqd->ppa_addr, "rqd", error);
+ print_ppa(&pblk->dev->geo, &rqd->ppa_addr, "rqd", error);
return;
}
while ((bit = find_next_bit((void *)&rqd->ppa_status, rqd->nr_ppas,
bit + 1)) < rqd->nr_ppas) {
- print_ppa(&rqd->ppa_list[bit], "rqd", error);
+ print_ppa(&pblk->dev->geo, &rqd->ppa_list[bit], "rqd", error);
}
pr_err("error:%d, ppa_status:%llx\n", error, rqd->ppa_status);
@@ -1233,16 +1334,25 @@ static inline int pblk_boundary_ppa_checks(struct nvm_tgt_dev *tgt_dev,
for (i = 0; i < nr_ppas; i++) {
ppa = &ppas[i];
- if (!ppa->c.is_cached &&
- ppa->g.ch < geo->num_ch &&
- ppa->g.lun < geo->num_lun &&
- ppa->g.pl < geo->c.num_pln &&
- ppa->g.blk < geo->c.num_chk &&
- ppa->g.pg < geo->c.num_pg &&
- ppa->g.sec < geo->c.ws_min)
- continue;
+ if (geo->c.version == NVM_OCSSD_SPEC_12) {
+ if (!ppa->c.is_cached &&
+ ppa->g.ch < geo->num_ch &&
+ ppa->g.lun < geo->num_lun &&
+ ppa->g.pl < geo->c.num_pln &&
+ ppa->g.blk < geo->c.num_chk &&
+ ppa->g.pg < geo->c.num_pg &&
+ ppa->g.sec < geo->c.ws_min)
+ continue;
+ } else {
+ if (!ppa->c.is_cached &&
+ ppa->m.grp < geo->num_ch &&
+ ppa->m.pu < geo->num_lun &&
+ ppa->m.chk < geo->c.num_chk &&
+ ppa->m.sec < geo->c.clba)
+ continue;
+ }
- print_ppa(ppa, "boundary", i);
+ print_ppa(geo, ppa, "boundary", i);
return 1;
}
--
2.7.4
^ permalink raw reply related [flat|nested] 29+ messages in thread
* Re: [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
` (18 preceding siblings ...)
2018-02-26 13:17 ` [PATCH 19/19] lightnvm: pblk: implement 2.0 support Javier González
@ 2018-02-26 18:19 ` Matias Bjørling
2018-02-26 18:21 ` Javier Gonzalez
19 siblings, 1 reply; 29+ messages in thread
From: Matias Bjørling @ 2018-02-26 18:19 UTC (permalink / raw)
To: Javier González
Cc: linux-block, linux-kernel, linux-nvme, Javier González
On 02/26/2018 02:16 PM, Javier González wrote:
> # Changes since V2:
>
> Apply Matias' feedback:
> - Remove generic nvm_id identify structure.
> - Do not remap capabilities (cap) to media and controlled capabilities
> (mccap). Instead, add a comment to prevent confusion when
> crosschecking with 2.0 spec.
> - Change maxoc and maxocpu defaults from 1 block to the max number of
> blocks.
> - Re-implement the generic geometry to use nvm_geo on both device and
> targets. Maintain nvm_common_geo to make it easier to copy the common
> part of the geometry (without having to overwrite target-specific
> fields, which is ugly and error prone). Matias, if you still want to
> get rid of this, we can do it.
I do, the variables should go directly in nvm_geo. Thanks.
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support
2018-02-26 18:19 ` [PATCH V3 00/19] " Matias Bjørling
@ 2018-02-26 18:21 ` Javier Gonzalez
2018-02-26 18:24 ` Matias Bjørling
0 siblings, 1 reply; 29+ messages in thread
From: Javier Gonzalez @ 2018-02-26 18:21 UTC (permalink / raw)
To: Matias Bjørling; +Cc: linux-block, linux-kernel, linux-nvme
[-- Attachment #1: Type: text/plain, Size: 940 bytes --]
> On 26 Feb 2018, at 19.19, Matias Bjørling <mb@lightnvm.io> wrote:
>
> On 02/26/2018 02:16 PM, Javier González wrote:
>> # Changes since V2:
>> Apply Matias' feedback:
>> - Remove generic nvm_id identify structure.
>> - Do not remap capabilities (cap) to media and controlled capabilities
>> (mccap). Instead, add a comment to prevent confusion when
>> crosschecking with 2.0 spec.
>> - Change maxoc and maxocpu defaults from 1 block to the max number of
>> blocks.
>> - Re-implement the generic geometry to use nvm_geo on both device and
>> targets. Maintain nvm_common_geo to make it easier to copy the common
>> part of the geometry (without having to overwrite target-specific
>> fields, which is ugly and error prone). Matias, if you still want to
>> get rid of this, we can do it.
>
> I do, the variables should go directly in nvm_geo. Thanks.
Ok. Is the rest ok with you?
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support
2018-02-26 18:21 ` Javier Gonzalez
@ 2018-02-26 18:24 ` Matias Bjørling
2018-02-26 18:27 ` Javier Gonzalez
0 siblings, 1 reply; 29+ messages in thread
From: Matias Bjørling @ 2018-02-26 18:24 UTC (permalink / raw)
To: Javier Gonzalez; +Cc: linux-block, linux-kernel, linux-nvme
On 02/26/2018 07:21 PM, Javier Gonzalez wrote:
>> On 26 Feb 2018, at 19.19, Matias Bjørling <mb@lightnvm.io> wrote:
>>
>> On 02/26/2018 02:16 PM, Javier González wrote:
>>> # Changes since V2:
>>> Apply Matias' feedback:
>>> - Remove generic nvm_id identify structure.
>>> - Do not remap capabilities (cap) to media and controlled capabilities
>>> (mccap). Instead, add a comment to prevent confusion when
>>> crosschecking with 2.0 spec.
>>> - Change maxoc and maxocpu defaults from 1 block to the max number of
>>> blocks.
>>> - Re-implement the generic geometry to use nvm_geo on both device and
>>> targets. Maintain nvm_common_geo to make it easier to copy the common
>>> part of the geometry (without having to overwrite target-specific
>>> fields, which is ugly and error prone). Matias, if you still want to
>>> get rid of this, we can do it.
>>
>> I do, the variables should go directly in nvm_geo. Thanks.
>
> Ok. Is the rest ok with you?
>
I'll go through it when the rebase is posted. Most of the patches is
dependent on the first patch.
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support
2018-02-26 18:24 ` Matias Bjørling
@ 2018-02-26 18:27 ` Javier Gonzalez
2018-02-26 18:33 ` Matias Bjørling
0 siblings, 1 reply; 29+ messages in thread
From: Javier Gonzalez @ 2018-02-26 18:27 UTC (permalink / raw)
To: Matias Bjørling; +Cc: linux-block, linux-kernel, linux-nvme
[-- Attachment #1: Type: text/plain, Size: 1395 bytes --]
> On 26 Feb 2018, at 19.24, Matias Bjørling <mb@lightnvm.io> wrote:
>
> On 02/26/2018 07:21 PM, Javier Gonzalez wrote:
>>> On 26 Feb 2018, at 19.19, Matias Bjørling <mb@lightnvm.io> wrote:
>>>
>>> On 02/26/2018 02:16 PM, Javier González wrote:
>>>> # Changes since V2:
>>>> Apply Matias' feedback:
>>>> - Remove generic nvm_id identify structure.
>>>> - Do not remap capabilities (cap) to media and controlled capabilities
>>>> (mccap). Instead, add a comment to prevent confusion when
>>>> crosschecking with 2.0 spec.
>>>> - Change maxoc and maxocpu defaults from 1 block to the max number of
>>>> blocks.
>>>> - Re-implement the generic geometry to use nvm_geo on both device and
>>>> targets. Maintain nvm_common_geo to make it easier to copy the common
>>>> part of the geometry (without having to overwrite target-specific
>>>> fields, which is ugly and error prone). Matias, if you still want to
>>>> get rid of this, we can do it.
>>>
>>> I do, the variables should go directly in nvm_geo. Thanks.
>> Ok. Is the rest ok with you?
> I'll go through it when the rebase is posted. Most of the patches is
> dependent on the first patch.
As it is now, it will be basically %s/geo->c.X/geo->X/g. If you can look
at the first patch now, I can put all comments in a single version more,
instead of going through one extra cycle...
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support
2018-02-26 18:27 ` Javier Gonzalez
@ 2018-02-26 18:33 ` Matias Bjørling
0 siblings, 0 replies; 29+ messages in thread
From: Matias Bjørling @ 2018-02-26 18:33 UTC (permalink / raw)
To: Javier Gonzalez; +Cc: linux-block, linux-kernel, linux-nvme
On 02/26/2018 07:27 PM, Javier Gonzalez wrote:
>> On 26 Feb 2018, at 19.24, Matias Bjørling <mb@lightnvm.io> wrote:
>>
>> On 02/26/2018 07:21 PM, Javier Gonzalez wrote:
>>>> On 26 Feb 2018, at 19.19, Matias Bjørling <mb@lightnvm.io> wrote:
>>>>
>>>> On 02/26/2018 02:16 PM, Javier González wrote:
>>>>> # Changes since V2:
>>>>> Apply Matias' feedback:
>>>>> - Remove generic nvm_id identify structure.
>>>>> - Do not remap capabilities (cap) to media and controlled capabilities
>>>>> (mccap). Instead, add a comment to prevent confusion when
>>>>> crosschecking with 2.0 spec.
>>>>> - Change maxoc and maxocpu defaults from 1 block to the max number of
>>>>> blocks.
>>>>> - Re-implement the generic geometry to use nvm_geo on both device and
>>>>> targets. Maintain nvm_common_geo to make it easier to copy the common
>>>>> part of the geometry (without having to overwrite target-specific
>>>>> fields, which is ugly and error prone). Matias, if you still want to
>>>>> get rid of this, we can do it.
>>>>
>>>> I do, the variables should go directly in nvm_geo. Thanks.
>>> Ok. Is the rest ok with you?
>> I'll go through it when the rebase is posted. Most of the patches is
>> dependent on the first patch.
>
> As it is now, it will be basically %s/geo->c.X/geo->X/g. If you can look
> at the first patch now, I can put all comments in a single version more,
> instead of going through one extra cycle...
>
>
Also the variable dev_geo -> geo. I'm good with just getting the updated
version before posting the full patch set again.
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 17/19] lightnvm: pblk: implement get log report chunk
2018-02-26 13:17 ` [PATCH 17/19] lightnvm: pblk: implement get log report chunk Javier González
@ 2018-02-26 19:04 ` Matias Bjørling
2018-02-27 14:40 ` Javier González
0 siblings, 1 reply; 29+ messages in thread
From: Matias Bjørling @ 2018-02-26 19:04 UTC (permalink / raw)
To: Javier González
Cc: linux-block, linux-kernel, linux-nvme, Javier González,
Javier González
On 02/26/2018 02:17 PM, Javier González wrote:
> From: Javier González <javier@javigon.com>
>
> In preparation of pblk supporting 2.0, implement the get log report
> chunk in pblk.
>
> This patch only replicates de bad block functionality as the rest of the
> metadata requires new pblk functionality (e.g., wear-index to implement
> wear-leveling).
"This functionality will come in future patches."
I think the above belongs in the cover letter, not in the patch
description.
>
> Signed-off-by: Javier González <javier@cnexlabs.com>
> ---
> drivers/lightnvm/pblk-core.c | 121 +++++++++++++++++++----
> drivers/lightnvm/pblk-init.c | 218 ++++++++++++++++++++++++++++++------------
> drivers/lightnvm/pblk-sysfs.c | 67 +++++++++++++
> drivers/lightnvm/pblk.h | 20 ++++
> 4 files changed, 346 insertions(+), 80 deletions(-)
>
> diff --git a/drivers/lightnvm/pblk-core.c b/drivers/lightnvm/pblk-core.c
> index 1d7c6313f3d9..ce6a7cfdba66 100644
> --- a/drivers/lightnvm/pblk-core.c
> +++ b/drivers/lightnvm/pblk-core.c
> @@ -44,11 +44,12 @@ static void pblk_line_mark_bb(struct work_struct *work)
> }
>
> static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
> - struct ppa_addr *ppa)
> + struct ppa_addr ppa_addr)
> {
> struct nvm_tgt_dev *dev = pblk->dev;
> struct nvm_geo *geo = &dev->geo;
> - int pos = pblk_ppa_to_pos(geo, *ppa);
> + struct ppa_addr *ppa;
> + int pos = pblk_ppa_to_pos(geo, ppa_addr);
>
> pr_debug("pblk: erase failed: line:%d, pos:%d\n", line->id, pos);
> atomic_long_inc(&pblk->erase_failed);
> @@ -58,6 +59,15 @@ static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
> pr_err("pblk: attempted to erase bb: line:%d, pos:%d\n",
> line->id, pos);
>
> + /* Not necessary to mark bad blocks on 2.0 spec. */
> + if (geo->c.version == NVM_OCSSD_SPEC_20)
> + return;
> +
> + ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
> + if (!ppa)
> + return;
> +
> + *ppa = ppa_addr;
> pblk_gen_run_ws(pblk, NULL, ppa, pblk_line_mark_bb,
> GFP_ATOMIC, pblk->bb_wq);
> }
> @@ -69,16 +79,8 @@ static void __pblk_end_io_erase(struct pblk *pblk, struct nvm_rq *rqd)
> line = &pblk->lines[pblk_ppa_to_line(rqd->ppa_addr)];
> atomic_dec(&line->left_seblks);
>
> - if (rqd->error) {
> - struct ppa_addr *ppa;
> -
> - ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
> - if (!ppa)
> - return;
> -
> - *ppa = rqd->ppa_addr;
> - pblk_mark_bb(pblk, line, ppa);
> - }
> + if (rqd->error)
> + pblk_mark_bb(pblk, line, rqd->ppa_addr);
>
> atomic_dec(&pblk->inflight_io);
> }
> @@ -92,6 +94,50 @@ static void pblk_end_io_erase(struct nvm_rq *rqd)
> mempool_free(rqd, pblk->e_rq_pool);
> }
>
> +/*
> + * Get information for all chunks from the device.
> + *
> + * The caller is responsible for freeing the returned structure
> + */
> +struct nvm_chk_meta *pblk_chunk_get_info(struct pblk *pblk)
> +{
> + struct nvm_tgt_dev *dev = pblk->dev;
> + struct nvm_geo *geo = &dev->geo;
> + struct nvm_chk_meta *meta;
> + struct ppa_addr ppa;
> + unsigned long len;
> + int ret;
> +
> + ppa.ppa = 0;
> +
> + len = geo->all_chunks * sizeof(*meta);
> + meta = kzalloc(len, GFP_KERNEL);
> + if (!meta)
> + return ERR_PTR(-ENOMEM);
> +
> + ret = nvm_get_chunk_meta(dev, meta, ppa, geo->all_chunks);
> + if (ret) {
> + pr_err("pblk: could not get chunk metadata (%d)\n", ret);
> + kfree(meta);
> + return ERR_PTR(-EIO);
> + }
> +
> + return meta;
> +}
> +
> +struct nvm_chk_meta *pblk_chunk_get_off(struct pblk *pblk,
> + struct nvm_chk_meta *meta,
> + struct ppa_addr ppa)
> +{
> + struct nvm_tgt_dev *dev = pblk->dev;
> + struct nvm_geo *geo = &dev->geo;
> + int ch_off = ppa.m.grp * geo->c.num_chk * geo->num_lun;
> + int lun_off = ppa.m.pu * geo->c.num_chk;
> + int chk_off = ppa.m.chk;
> +
> + return meta + ch_off + lun_off + chk_off;
> +}
> +
> void __pblk_map_invalidate(struct pblk *pblk, struct pblk_line *line,
> u64 paddr)
> {
> @@ -1094,10 +1140,38 @@ static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
> return 1;
> }
>
> +static int pblk_prepare_new_line(struct pblk *pblk, struct pblk_line *line)
> +{
> + struct pblk_line_meta *lm = &pblk->lm;
> + struct nvm_tgt_dev *dev = pblk->dev;
> + struct nvm_geo *geo = &dev->geo;
> + int blk_to_erase = atomic_read(&line->blk_in_line);
> + int i;
> +
> + for (i = 0; i < lm->blk_per_line; i++) {
> + int state = line->chks[i].state;
> + struct pblk_lun *rlun = &pblk->luns[i];
> +
> + /* Free chunks should not be erased */
> + if (state & NVM_CHK_ST_FREE) {
> + set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
> + line->erase_bitmap);
> + blk_to_erase--;
> + line->chks[i].state = NVM_CHK_ST_HOST_USE;
Can you help me understand why you want to use The NVM_CHK_ST_HOST_USE?
Why would I care if the chunk state is HOST_USE? A target instance
should not be able to see states from other chunks it doesn't own. and
in that case, why have a separate state?
> + }
> +
> + WARN_ONCE(state & NVM_CHK_ST_OPEN,
> + "pblk: open chunk in new line: %d\n",
> + line->id);
> + }
> +
> + return blk_to_erase;
> +}
> +
> static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
> {
> struct pblk_line_meta *lm = &pblk->lm;
> - int blk_in_line = atomic_read(&line->blk_in_line);
> + int blk_to_erase;
>
> line->map_bitmap = kzalloc(lm->sec_bitmap_len, GFP_ATOMIC);
> if (!line->map_bitmap)
> @@ -1110,7 +1184,21 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
> return -ENOMEM;
> }
>
> + /* Bad blocks do not need to be erased */
> + bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
> +
> spin_lock(&line->lock);
> +
> + /* If we have not written to this line, we need to mark up free chunks
> + * as already erased
> + */
> + if (line->state == PBLK_LINESTATE_NEW) {
> + blk_to_erase = pblk_prepare_new_line(pblk, line);
> + line->state = PBLK_LINESTATE_FREE;
> + } else {
> + blk_to_erase = atomic_read(&line->blk_in_line);
> + }
> +
> if (line->state != PBLK_LINESTATE_FREE) {
> kfree(line->map_bitmap);
> kfree(line->invalid_bitmap);
> @@ -1122,15 +1210,12 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>
> line->state = PBLK_LINESTATE_OPEN;
>
> - atomic_set(&line->left_eblks, blk_in_line);
> - atomic_set(&line->left_seblks, blk_in_line);
> + atomic_set(&line->left_eblks, blk_to_erase);
> + atomic_set(&line->left_seblks, blk_to_erase);
>
> line->meta_distance = lm->meta_distance;
> spin_unlock(&line->lock);
>
> - /* Bad blocks do not need to be erased */
> - bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
> -
> kref_init(&line->ref);
>
> return 0;
> diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
> index ec39800eea42..cf4f49d48aed 100644
> --- a/drivers/lightnvm/pblk-init.c
> +++ b/drivers/lightnvm/pblk-init.c
> @@ -401,6 +401,7 @@ static void pblk_line_meta_free(struct pblk_line *line)
> {
> kfree(line->blk_bitmap);
> kfree(line->erase_bitmap);
> + kfree(line->chks);
> }
>
> static void pblk_lines_free(struct pblk *pblk)
> @@ -440,54 +441,44 @@ static int pblk_bb_get_tbl(struct nvm_tgt_dev *dev, struct pblk_lun *rlun,
> return 0;
> }
>
> -static void *pblk_bb_get_log(struct pblk *pblk)
> +static void *pblk_bb_get_meta(struct pblk *pblk)
> {
> struct nvm_tgt_dev *dev = pblk->dev;
> struct nvm_geo *geo = &dev->geo;
> - u8 *log;
> + u8 *meta;
> int i, nr_blks, blk_per_lun;
> int ret;
>
> blk_per_lun = geo->c.num_chk * geo->c.pln_mode;
> nr_blks = blk_per_lun * geo->all_luns;
>
> - log = kmalloc(nr_blks, GFP_KERNEL);
> - if (!log)
> + meta = kmalloc(nr_blks, GFP_KERNEL);
> + if (!meta)
> return ERR_PTR(-ENOMEM);
>
> for (i = 0; i < geo->all_luns; i++) {
> struct pblk_lun *rlun = &pblk->luns[i];
> - u8 *log_pos = log + i * blk_per_lun;
> + u8 *meta_pos = meta + i * blk_per_lun;
>
> - ret = pblk_bb_get_tbl(dev, rlun, log_pos, blk_per_lun);
> + ret = pblk_bb_get_tbl(dev, rlun, meta_pos, blk_per_lun);
> if (ret) {
> - kfree(log);
> + kfree(meta);
> return ERR_PTR(-EIO);
> }
> }
>
> - return log;
> + return meta;
> }
>
> -static int pblk_bb_line(struct pblk *pblk, struct pblk_line *line,
> - u8 *bb_log, int blk_per_line)
> +static void *pblk_chunk_get_meta(struct pblk *pblk)
> {
> struct nvm_tgt_dev *dev = pblk->dev;
> struct nvm_geo *geo = &dev->geo;
> - int i, bb_cnt = 0;
>
> - for (i = 0; i < blk_per_line; i++) {
> - struct pblk_lun *rlun = &pblk->luns[i];
> - u8 *lun_bb_log = bb_log + i * blk_per_line;
> -
> - if (lun_bb_log[line->id] == NVM_BLK_T_FREE)
> - continue;
> -
> - set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
> - bb_cnt++;
> - }
> -
> - return bb_cnt;
> + if (geo->c.version == NVM_OCSSD_SPEC_12)
> + return pblk_bb_get_meta(pblk);
> + else
> + return pblk_chunk_get_info(pblk);
This 1.2 or 2.0 would be nice to have inside the lightnvm core. A target
should not care weather it is a 1.2 or 2.0 device.
> }
>
> static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
> @@ -516,6 +507,7 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>
> rlun = &pblk->luns[i];
> rlun->bppa = luns[lunid];
> + rlun->chunk_bppa = luns[i];
>
> sema_init(&rlun->wr_sem, 1);
> }
> @@ -695,8 +687,125 @@ static int pblk_lines_alloc_metadata(struct pblk *pblk)
> return -ENOMEM;
> }
>
> -static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
> - void *chunk_log, long *nr_bad_blks)
> +static int pblk_setup_line_meta_12(struct pblk *pblk, struct pblk_line *line,
> + void *chunk_meta)
> +{
> + struct nvm_tgt_dev *dev = pblk->dev;
> + struct nvm_geo *geo = &dev->geo;
> + struct pblk_line_meta *lm = &pblk->lm;
> + int i, chk_per_lun, nr_bad_chks = 0;
> +
> + chk_per_lun = geo->c.num_chk * geo->c.pln_mode;
> +
> + for (i = 0; i < lm->blk_per_line; i++) {
> + struct pblk_chunk *chunk = &line->chks[i];
> + struct pblk_lun *rlun = &pblk->luns[i];
> + u8 *lun_bb_meta = chunk_meta + i * chk_per_lun;
> +
> + /*
> + * In 1.2 spec. chunk state is not persisted by the device. Thus
> + * some of the values are reset each time pblk is instantiated.
> + */
> + if (lun_bb_meta[line->id] == NVM_BLK_T_FREE)
> + chunk->state = NVM_CHK_ST_HOST_USE;
> + else
> + chunk->state = NVM_CHK_ST_OFFLINE;
> +
> + chunk->type = NVM_CHK_TP_W_SEQ;
> + chunk->wi = 0;
> + chunk->slba = -1;
> + chunk->cnlb = geo->c.clba;
> + chunk->wp = 0;
> +
> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
> + continue;
> +
> + set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
> + nr_bad_chks++;
> + }
> +
> + return nr_bad_chks;
> +}
> +
> +static int pblk_setup_line_meta_20(struct pblk *pblk, struct pblk_line *line,
> + struct nvm_chk_meta *meta)
> +{
> + struct nvm_tgt_dev *dev = pblk->dev;
> + struct nvm_geo *geo = &dev->geo;
> + struct pblk_line_meta *lm = &pblk->lm;
> + int i, nr_bad_chks = 0;
> +
> + for (i = 0; i < lm->blk_per_line; i++) {
> + struct pblk_chunk *chunk = &line->chks[i];
> + struct pblk_lun *rlun = &pblk->luns[i];
> + struct nvm_chk_meta *chunk_meta;
> + struct ppa_addr ppa;
> +
> + ppa = rlun->chunk_bppa;
> + ppa.m.chk = line->id;
> + chunk_meta = pblk_chunk_get_off(pblk, meta, ppa);
> +
> + chunk->state = chunk_meta->state;
> + chunk->type = chunk_meta->type;
> + chunk->wi = chunk_meta->wi;
> + chunk->slba = chunk_meta->slba;
> + chunk->cnlb = chunk_meta->cnlb;
> + chunk->wp = chunk_meta->wp;
> +
> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
> + continue;
> +
> + if (chunk->type & NVM_CHK_TP_SZ_SPEC) {
> + WARN_ONCE(1, "pblk: custom-sized chunks unsupported\n");
> + continue;
> + }
> +
> + set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
> + line->blk_bitmap);
> + nr_bad_chks++;
> + }
> +
> + return nr_bad_chks;
> +}
> +
> +static long pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
> + void *chunk_meta, int line_id)
> +{
> + struct nvm_tgt_dev *dev = pblk->dev;
> + struct nvm_geo *geo = &dev->geo;
> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
> + struct pblk_line_meta *lm = &pblk->lm;
> + long nr_bad_chks, chk_in_line;
> +
> + line->pblk = pblk;
> + line->id = line_id;
> + line->type = PBLK_LINETYPE_FREE;
> + line->state = PBLK_LINESTATE_NEW;
> + line->gc_group = PBLK_LINEGC_NONE;
> + line->vsc = &l_mg->vsc_list[line_id];
> + spin_lock_init(&line->lock);
> +
> + if (geo->c.version == NVM_OCSSD_SPEC_12)
> + nr_bad_chks = pblk_setup_line_meta_12(pblk, line, chunk_meta);
> + else
> + nr_bad_chks = pblk_setup_line_meta_20(pblk, line, chunk_meta);
> +
> + chk_in_line = lm->blk_per_line - nr_bad_chks;
> + if (nr_bad_chks < 0 || nr_bad_chks > lm->blk_per_line ||
> + chk_in_line < lm->min_blk_line) {
> + line->state = PBLK_LINESTATE_BAD;
> + list_add_tail(&line->list, &l_mg->bad_list);
> + return 0;
> + }
> +
> + atomic_set(&line->blk_in_line, chk_in_line);
> + list_add_tail(&line->list, &l_mg->free_list);
> + l_mg->nr_free_lines++;
> +
> + return chk_in_line;
> +}
> +
> +static int pblk_alloc_line_meta(struct pblk *pblk, struct pblk_line *line)
> {
> struct pblk_line_meta *lm = &pblk->lm;
>
> @@ -710,7 +819,13 @@ static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
> return -ENOMEM;
> }
>
> - *nr_bad_blks = pblk_bb_line(pblk, line, chunk_log, lm->blk_per_line);
> + line->chks = kmalloc(lm->blk_per_line * sizeof(struct pblk_chunk),
> + GFP_KERNEL);
> + if (!line->chks) {
> + kfree(line->erase_bitmap);
> + kfree(line->blk_bitmap);
> + return -ENOMEM;
> + }
>
> return 0;
> }
> @@ -722,9 +837,9 @@ static int pblk_lines_init(struct pblk *pblk)
> struct pblk_line_mgmt *l_mg = &pblk->l_mg;
> struct pblk_line_meta *lm = &pblk->lm;
> struct pblk_line *line;
> - void *chunk_log;
> + void *chunk_meta;
> unsigned int smeta_len, emeta_len;
> - long nr_bad_blks = 0, nr_free_blks = 0;
> + long nr_free_chks = 0;
> int bb_distance, max_write_ppas;
> int i, ret;
>
> @@ -743,6 +858,7 @@ static int pblk_lines_init(struct pblk *pblk)
> l_mg->log_line = l_mg->data_line = NULL;
> l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
> l_mg->nr_free_lines = 0;
> + atomic_set(&l_mg->sysfs_line_state, -1);
> bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
>
> lm->sec_per_line = geo->c.clba * geo->all_luns;
> @@ -841,53 +957,31 @@ static int pblk_lines_init(struct pblk *pblk)
> goto fail_free_bb_aux;
> }
>
> - chunk_log = pblk_bb_get_log(pblk);
> - if (IS_ERR(chunk_log)) {
> - pr_err("pblk: could not get bad block log (%lu)\n",
> - PTR_ERR(chunk_log));
> - ret = PTR_ERR(chunk_log);
> + chunk_meta = pblk_chunk_get_meta(pblk);
> + if (IS_ERR(chunk_meta)) {
> + pr_err("pblk: could not get chunk log (%lu)\n",
> + PTR_ERR(chunk_meta));
> + ret = PTR_ERR(chunk_meta);
> goto fail_free_lines;
> }
>
> for (i = 0; i < l_mg->nr_lines; i++) {
> - int chk_in_line;
> -
> line = &pblk->lines[i];
>
> - line->pblk = pblk;
> - line->id = i;
> - line->type = PBLK_LINETYPE_FREE;
> - line->state = PBLK_LINESTATE_FREE;
> - line->gc_group = PBLK_LINEGC_NONE;
> - line->vsc = &l_mg->vsc_list[i];
> - spin_lock_init(&line->lock);
> -
> - ret = pblk_setup_line_meta(pblk, line, chunk_log, &nr_bad_blks);
> + ret = pblk_alloc_line_meta(pblk, line);
> if (ret)
> - goto fail_free_chunk_log;
> + goto fail_free_chunk_meta;
>
> - chk_in_line = lm->blk_per_line - nr_bad_blks;
> - if (nr_bad_blks < 0 || nr_bad_blks > lm->blk_per_line ||
> - chk_in_line < lm->min_blk_line) {
> - line->state = PBLK_LINESTATE_BAD;
> - list_add_tail(&line->list, &l_mg->bad_list);
> - continue;
> - }
> -
> - nr_free_blks += chk_in_line;
> - atomic_set(&line->blk_in_line, chk_in_line);
> -
> - l_mg->nr_free_lines++;
> - list_add_tail(&line->list, &l_mg->free_list);
> + nr_free_chks += pblk_setup_line_meta(pblk, line, chunk_meta, i);
> }
>
> - pblk_set_provision(pblk, nr_free_blks);
> + pblk_set_provision(pblk, nr_free_chks);
>
> - kfree(chunk_log);
> + kfree(chunk_meta);
> return 0;
>
> -fail_free_chunk_log:
> - kfree(chunk_log);
> +fail_free_chunk_meta:
> + kfree(chunk_meta);
> while (--i >= 0)
> pblk_line_meta_free(&pblk->lines[i]);
> fail_free_lines:
> diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
> index ccfb3abd2773..1ce5b956c622 100644
> --- a/drivers/lightnvm/pblk-sysfs.c
> +++ b/drivers/lightnvm/pblk-sysfs.c
> @@ -142,6 +142,40 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
> return sz;
> }
>
> +static ssize_t pblk_sysfs_line_state_show(struct pblk *pblk, char *page)
> +{
> + struct pblk_line_meta *lm = &pblk->lm;
> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
> + struct pblk_line *line;
> + int line_id = atomic_read(&l_mg->sysfs_line_state);
> + ssize_t sz = 0;
> + int i;
> +
> + if (line_id < 0 || line_id >= l_mg->nr_lines)
> + return 0;
> +
> + sz = snprintf(page, PAGE_SIZE,
> + "line\tchunk\tstate\ttype\twear-index\tslba\t\tcnlb\twp\n");
> +
> + line = &pblk->lines[line_id];
> +
> + for (i = 0; i < lm->blk_per_line; i++) {
> + struct pblk_chunk *chunk = &line->chks[i];
> +
> + sz += snprintf(page + sz, PAGE_SIZE - sz,
> + "%d\t%d\t%d\t%d\t%d\t\t%llu\t\t%llu\t%llu\n",
> + line->id, i,
> + chunk->state,
> + chunk->type,
> + chunk->wi,
> + chunk->slba,
> + chunk->cnlb,
> + chunk->wp);
> + }
> +
> + return sz;
> +}
> +
> static ssize_t pblk_sysfs_lines(struct pblk *pblk, char *page)
> {
> struct nvm_tgt_dev *dev = pblk->dev;
> @@ -398,6 +432,29 @@ static ssize_t pblk_sysfs_stats_debug(struct pblk *pblk, char *page)
> }
> #endif
>
> +
> +static ssize_t pblk_sysfs_line_state_store(struct pblk *pblk, const char *page,
> + size_t len)
> +{
> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
> + size_t c_len;
> + int line_id;
> +
> + c_len = strcspn(page, "\n");
> + if (c_len >= len)
> + return -EINVAL;
> +
> + if (kstrtouint(page, 0, &line_id))
> + return -EINVAL;
> +
> + if (line_id < 0 || line_id >= l_mg->nr_lines)
> + return -EINVAL;
> +
> + atomic_set(&l_mg->sysfs_line_state, line_id);
> +
> + return len;
> +}
> +
> static ssize_t pblk_sysfs_gc_force(struct pblk *pblk, const char *page,
> size_t len)
> {
> @@ -529,6 +586,11 @@ static struct attribute sys_lines_info_attr = {
> .mode = 0444,
> };
>
> +static struct attribute sys_line_state_attr = {
> + .name = "line_state",
> + .mode = 0644,
> +};
> +
> static struct attribute sys_gc_force = {
> .name = "gc_force",
> .mode = 0200,
> @@ -572,6 +634,7 @@ static struct attribute *pblk_attrs[] = {
> &sys_stats_ppaf_attr,
> &sys_lines_attr,
> &sys_lines_info_attr,
> + &sys_line_state_attr,
> &sys_write_amp_mileage,
> &sys_write_amp_trip,
> &sys_padding_dist,
> @@ -602,6 +665,8 @@ static ssize_t pblk_sysfs_show(struct kobject *kobj, struct attribute *attr,
> return pblk_sysfs_lines(pblk, buf);
> else if (strcmp(attr->name, "lines_info") == 0)
> return pblk_sysfs_lines_info(pblk, buf);
> + else if (strcmp(attr->name, "line_state") == 0)
> + return pblk_sysfs_line_state_show(pblk, buf);
> else if (strcmp(attr->name, "max_sec_per_write") == 0)
> return pblk_sysfs_get_sec_per_write(pblk, buf);
> else if (strcmp(attr->name, "write_amp_mileage") == 0)
> @@ -628,6 +693,8 @@ static ssize_t pblk_sysfs_store(struct kobject *kobj, struct attribute *attr,
> return pblk_sysfs_set_sec_per_write(pblk, buf, len);
> else if (strcmp(attr->name, "write_amp_trip") == 0)
> return pblk_sysfs_set_write_amp_trip(pblk, buf, len);
> + else if (strcmp(attr->name, "line_state") == 0)
> + return pblk_sysfs_line_state_store(pblk, buf, len);
> else if (strcmp(attr->name, "padding_dist") == 0)
> return pblk_sysfs_set_padding_dist(pblk, buf, len);
> return 0;
> diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
> index 6e1fcd1a538a..bc31c67b725f 100644
> --- a/drivers/lightnvm/pblk.h
> +++ b/drivers/lightnvm/pblk.h
> @@ -201,6 +201,8 @@ struct pblk_rb {
>
> struct pblk_lun {
> struct ppa_addr bppa;
> + struct ppa_addr chunk_bppa;
> +
> struct semaphore wr_sem;
> };
>
> @@ -297,6 +299,7 @@ enum {
> PBLK_LINETYPE_DATA = 2,
>
> /* Line state */
> + PBLK_LINESTATE_NEW = 9,
> PBLK_LINESTATE_FREE = 10,
> PBLK_LINESTATE_OPEN = 11,
> PBLK_LINESTATE_CLOSED = 12,
> @@ -412,6 +415,15 @@ struct pblk_smeta {
> struct line_smeta *buf; /* smeta buffer in persistent format */
> };
>
> +struct pblk_chunk {
> + int state;
> + int type;
> + int wi;
> + u64 slba;
> + u64 cnlb;
> + u64 wp;
> +};
> +
How come the code replicate the nvm_chk_meta data structure?
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 17/19] lightnvm: pblk: implement get log report chunk
2018-02-26 19:04 ` Matias Bjørling
@ 2018-02-27 14:40 ` Javier González
2018-02-27 18:46 ` Matias Bjørling
0 siblings, 1 reply; 29+ messages in thread
From: Javier González @ 2018-02-27 14:40 UTC (permalink / raw)
To: Matias Bjørling; +Cc: linux-block, linux-kernel, linux-nvme
[-- Attachment #1: Type: text/plain, Size: 23272 bytes --]
> On 26 Feb 2018, at 20.04, Matias Bjørling <mb@lightnvm.io> wrote:
>
> On 02/26/2018 02:17 PM, Javier González wrote:
>> From: Javier González <javier@javigon.com>
>> In preparation of pblk supporting 2.0, implement the get log report
>> chunk in pblk.
>> This patch only replicates de bad block functionality as the rest of the
>> metadata requires new pblk functionality (e.g., wear-index to implement
>> wear-leveling).
>
>
> "This functionality will come in future patches."
>
> I think the above belongs in the cover letter, not in the patch
> description.
>
Ok.
>> Signed-off-by: Javier González <javier@cnexlabs.com>
>> ---
>> drivers/lightnvm/pblk-core.c | 121 +++++++++++++++++++----
>> drivers/lightnvm/pblk-init.c | 218 ++++++++++++++++++++++++++++++------------
>> drivers/lightnvm/pblk-sysfs.c | 67 +++++++++++++
>> drivers/lightnvm/pblk.h | 20 ++++
>> 4 files changed, 346 insertions(+), 80 deletions(-)
>> diff --git a/drivers/lightnvm/pblk-core.c b/drivers/lightnvm/pblk-core.c
>> index 1d7c6313f3d9..ce6a7cfdba66 100644
>> --- a/drivers/lightnvm/pblk-core.c
>> +++ b/drivers/lightnvm/pblk-core.c
>> @@ -44,11 +44,12 @@ static void pblk_line_mark_bb(struct work_struct *work)
>> }
>> static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
>> - struct ppa_addr *ppa)
>> + struct ppa_addr ppa_addr)
>> {
>> struct nvm_tgt_dev *dev = pblk->dev;
>> struct nvm_geo *geo = &dev->geo;
>> - int pos = pblk_ppa_to_pos(geo, *ppa);
>> + struct ppa_addr *ppa;
>> + int pos = pblk_ppa_to_pos(geo, ppa_addr);
>> pr_debug("pblk: erase failed: line:%d, pos:%d\n", line->id, pos);
>> atomic_long_inc(&pblk->erase_failed);
>> @@ -58,6 +59,15 @@ static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
>> pr_err("pblk: attempted to erase bb: line:%d, pos:%d\n",
>> line->id, pos);
>> + /* Not necessary to mark bad blocks on 2.0 spec. */
>> + if (geo->c.version == NVM_OCSSD_SPEC_20)
>> + return;
>> +
>> + ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
>> + if (!ppa)
>> + return;
>> +
>> + *ppa = ppa_addr;
>> pblk_gen_run_ws(pblk, NULL, ppa, pblk_line_mark_bb,
>> GFP_ATOMIC, pblk->bb_wq);
>> }
>> @@ -69,16 +79,8 @@ static void __pblk_end_io_erase(struct pblk *pblk, struct nvm_rq *rqd)
>> line = &pblk->lines[pblk_ppa_to_line(rqd->ppa_addr)];
>> atomic_dec(&line->left_seblks);
>> - if (rqd->error) {
>> - struct ppa_addr *ppa;
>> -
>> - ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
>> - if (!ppa)
>> - return;
>> -
>> - *ppa = rqd->ppa_addr;
>> - pblk_mark_bb(pblk, line, ppa);
>> - }
>> + if (rqd->error)
>> + pblk_mark_bb(pblk, line, rqd->ppa_addr);
>> atomic_dec(&pblk->inflight_io);
>> }
>> @@ -92,6 +94,50 @@ static void pblk_end_io_erase(struct nvm_rq *rqd)
>> mempool_free(rqd, pblk->e_rq_pool);
>> }
>> +/*
>> + * Get information for all chunks from the device.
>> + *
>> + * The caller is responsible for freeing the returned structure
>> + */
>> +struct nvm_chk_meta *pblk_chunk_get_info(struct pblk *pblk)
>> +{
>> + struct nvm_tgt_dev *dev = pblk->dev;
>> + struct nvm_geo *geo = &dev->geo;
>> + struct nvm_chk_meta *meta;
>> + struct ppa_addr ppa;
>> + unsigned long len;
>> + int ret;
>> +
>> + ppa.ppa = 0;
>> +
>> + len = geo->all_chunks * sizeof(*meta);
>> + meta = kzalloc(len, GFP_KERNEL);
>> + if (!meta)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + ret = nvm_get_chunk_meta(dev, meta, ppa, geo->all_chunks);
>> + if (ret) {
>> + pr_err("pblk: could not get chunk metadata (%d)\n", ret);
>> + kfree(meta);
>> + return ERR_PTR(-EIO);
>> + }
>> +
>> + return meta;
>> +}
>> +
>> +struct nvm_chk_meta *pblk_chunk_get_off(struct pblk *pblk,
>> + struct nvm_chk_meta *meta,
>> + struct ppa_addr ppa)
>> +{
>> + struct nvm_tgt_dev *dev = pblk->dev;
>> + struct nvm_geo *geo = &dev->geo;
>> + int ch_off = ppa.m.grp * geo->c.num_chk * geo->num_lun;
>> + int lun_off = ppa.m.pu * geo->c.num_chk;
>> + int chk_off = ppa.m.chk;
>> +
>> + return meta + ch_off + lun_off + chk_off;
>> +}
>> +
>> void __pblk_map_invalidate(struct pblk *pblk, struct pblk_line *line,
>> u64 paddr)
>> {
>> @@ -1094,10 +1140,38 @@ static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
>> return 1;
>> }
>> +static int pblk_prepare_new_line(struct pblk *pblk, struct pblk_line *line)
>> +{
>> + struct pblk_line_meta *lm = &pblk->lm;
>> + struct nvm_tgt_dev *dev = pblk->dev;
>> + struct nvm_geo *geo = &dev->geo;
>> + int blk_to_erase = atomic_read(&line->blk_in_line);
>> + int i;
>> +
>> + for (i = 0; i < lm->blk_per_line; i++) {
>> + int state = line->chks[i].state;
>> + struct pblk_lun *rlun = &pblk->luns[i];
>> +
>> + /* Free chunks should not be erased */
>> + if (state & NVM_CHK_ST_FREE) {
>> + set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
>> + line->erase_bitmap);
>> + blk_to_erase--;
>> + line->chks[i].state = NVM_CHK_ST_HOST_USE;
>
>
> Can you help me understand why you want to use The
> NVM_CHK_ST_HOST_USE? Why would I care if the chunk state is HOST_USE?
> A target instance should not be able to see states from other chunks
> it doesn't own. and in that case, why have a separate state?
>
The motivation for this state is that pblk does not maintain a per
block/chunk state. On the first disk pass however, the state of the
chunks is unknown, thus we need to check for each one individually;
after that, we know that good chunks need to be erased. Having this
extra chunk state, frees us from maintaining this unnecessary intra-line
state. Also, at the point I wrote this code, we did not have the report
chunk on the erase path, so updating the chunk metadata would be
something periodic.
I have taken a look at this again and I solved it in a different way -
basically, only maintaining free/closed states, which is not much more
expensive. It is included in the next version.
>
>> + }
>> +
>> + WARN_ONCE(state & NVM_CHK_ST_OPEN,
>> + "pblk: open chunk in new line: %d\n",
>> + line->id);
>> + }
>> +
>> + return blk_to_erase;
>> +}
>> +
>> static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>> {
>> struct pblk_line_meta *lm = &pblk->lm;
>> - int blk_in_line = atomic_read(&line->blk_in_line);
>> + int blk_to_erase;
>> line->map_bitmap = kzalloc(lm->sec_bitmap_len, GFP_ATOMIC);
>> if (!line->map_bitmap)
>> @@ -1110,7 +1184,21 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>> return -ENOMEM;
>> }
>> + /* Bad blocks do not need to be erased */
>> + bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
>> +
>> spin_lock(&line->lock);
>> +
>> + /* If we have not written to this line, we need to mark up free chunks
>> + * as already erased
>> + */
>> + if (line->state == PBLK_LINESTATE_NEW) {
>> + blk_to_erase = pblk_prepare_new_line(pblk, line);
>> + line->state = PBLK_LINESTATE_FREE;
>> + } else {
>> + blk_to_erase = atomic_read(&line->blk_in_line);
>> + }
>> +
>> if (line->state != PBLK_LINESTATE_FREE) {
>> kfree(line->map_bitmap);
>> kfree(line->invalid_bitmap);
>> @@ -1122,15 +1210,12 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>> line->state = PBLK_LINESTATE_OPEN;
>> - atomic_set(&line->left_eblks, blk_in_line);
>> - atomic_set(&line->left_seblks, blk_in_line);
>> + atomic_set(&line->left_eblks, blk_to_erase);
>> + atomic_set(&line->left_seblks, blk_to_erase);
>> line->meta_distance = lm->meta_distance;
>> spin_unlock(&line->lock);
>> - /* Bad blocks do not need to be erased */
>> - bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
>> -
>> kref_init(&line->ref);
>> return 0;
>> diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
>> index ec39800eea42..cf4f49d48aed 100644
>> --- a/drivers/lightnvm/pblk-init.c
>> +++ b/drivers/lightnvm/pblk-init.c
>> @@ -401,6 +401,7 @@ static void pblk_line_meta_free(struct pblk_line *line)
>> {
>> kfree(line->blk_bitmap);
>> kfree(line->erase_bitmap);
>> + kfree(line->chks);
>> }
>> static void pblk_lines_free(struct pblk *pblk)
>> @@ -440,54 +441,44 @@ static int pblk_bb_get_tbl(struct nvm_tgt_dev *dev, struct pblk_lun *rlun,
>> return 0;
>> }
>> -static void *pblk_bb_get_log(struct pblk *pblk)
>> +static void *pblk_bb_get_meta(struct pblk *pblk)
>> {
>> struct nvm_tgt_dev *dev = pblk->dev;
>> struct nvm_geo *geo = &dev->geo;
>> - u8 *log;
>> + u8 *meta;
>> int i, nr_blks, blk_per_lun;
>> int ret;
>> blk_per_lun = geo->c.num_chk * geo->c.pln_mode;
>> nr_blks = blk_per_lun * geo->all_luns;
>> - log = kmalloc(nr_blks, GFP_KERNEL);
>> - if (!log)
>> + meta = kmalloc(nr_blks, GFP_KERNEL);
>> + if (!meta)
>> return ERR_PTR(-ENOMEM);
>> for (i = 0; i < geo->all_luns; i++) {
>> struct pblk_lun *rlun = &pblk->luns[i];
>> - u8 *log_pos = log + i * blk_per_lun;
>> + u8 *meta_pos = meta + i * blk_per_lun;
>> - ret = pblk_bb_get_tbl(dev, rlun, log_pos, blk_per_lun);
>> + ret = pblk_bb_get_tbl(dev, rlun, meta_pos, blk_per_lun);
>> if (ret) {
>> - kfree(log);
>> + kfree(meta);
>> return ERR_PTR(-EIO);
>> }
>> }
>> - return log;
>> + return meta;
>> }
>> -static int pblk_bb_line(struct pblk *pblk, struct pblk_line *line,
>> - u8 *bb_log, int blk_per_line)
>> +static void *pblk_chunk_get_meta(struct pblk *pblk)
>> {
>> struct nvm_tgt_dev *dev = pblk->dev;
>> struct nvm_geo *geo = &dev->geo;
>> - int i, bb_cnt = 0;
>> - for (i = 0; i < blk_per_line; i++) {
>> - struct pblk_lun *rlun = &pblk->luns[i];
>> - u8 *lun_bb_log = bb_log + i * blk_per_line;
>> -
>> - if (lun_bb_log[line->id] == NVM_BLK_T_FREE)
>> - continue;
>> -
>> - set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
>> - bb_cnt++;
>> - }
>> -
>> - return bb_cnt;
>> + if (geo->c.version == NVM_OCSSD_SPEC_12)
>> + return pblk_bb_get_meta(pblk);
>> + else
>> + return pblk_chunk_get_info(pblk);
>
> This 1.2 or 2.0 would be nice to have inside the lightnvm core. A
> target should not care weather it is a 1.2 or 2.0 device.
>
It is not an easy thing to do since the bad block format and chunk
report return completely different formats. As I explained in a
different thread, doing this in core would force 1.2 to understand fake
chunk formats, which are not meaningful. It is better to leave the logic
at the target level, where we know what to do with the format.
>> }
>> static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>> @@ -516,6 +507,7 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>> rlun = &pblk->luns[i];
>> rlun->bppa = luns[lunid];
>> + rlun->chunk_bppa = luns[i];
>> sema_init(&rlun->wr_sem, 1);
>> }
>> @@ -695,8 +687,125 @@ static int pblk_lines_alloc_metadata(struct pblk *pblk)
>> return -ENOMEM;
>> }
>> -static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>> - void *chunk_log, long *nr_bad_blks)
>> +static int pblk_setup_line_meta_12(struct pblk *pblk, struct pblk_line *line,
>> + void *chunk_meta)
>> +{
>> + struct nvm_tgt_dev *dev = pblk->dev;
>> + struct nvm_geo *geo = &dev->geo;
>> + struct pblk_line_meta *lm = &pblk->lm;
>> + int i, chk_per_lun, nr_bad_chks = 0;
>> +
>> + chk_per_lun = geo->c.num_chk * geo->c.pln_mode;
>> +
>> + for (i = 0; i < lm->blk_per_line; i++) {
>> + struct pblk_chunk *chunk = &line->chks[i];
>> + struct pblk_lun *rlun = &pblk->luns[i];
>> + u8 *lun_bb_meta = chunk_meta + i * chk_per_lun;
>> +
>> + /*
>> + * In 1.2 spec. chunk state is not persisted by the device. Thus
>> + * some of the values are reset each time pblk is instantiated.
>> + */
>> + if (lun_bb_meta[line->id] == NVM_BLK_T_FREE)
>> + chunk->state = NVM_CHK_ST_HOST_USE;
>> + else
>> + chunk->state = NVM_CHK_ST_OFFLINE;
>> +
>> + chunk->type = NVM_CHK_TP_W_SEQ;
>> + chunk->wi = 0;
>> + chunk->slba = -1;
>> + chunk->cnlb = geo->c.clba;
>> + chunk->wp = 0;
>> +
>> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
>> + continue;
>> +
>> + set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
>> + nr_bad_chks++;
>> + }
>> +
>> + return nr_bad_chks;
>> +}
>> +
>> +static int pblk_setup_line_meta_20(struct pblk *pblk, struct pblk_line *line,
>> + struct nvm_chk_meta *meta)
>> +{
>> + struct nvm_tgt_dev *dev = pblk->dev;
>> + struct nvm_geo *geo = &dev->geo;
>> + struct pblk_line_meta *lm = &pblk->lm;
>> + int i, nr_bad_chks = 0;
>> +
>> + for (i = 0; i < lm->blk_per_line; i++) {
>> + struct pblk_chunk *chunk = &line->chks[i];
>> + struct pblk_lun *rlun = &pblk->luns[i];
>> + struct nvm_chk_meta *chunk_meta;
>> + struct ppa_addr ppa;
>> +
>> + ppa = rlun->chunk_bppa;
>> + ppa.m.chk = line->id;
>> + chunk_meta = pblk_chunk_get_off(pblk, meta, ppa);
>> +
>> + chunk->state = chunk_meta->state;
>> + chunk->type = chunk_meta->type;
>> + chunk->wi = chunk_meta->wi;
>> + chunk->slba = chunk_meta->slba;
>> + chunk->cnlb = chunk_meta->cnlb;
>> + chunk->wp = chunk_meta->wp;
>> +
>> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
>> + continue;
>> +
>> + if (chunk->type & NVM_CHK_TP_SZ_SPEC) {
>> + WARN_ONCE(1, "pblk: custom-sized chunks unsupported\n");
>> + continue;
>> + }
>> +
>> + set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
>> + line->blk_bitmap);
>> + nr_bad_chks++;
>> + }
>> +
>> + return nr_bad_chks;
>> +}
>> +
>> +static long pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>> + void *chunk_meta, int line_id)
>> +{
>> + struct nvm_tgt_dev *dev = pblk->dev;
>> + struct nvm_geo *geo = &dev->geo;
>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>> + struct pblk_line_meta *lm = &pblk->lm;
>> + long nr_bad_chks, chk_in_line;
>> +
>> + line->pblk = pblk;
>> + line->id = line_id;
>> + line->type = PBLK_LINETYPE_FREE;
>> + line->state = PBLK_LINESTATE_NEW;
>> + line->gc_group = PBLK_LINEGC_NONE;
>> + line->vsc = &l_mg->vsc_list[line_id];
>> + spin_lock_init(&line->lock);
>> +
>> + if (geo->c.version == NVM_OCSSD_SPEC_12)
>> + nr_bad_chks = pblk_setup_line_meta_12(pblk, line, chunk_meta);
>> + else
>> + nr_bad_chks = pblk_setup_line_meta_20(pblk, line, chunk_meta);
>> +
>> + chk_in_line = lm->blk_per_line - nr_bad_chks;
>> + if (nr_bad_chks < 0 || nr_bad_chks > lm->blk_per_line ||
>> + chk_in_line < lm->min_blk_line) {
>> + line->state = PBLK_LINESTATE_BAD;
>> + list_add_tail(&line->list, &l_mg->bad_list);
>> + return 0;
>> + }
>> +
>> + atomic_set(&line->blk_in_line, chk_in_line);
>> + list_add_tail(&line->list, &l_mg->free_list);
>> + l_mg->nr_free_lines++;
>> +
>> + return chk_in_line;
>> +}
>> +
>> +static int pblk_alloc_line_meta(struct pblk *pblk, struct pblk_line *line)
>> {
>> struct pblk_line_meta *lm = &pblk->lm;
>> @@ -710,7 +819,13 @@ static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>> return -ENOMEM;
>> }
>> - *nr_bad_blks = pblk_bb_line(pblk, line, chunk_log, lm->blk_per_line);
>> + line->chks = kmalloc(lm->blk_per_line * sizeof(struct pblk_chunk),
>> + GFP_KERNEL);
>> + if (!line->chks) {
>> + kfree(line->erase_bitmap);
>> + kfree(line->blk_bitmap);
>> + return -ENOMEM;
>> + }
>> return 0;
>> }
>> @@ -722,9 +837,9 @@ static int pblk_lines_init(struct pblk *pblk)
>> struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>> struct pblk_line_meta *lm = &pblk->lm;
>> struct pblk_line *line;
>> - void *chunk_log;
>> + void *chunk_meta;
>> unsigned int smeta_len, emeta_len;
>> - long nr_bad_blks = 0, nr_free_blks = 0;
>> + long nr_free_chks = 0;
>> int bb_distance, max_write_ppas;
>> int i, ret;
>> @@ -743,6 +858,7 @@ static int pblk_lines_init(struct pblk *pblk)
>> l_mg->log_line = l_mg->data_line = NULL;
>> l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
>> l_mg->nr_free_lines = 0;
>> + atomic_set(&l_mg->sysfs_line_state, -1);
>> bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
>> lm->sec_per_line = geo->c.clba * geo->all_luns;
>> @@ -841,53 +957,31 @@ static int pblk_lines_init(struct pblk *pblk)
>> goto fail_free_bb_aux;
>> }
>> - chunk_log = pblk_bb_get_log(pblk);
>> - if (IS_ERR(chunk_log)) {
>> - pr_err("pblk: could not get bad block log (%lu)\n",
>> - PTR_ERR(chunk_log));
>> - ret = PTR_ERR(chunk_log);
>> + chunk_meta = pblk_chunk_get_meta(pblk);
>> + if (IS_ERR(chunk_meta)) {
>> + pr_err("pblk: could not get chunk log (%lu)\n",
>> + PTR_ERR(chunk_meta));
>> + ret = PTR_ERR(chunk_meta);
>> goto fail_free_lines;
>> }
>> for (i = 0; i < l_mg->nr_lines; i++) {
>> - int chk_in_line;
>> -
>> line = &pblk->lines[i];
>> - line->pblk = pblk;
>> - line->id = i;
>> - line->type = PBLK_LINETYPE_FREE;
>> - line->state = PBLK_LINESTATE_FREE;
>> - line->gc_group = PBLK_LINEGC_NONE;
>> - line->vsc = &l_mg->vsc_list[i];
>> - spin_lock_init(&line->lock);
>> -
>> - ret = pblk_setup_line_meta(pblk, line, chunk_log, &nr_bad_blks);
>> + ret = pblk_alloc_line_meta(pblk, line);
>> if (ret)
>> - goto fail_free_chunk_log;
>> + goto fail_free_chunk_meta;
>> - chk_in_line = lm->blk_per_line - nr_bad_blks;
>> - if (nr_bad_blks < 0 || nr_bad_blks > lm->blk_per_line ||
>> - chk_in_line < lm->min_blk_line) {
>> - line->state = PBLK_LINESTATE_BAD;
>> - list_add_tail(&line->list, &l_mg->bad_list);
>> - continue;
>> - }
>> -
>> - nr_free_blks += chk_in_line;
>> - atomic_set(&line->blk_in_line, chk_in_line);
>> -
>> - l_mg->nr_free_lines++;
>> - list_add_tail(&line->list, &l_mg->free_list);
>> + nr_free_chks += pblk_setup_line_meta(pblk, line, chunk_meta, i);
>> }
>> - pblk_set_provision(pblk, nr_free_blks);
>> + pblk_set_provision(pblk, nr_free_chks);
>> - kfree(chunk_log);
>> + kfree(chunk_meta);
>> return 0;
>> -fail_free_chunk_log:
>> - kfree(chunk_log);
>> +fail_free_chunk_meta:
>> + kfree(chunk_meta);
>> while (--i >= 0)
>> pblk_line_meta_free(&pblk->lines[i]);
>> fail_free_lines:
>> diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
>> index ccfb3abd2773..1ce5b956c622 100644
>> --- a/drivers/lightnvm/pblk-sysfs.c
>> +++ b/drivers/lightnvm/pblk-sysfs.c
>> @@ -142,6 +142,40 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
>> return sz;
>> }
>> +static ssize_t pblk_sysfs_line_state_show(struct pblk *pblk, char *page)
>> +{
>> + struct pblk_line_meta *lm = &pblk->lm;
>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>> + struct pblk_line *line;
>> + int line_id = atomic_read(&l_mg->sysfs_line_state);
>> + ssize_t sz = 0;
>> + int i;
>> +
>> + if (line_id < 0 || line_id >= l_mg->nr_lines)
>> + return 0;
>> +
>> + sz = snprintf(page, PAGE_SIZE,
>> + "line\tchunk\tstate\ttype\twear-index\tslba\t\tcnlb\twp\n");
>> +
>> + line = &pblk->lines[line_id];
>> +
>> + for (i = 0; i < lm->blk_per_line; i++) {
>> + struct pblk_chunk *chunk = &line->chks[i];
>> +
>> + sz += snprintf(page + sz, PAGE_SIZE - sz,
>> + "%d\t%d\t%d\t%d\t%d\t\t%llu\t\t%llu\t%llu\n",
>> + line->id, i,
>> + chunk->state,
>> + chunk->type,
>> + chunk->wi,
>> + chunk->slba,
>> + chunk->cnlb,
>> + chunk->wp);
>> + }
>> +
>> + return sz;
>> +}
>> +
>> static ssize_t pblk_sysfs_lines(struct pblk *pblk, char *page)
>> {
>> struct nvm_tgt_dev *dev = pblk->dev;
>> @@ -398,6 +432,29 @@ static ssize_t pblk_sysfs_stats_debug(struct pblk *pblk, char *page)
>> }
>> #endif
>> +
>> +static ssize_t pblk_sysfs_line_state_store(struct pblk *pblk, const char *page,
>> + size_t len)
>> +{
>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>> + size_t c_len;
>> + int line_id;
>> +
>> + c_len = strcspn(page, "\n");
>> + if (c_len >= len)
>> + return -EINVAL;
>> +
>> + if (kstrtouint(page, 0, &line_id))
>> + return -EINVAL;
>> +
>> + if (line_id < 0 || line_id >= l_mg->nr_lines)
>> + return -EINVAL;
>> +
>> + atomic_set(&l_mg->sysfs_line_state, line_id);
>> +
>> + return len;
>> +}
>> +
>> static ssize_t pblk_sysfs_gc_force(struct pblk *pblk, const char *page,
>> size_t len)
>> {
>> @@ -529,6 +586,11 @@ static struct attribute sys_lines_info_attr = {
>> .mode = 0444,
>> };
>> +static struct attribute sys_line_state_attr = {
>> + .name = "line_state",
>> + .mode = 0644,
>> +};
>> +
>> static struct attribute sys_gc_force = {
>> .name = "gc_force",
>> .mode = 0200,
>> @@ -572,6 +634,7 @@ static struct attribute *pblk_attrs[] = {
>> &sys_stats_ppaf_attr,
>> &sys_lines_attr,
>> &sys_lines_info_attr,
>> + &sys_line_state_attr,
>> &sys_write_amp_mileage,
>> &sys_write_amp_trip,
>> &sys_padding_dist,
>> @@ -602,6 +665,8 @@ static ssize_t pblk_sysfs_show(struct kobject *kobj, struct attribute *attr,
>> return pblk_sysfs_lines(pblk, buf);
>> else if (strcmp(attr->name, "lines_info") == 0)
>> return pblk_sysfs_lines_info(pblk, buf);
>> + else if (strcmp(attr->name, "line_state") == 0)
>> + return pblk_sysfs_line_state_show(pblk, buf);
>> else if (strcmp(attr->name, "max_sec_per_write") == 0)
>> return pblk_sysfs_get_sec_per_write(pblk, buf);
>> else if (strcmp(attr->name, "write_amp_mileage") == 0)
>> @@ -628,6 +693,8 @@ static ssize_t pblk_sysfs_store(struct kobject *kobj, struct attribute *attr,
>> return pblk_sysfs_set_sec_per_write(pblk, buf, len);
>> else if (strcmp(attr->name, "write_amp_trip") == 0)
>> return pblk_sysfs_set_write_amp_trip(pblk, buf, len);
>> + else if (strcmp(attr->name, "line_state") == 0)
>> + return pblk_sysfs_line_state_store(pblk, buf, len);
>> else if (strcmp(attr->name, "padding_dist") == 0)
>> return pblk_sysfs_set_padding_dist(pblk, buf, len);
>> return 0;
>> diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
>> index 6e1fcd1a538a..bc31c67b725f 100644
>> --- a/drivers/lightnvm/pblk.h
>> +++ b/drivers/lightnvm/pblk.h
>> @@ -201,6 +201,8 @@ struct pblk_rb {
>> struct pblk_lun {
>> struct ppa_addr bppa;
>> + struct ppa_addr chunk_bppa;
>> +
>> struct semaphore wr_sem;
>> };
>> @@ -297,6 +299,7 @@ enum {
>> PBLK_LINETYPE_DATA = 2,
>> /* Line state */
>> + PBLK_LINESTATE_NEW = 9,
>> PBLK_LINESTATE_FREE = 10,
>> PBLK_LINESTATE_OPEN = 11,
>> PBLK_LINESTATE_CLOSED = 12,
>> @@ -412,6 +415,15 @@ struct pblk_smeta {
>> struct line_smeta *buf; /* smeta buffer in persistent format */
>> };
>> +struct pblk_chunk {
>> + int state;
>> + int type;
>> + int wi;
>> + u64 slba;
>> + u64 cnlb;
>> + u64 wp;
>> +};
>> +
>
> How come the code replicate the nvm_chk_meta data structure?
There is no need to maintain the chunk reserved bits in nvm_chk_meta and
waste memory in the process, as these are alive for the whole life of
the pblk instance. This structure only maintains the necessary bits.
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 17/19] lightnvm: pblk: implement get log report chunk
2018-02-27 14:40 ` Javier González
@ 2018-02-27 18:46 ` Matias Bjørling
2018-02-27 19:50 ` Javier González
0 siblings, 1 reply; 29+ messages in thread
From: Matias Bjørling @ 2018-02-27 18:46 UTC (permalink / raw)
To: Javier González; +Cc: linux-block, linux-kernel, linux-nvme
On 02/27/2018 03:40 PM, Javier González wrote:
>> On 26 Feb 2018, at 20.04, Matias Bjørling <mb@lightnvm.io> wrote:
>>
<snip>
>>
>>
>> Can you help me understand why you want to use The
>> NVM_CHK_ST_HOST_USE? Why would I care if the chunk state is HOST_USE?
>> A target instance should not be able to see states from other chunks
>> it doesn't own. and in that case, why have a separate state?
>>
>
> The motivation for this state is that pblk does not maintain a per
> block/chunk state. On the first disk pass however, the state of the
> chunks is unknown,
I don't understand this. Why is the state unknown? If assuming 1.2
drive, one can do a scan of first page of each block. If the block
returns success (it is in use), if 0x42ff, the block is empty, and if
reading and it returns 0x4281, it is open.
thus we need to check for each one individually;
> after that, we know that good chunks need to be erased. Having this
> extra chunk state, frees us from maintaining this unnecessary intra-line
> state. Also, at the point I wrote this code, we did not have the report
> chunk on the erase path, so updating the chunk metadata would be
> something periodic.
>
> I have taken a look at this again and I solved it in a different way -
> basically, only maintaining free/closed states, which is not much more
> expensive. It is included in the next version.
>
>>
>>> + }
>>> +
>>> + WARN_ONCE(state & NVM_CHK_ST_OPEN,
>>> + "pblk: open chunk in new line: %d\n",
>>> + line->id);
>>> + }
>>> +
>>> + return blk_to_erase;
>>> +}
>>> +
>>> static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>>> {
>>> struct pblk_line_meta *lm = &pblk->lm;
>>> - int blk_in_line = atomic_read(&line->blk_in_line);
>>> + int blk_to_erase;
>>> line->map_bitmap = kzalloc(lm->sec_bitmap_len, GFP_ATOMIC);
>>> if (!line->map_bitmap)
>>> @@ -1110,7 +1184,21 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>>> return -ENOMEM;
>>> }
>>> + /* Bad blocks do not need to be erased */
>>> + bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
>>> +
>>> spin_lock(&line->lock);
>>> +
>>> + /* If we have not written to this line, we need to mark up free chunks
>>> + * as already erased
>>> + */
>>> + if (line->state == PBLK_LINESTATE_NEW) {
>>> + blk_to_erase = pblk_prepare_new_line(pblk, line);
>>> + line->state = PBLK_LINESTATE_FREE;
>>> + } else {
>>> + blk_to_erase = atomic_read(&line->blk_in_line);
>>> + }
>>> +
>>> if (line->state != PBLK_LINESTATE_FREE) {
>>> kfree(line->map_bitmap);
>>> kfree(line->invalid_bitmap);
>>> @@ -1122,15 +1210,12 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>>> line->state = PBLK_LINESTATE_OPEN;
>>> - atomic_set(&line->left_eblks, blk_in_line);
>>> - atomic_set(&line->left_seblks, blk_in_line);
>>> + atomic_set(&line->left_eblks, blk_to_erase);
>>> + atomic_set(&line->left_seblks, blk_to_erase);
>>> line->meta_distance = lm->meta_distance;
>>> spin_unlock(&line->lock);
>>> - /* Bad blocks do not need to be erased */
>>> - bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
>>> -
>>> kref_init(&line->ref);
>>> return 0;
>>> diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
>>> index ec39800eea42..cf4f49d48aed 100644
>>> --- a/drivers/lightnvm/pblk-init.c
>>> +++ b/drivers/lightnvm/pblk-init.c
>>> @@ -401,6 +401,7 @@ static void pblk_line_meta_free(struct pblk_line *line)
>>> {
>>> kfree(line->blk_bitmap);
>>> kfree(line->erase_bitmap);
>>> + kfree(line->chks);
>>> }
>>> static void pblk_lines_free(struct pblk *pblk)
>>> @@ -440,54 +441,44 @@ static int pblk_bb_get_tbl(struct nvm_tgt_dev *dev, struct pblk_lun *rlun,
>>> return 0;
>>> }
>>> -static void *pblk_bb_get_log(struct pblk *pblk)
>>> +static void *pblk_bb_get_meta(struct pblk *pblk)
>>> {
>>> struct nvm_tgt_dev *dev = pblk->dev;
>>> struct nvm_geo *geo = &dev->geo;
>>> - u8 *log;
>>> + u8 *meta;
>>> int i, nr_blks, blk_per_lun;
>>> int ret;
>>> blk_per_lun = geo->c.num_chk * geo->c.pln_mode;
>>> nr_blks = blk_per_lun * geo->all_luns;
>>> - log = kmalloc(nr_blks, GFP_KERNEL);
>>> - if (!log)
>>> + meta = kmalloc(nr_blks, GFP_KERNEL);
>>> + if (!meta)
>>> return ERR_PTR(-ENOMEM);
>>> for (i = 0; i < geo->all_luns; i++) {
>>> struct pblk_lun *rlun = &pblk->luns[i];
>>> - u8 *log_pos = log + i * blk_per_lun;
>>> + u8 *meta_pos = meta + i * blk_per_lun;
>>> - ret = pblk_bb_get_tbl(dev, rlun, log_pos, blk_per_lun);
>>> + ret = pblk_bb_get_tbl(dev, rlun, meta_pos, blk_per_lun);
>>> if (ret) {
>>> - kfree(log);
>>> + kfree(meta);
>>> return ERR_PTR(-EIO);
>>> }
>>> }
>>> - return log;
>>> + return meta;
>>> }
>>> -static int pblk_bb_line(struct pblk *pblk, struct pblk_line *line,
>>> - u8 *bb_log, int blk_per_line)
>>> +static void *pblk_chunk_get_meta(struct pblk *pblk)
>>> {
>>> struct nvm_tgt_dev *dev = pblk->dev;
>>> struct nvm_geo *geo = &dev->geo;
>>> - int i, bb_cnt = 0;
>>> - for (i = 0; i < blk_per_line; i++) {
>>> - struct pblk_lun *rlun = &pblk->luns[i];
>>> - u8 *lun_bb_log = bb_log + i * blk_per_line;
>>> -
>>> - if (lun_bb_log[line->id] == NVM_BLK_T_FREE)
>>> - continue;
>>> -
>>> - set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
>>> - bb_cnt++;
>>> - }
>>> -
>>> - return bb_cnt;
>>> + if (geo->c.version == NVM_OCSSD_SPEC_12)
>>> + return pblk_bb_get_meta(pblk);
>>> + else
>>> + return pblk_chunk_get_info(pblk);
>>
>> This 1.2 or 2.0 would be nice to have inside the lightnvm core. A
>> target should not care weather it is a 1.2 or 2.0 device.
>>
>
> It is not an easy thing to do since the bad block format and chunk
> report return completely different formats. As I explained in a
> different thread, doing this in core would force 1.2 to understand fake
> chunk formats, which are not meaningful. It is better to leave the logic
> at the target level, where we know what to do with the format.
What if the core layer exported a generic "chunk meta" interface, and
then filled it as is done in pblk. In the core, it will then fill out
what it can fill out, and leave the rest of the fields untouched. The
pblk code can then if () in certain places if for example wp attribute
can,'t be used?
In the interest of time, and to get this moving, we can ignore it for now.
>
>>> }
>>> static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>>> @@ -516,6 +507,7 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>>> rlun = &pblk->luns[i];
>>> rlun->bppa = luns[lunid];
>>> + rlun->chunk_bppa = luns[i];
>>> sema_init(&rlun->wr_sem, 1);
>>> }
>>> @@ -695,8 +687,125 @@ static int pblk_lines_alloc_metadata(struct pblk *pblk)
>>> return -ENOMEM;
>>> }
>>> -static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>>> - void *chunk_log, long *nr_bad_blks)
>>> +static int pblk_setup_line_meta_12(struct pblk *pblk, struct pblk_line *line,
>>> + void *chunk_meta)
>>> +{
>>> + struct nvm_tgt_dev *dev = pblk->dev;
>>> + struct nvm_geo *geo = &dev->geo;
>>> + struct pblk_line_meta *lm = &pblk->lm;
>>> + int i, chk_per_lun, nr_bad_chks = 0;
>>> +
>>> + chk_per_lun = geo->c.num_chk * geo->c.pln_mode;
>>> +
>>> + for (i = 0; i < lm->blk_per_line; i++) {
>>> + struct pblk_chunk *chunk = &line->chks[i];
>>> + struct pblk_lun *rlun = &pblk->luns[i];
>>> + u8 *lun_bb_meta = chunk_meta + i * chk_per_lun;
>>> +
>>> + /*
>>> + * In 1.2 spec. chunk state is not persisted by the device. Thus
>>> + * some of the values are reset each time pblk is instantiated.
>>> + */
>>> + if (lun_bb_meta[line->id] == NVM_BLK_T_FREE)
>>> + chunk->state = NVM_CHK_ST_HOST_USE;
>>> + else
>>> + chunk->state = NVM_CHK_ST_OFFLINE;
>>> +
>>> + chunk->type = NVM_CHK_TP_W_SEQ;
>>> + chunk->wi = 0;
>>> + chunk->slba = -1;
>>> + chunk->cnlb = geo->c.clba;
>>> + chunk->wp = 0;
>>> +
>>> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
>>> + continue;
>>> +
>>> + set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
>>> + nr_bad_chks++;
>>> + }
>>> +
>>> + return nr_bad_chks;
>>> +}
>>> +
>>> +static int pblk_setup_line_meta_20(struct pblk *pblk, struct pblk_line *line,
>>> + struct nvm_chk_meta *meta)
>>> +{
>>> + struct nvm_tgt_dev *dev = pblk->dev;
>>> + struct nvm_geo *geo = &dev->geo;
>>> + struct pblk_line_meta *lm = &pblk->lm;
>>> + int i, nr_bad_chks = 0;
>>> +
>>> + for (i = 0; i < lm->blk_per_line; i++) {
>>> + struct pblk_chunk *chunk = &line->chks[i];
>>> + struct pblk_lun *rlun = &pblk->luns[i];
>>> + struct nvm_chk_meta *chunk_meta;
>>> + struct ppa_addr ppa;
>>> +
>>> + ppa = rlun->chunk_bppa;
>>> + ppa.m.chk = line->id;
>>> + chunk_meta = pblk_chunk_get_off(pblk, meta, ppa);
>>> +
>>> + chunk->state = chunk_meta->state;
>>> + chunk->type = chunk_meta->type;
>>> + chunk->wi = chunk_meta->wi;
>>> + chunk->slba = chunk_meta->slba;
>>> + chunk->cnlb = chunk_meta->cnlb;
>>> + chunk->wp = chunk_meta->wp;
>>> +
>>> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
>>> + continue;
>>> +
>>> + if (chunk->type & NVM_CHK_TP_SZ_SPEC) {
>>> + WARN_ONCE(1, "pblk: custom-sized chunks unsupported\n");
>>> + continue;
>>> + }
>>> +
>>> + set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
>>> + line->blk_bitmap);
>>> + nr_bad_chks++;
>>> + }
>>> +
>>> + return nr_bad_chks;
>>> +}
>>> +
>>> +static long pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>>> + void *chunk_meta, int line_id)
>>> +{
>>> + struct nvm_tgt_dev *dev = pblk->dev;
>>> + struct nvm_geo *geo = &dev->geo;
>>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>> + struct pblk_line_meta *lm = &pblk->lm;
>>> + long nr_bad_chks, chk_in_line;
>>> +
>>> + line->pblk = pblk;
>>> + line->id = line_id;
>>> + line->type = PBLK_LINETYPE_FREE;
>>> + line->state = PBLK_LINESTATE_NEW;
>>> + line->gc_group = PBLK_LINEGC_NONE;
>>> + line->vsc = &l_mg->vsc_list[line_id];
>>> + spin_lock_init(&line->lock);
>>> +
>>> + if (geo->c.version == NVM_OCSSD_SPEC_12)
>>> + nr_bad_chks = pblk_setup_line_meta_12(pblk, line, chunk_meta);
>>> + else
>>> + nr_bad_chks = pblk_setup_line_meta_20(pblk, line, chunk_meta);
>>> +
>>> + chk_in_line = lm->blk_per_line - nr_bad_chks;
>>> + if (nr_bad_chks < 0 || nr_bad_chks > lm->blk_per_line ||
>>> + chk_in_line < lm->min_blk_line) {
>>> + line->state = PBLK_LINESTATE_BAD;
>>> + list_add_tail(&line->list, &l_mg->bad_list);
>>> + return 0;
>>> + }
>>> +
>>> + atomic_set(&line->blk_in_line, chk_in_line);
>>> + list_add_tail(&line->list, &l_mg->free_list);
>>> + l_mg->nr_free_lines++;
>>> +
>>> + return chk_in_line;
>>> +}
>>> +
>>> +static int pblk_alloc_line_meta(struct pblk *pblk, struct pblk_line *line)
>>> {
>>> struct pblk_line_meta *lm = &pblk->lm;
>>> @@ -710,7 +819,13 @@ static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>>> return -ENOMEM;
>>> }
>>> - *nr_bad_blks = pblk_bb_line(pblk, line, chunk_log, lm->blk_per_line);
>>> + line->chks = kmalloc(lm->blk_per_line * sizeof(struct pblk_chunk),
>>> + GFP_KERNEL);
>>> + if (!line->chks) {
>>> + kfree(line->erase_bitmap);
>>> + kfree(line->blk_bitmap);
>>> + return -ENOMEM;
>>> + }
>>> return 0;
>>> }
>>> @@ -722,9 +837,9 @@ static int pblk_lines_init(struct pblk *pblk)
>>> struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>> struct pblk_line_meta *lm = &pblk->lm;
>>> struct pblk_line *line;
>>> - void *chunk_log;
>>> + void *chunk_meta;
>>> unsigned int smeta_len, emeta_len;
>>> - long nr_bad_blks = 0, nr_free_blks = 0;
>>> + long nr_free_chks = 0;
>>> int bb_distance, max_write_ppas;
>>> int i, ret;
>>> @@ -743,6 +858,7 @@ static int pblk_lines_init(struct pblk *pblk)
>>> l_mg->log_line = l_mg->data_line = NULL;
>>> l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
>>> l_mg->nr_free_lines = 0;
>>> + atomic_set(&l_mg->sysfs_line_state, -1);
>>> bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
>>> lm->sec_per_line = geo->c.clba * geo->all_luns;
>>> @@ -841,53 +957,31 @@ static int pblk_lines_init(struct pblk *pblk)
>>> goto fail_free_bb_aux;
>>> }
>>> - chunk_log = pblk_bb_get_log(pblk);
>>> - if (IS_ERR(chunk_log)) {
>>> - pr_err("pblk: could not get bad block log (%lu)\n",
>>> - PTR_ERR(chunk_log));
>>> - ret = PTR_ERR(chunk_log);
>>> + chunk_meta = pblk_chunk_get_meta(pblk);
>>> + if (IS_ERR(chunk_meta)) {
>>> + pr_err("pblk: could not get chunk log (%lu)\n",
>>> + PTR_ERR(chunk_meta));
>>> + ret = PTR_ERR(chunk_meta);
>>> goto fail_free_lines;
>>> }
>>> for (i = 0; i < l_mg->nr_lines; i++) {
>>> - int chk_in_line;
>>> -
>>> line = &pblk->lines[i];
>>> - line->pblk = pblk;
>>> - line->id = i;
>>> - line->type = PBLK_LINETYPE_FREE;
>>> - line->state = PBLK_LINESTATE_FREE;
>>> - line->gc_group = PBLK_LINEGC_NONE;
>>> - line->vsc = &l_mg->vsc_list[i];
>>> - spin_lock_init(&line->lock);
>>> -
>>> - ret = pblk_setup_line_meta(pblk, line, chunk_log, &nr_bad_blks);
>>> + ret = pblk_alloc_line_meta(pblk, line);
>>> if (ret)
>>> - goto fail_free_chunk_log;
>>> + goto fail_free_chunk_meta;
>>> - chk_in_line = lm->blk_per_line - nr_bad_blks;
>>> - if (nr_bad_blks < 0 || nr_bad_blks > lm->blk_per_line ||
>>> - chk_in_line < lm->min_blk_line) {
>>> - line->state = PBLK_LINESTATE_BAD;
>>> - list_add_tail(&line->list, &l_mg->bad_list);
>>> - continue;
>>> - }
>>> -
>>> - nr_free_blks += chk_in_line;
>>> - atomic_set(&line->blk_in_line, chk_in_line);
>>> -
>>> - l_mg->nr_free_lines++;
>>> - list_add_tail(&line->list, &l_mg->free_list);
>>> + nr_free_chks += pblk_setup_line_meta(pblk, line, chunk_meta, i);
>>> }
>>> - pblk_set_provision(pblk, nr_free_blks);
>>> + pblk_set_provision(pblk, nr_free_chks);
>>> - kfree(chunk_log);
>>> + kfree(chunk_meta);
>>> return 0;
>>> -fail_free_chunk_log:
>>> - kfree(chunk_log);
>>> +fail_free_chunk_meta:
>>> + kfree(chunk_meta);
>>> while (--i >= 0)
>>> pblk_line_meta_free(&pblk->lines[i]);
>>> fail_free_lines:
>>> diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
>>> index ccfb3abd2773..1ce5b956c622 100644
>>> --- a/drivers/lightnvm/pblk-sysfs.c
>>> +++ b/drivers/lightnvm/pblk-sysfs.c
>>> @@ -142,6 +142,40 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
>>> return sz;
>>> }
>>> +static ssize_t pblk_sysfs_line_state_show(struct pblk *pblk, char *page)
>>> +{
>>> + struct pblk_line_meta *lm = &pblk->lm;
>>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>> + struct pblk_line *line;
>>> + int line_id = atomic_read(&l_mg->sysfs_line_state);
>>> + ssize_t sz = 0;
>>> + int i;
>>> +
>>> + if (line_id < 0 || line_id >= l_mg->nr_lines)
>>> + return 0;
>>> +
>>> + sz = snprintf(page, PAGE_SIZE,
>>> + "line\tchunk\tstate\ttype\twear-index\tslba\t\tcnlb\twp\n");
>>> +
>>> + line = &pblk->lines[line_id];
>>> +
>>> + for (i = 0; i < lm->blk_per_line; i++) {
>>> + struct pblk_chunk *chunk = &line->chks[i];
>>> +
>>> + sz += snprintf(page + sz, PAGE_SIZE - sz,
>>> + "%d\t%d\t%d\t%d\t%d\t\t%llu\t\t%llu\t%llu\n",
>>> + line->id, i,
>>> + chunk->state,
>>> + chunk->type,
>>> + chunk->wi,
>>> + chunk->slba,
>>> + chunk->cnlb,
>>> + chunk->wp);
>>> + }
>>> +
>>> + return sz;
>>> +}
>>> +
>>> static ssize_t pblk_sysfs_lines(struct pblk *pblk, char *page)
>>> {
>>> struct nvm_tgt_dev *dev = pblk->dev;
>>> @@ -398,6 +432,29 @@ static ssize_t pblk_sysfs_stats_debug(struct pblk *pblk, char *page)
>>> }
>>> #endif
>>> +
>>> +static ssize_t pblk_sysfs_line_state_store(struct pblk *pblk, const char *page,
>>> + size_t len)
>>> +{
>>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>> + size_t c_len;
>>> + int line_id;
>>> +
>>> + c_len = strcspn(page, "\n");
>>> + if (c_len >= len)
>>> + return -EINVAL;
>>> +
>>> + if (kstrtouint(page, 0, &line_id))
>>> + return -EINVAL;
>>> +
>>> + if (line_id < 0 || line_id >= l_mg->nr_lines)
>>> + return -EINVAL;
>>> +
>>> + atomic_set(&l_mg->sysfs_line_state, line_id);
>>> +
>>> + return len;
>>> +}
>>> +
>>> static ssize_t pblk_sysfs_gc_force(struct pblk *pblk, const char *page,
>>> size_t len)
>>> {
>>> @@ -529,6 +586,11 @@ static struct attribute sys_lines_info_attr = {
>>> .mode = 0444,
>>> };
>>> +static struct attribute sys_line_state_attr = {
>>> + .name = "line_state",
>>> + .mode = 0644,
>>> +};
>>> +
>>> static struct attribute sys_gc_force = {
>>> .name = "gc_force",
>>> .mode = 0200,
>>> @@ -572,6 +634,7 @@ static struct attribute *pblk_attrs[] = {
>>> &sys_stats_ppaf_attr,
>>> &sys_lines_attr,
>>> &sys_lines_info_attr,
>>> + &sys_line_state_attr,
>>> &sys_write_amp_mileage,
>>> &sys_write_amp_trip,
>>> &sys_padding_dist,
>>> @@ -602,6 +665,8 @@ static ssize_t pblk_sysfs_show(struct kobject *kobj, struct attribute *attr,
>>> return pblk_sysfs_lines(pblk, buf);
>>> else if (strcmp(attr->name, "lines_info") == 0)
>>> return pblk_sysfs_lines_info(pblk, buf);
>>> + else if (strcmp(attr->name, "line_state") == 0)
>>> + return pblk_sysfs_line_state_show(pblk, buf);
>>> else if (strcmp(attr->name, "max_sec_per_write") == 0)
>>> return pblk_sysfs_get_sec_per_write(pblk, buf);
>>> else if (strcmp(attr->name, "write_amp_mileage") == 0)
>>> @@ -628,6 +693,8 @@ static ssize_t pblk_sysfs_store(struct kobject *kobj, struct attribute *attr,
>>> return pblk_sysfs_set_sec_per_write(pblk, buf, len);
>>> else if (strcmp(attr->name, "write_amp_trip") == 0)
>>> return pblk_sysfs_set_write_amp_trip(pblk, buf, len);
>>> + else if (strcmp(attr->name, "line_state") == 0)
>>> + return pblk_sysfs_line_state_store(pblk, buf, len);
>>> else if (strcmp(attr->name, "padding_dist") == 0)
>>> return pblk_sysfs_set_padding_dist(pblk, buf, len);
>>> return 0;
>>> diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
>>> index 6e1fcd1a538a..bc31c67b725f 100644
>>> --- a/drivers/lightnvm/pblk.h
>>> +++ b/drivers/lightnvm/pblk.h
>>> @@ -201,6 +201,8 @@ struct pblk_rb {
>>> struct pblk_lun {
>>> struct ppa_addr bppa;
>>> + struct ppa_addr chunk_bppa;
>>> +
>>> struct semaphore wr_sem;
>>> };
>>> @@ -297,6 +299,7 @@ enum {
>>> PBLK_LINETYPE_DATA = 2,
>>> /* Line state */
>>> + PBLK_LINESTATE_NEW = 9,
>>> PBLK_LINESTATE_FREE = 10,
>>> PBLK_LINESTATE_OPEN = 11,
>>> PBLK_LINESTATE_CLOSED = 12,
>>> @@ -412,6 +415,15 @@ struct pblk_smeta {
>>> struct line_smeta *buf; /* smeta buffer in persistent format */
>>> };
>>> +struct pblk_chunk {
>>> + int state;
>>> + int type;
>>> + int wi;
>>> + u64 slba;
>>> + u64 cnlb;
>>> + u64 wp;
>>> +};
>>> +
>>
>> How come the code replicate the nvm_chk_meta data structure?
>
> There is no need to maintain the chunk reserved bits in nvm_chk_meta and
> waste memory in the process, as these are alive for the whole life of
> the pblk instance. This structure only maintains the necessary bits.
>
nvm_chk_meta is 32 bytes:
struct nvm_chk_meta {
u8 state; /* 0 1 */
u8 type; /* 1 1 */
u8 wli; /* 2 1 */
u8 rsvd[5]; /* 3 5 */
u64 slba; /* 8 8 */
u64 cnlb; /* 16 8 */
u64 wp; /* 24 8 */
/* size: 32, cachelines: 1, members: 7 */
/* last cacheline: 32 bytes */
};
pblk_chunk is 40 bytes:
struct pblk_chunk {
int state; /* 0 4 */
int type; /* 4 4 */
int wi; /* 8 4 */
/* XXX 4 bytes hole, try to pack */
u64 slba; /* 16 8 */
u64 cnlb; /* 24 8 */
u64 wp; /* 32 8 */
/* size: 40, cachelines: 1, members: 6 */
/* sum members: 36, holes: 1, sum holes: 4 */
/* last cacheline: 40 bytes */
};
If assuming 64K chunks, pblk chunk uses 2560KB (and requires an extra
traversal when the chunk meta information is returned to copy over the
data to the new data structure), while nvm_chk_meta uses 2048KB.
Granted, the nvm_chk_meta state/type/wli fields will be slightly slower
to access due to the bit shifting, but I think it is worth it. One can
measure and see if there is any performance difference.
^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 17/19] lightnvm: pblk: implement get log report chunk
2018-02-27 18:46 ` Matias Bjørling
@ 2018-02-27 19:50 ` Javier González
0 siblings, 0 replies; 29+ messages in thread
From: Javier González @ 2018-02-27 19:50 UTC (permalink / raw)
To: Matias Bjørling; +Cc: linux-block, linux-kernel, linux-nvme
[-- Attachment #1: Type: text/plain, Size: 22461 bytes --]
> On 27 Feb 2018, at 19.46, Matias Bjørling <mb@lightnvm.io> wrote:
>
> On 02/27/2018 03:40 PM, Javier González wrote:
>>> On 26 Feb 2018, at 20.04, Matias Bjørling <mb@lightnvm.io> wrote:
> <snip>
>>> Can you help me understand why you want to use The
>>> NVM_CHK_ST_HOST_USE? Why would I care if the chunk state is HOST_USE?
>>> A target instance should not be able to see states from other chunks
>>> it doesn't own. and in that case, why have a separate state?
>> The motivation for this state is that pblk does not maintain a per
>> block/chunk state. On the first disk pass however, the state of the
>> chunks is unknown,
>
> I don't understand this. Why is the state unknown? If assuming 1.2
> drive, one can do a scan of first page of each block. If the block
> returns success (it is in use), if 0x42ff, the block is empty, and if
> reading and it returns 0x4281, it is open.
>
> thus we need to check for each one individually;
Yes, we could do this on init - this is actually how we do recovery
after all..
>> after that, we know that good chunks need to be erased. Having this
>> extra chunk state, frees us from maintaining this unnecessary intra-line
>> state. Also, at the point I wrote this code, we did not have the report
>> chunk on the erase path, so updating the chunk metadata would be
>> something periodic.
>> I have taken a look at this again and I solved it in a different way -
>> basically, only maintaining free/closed states, which is not much more
>> expensive. It is included in the next version.
>>>> + }
>>>> +
>>>> + WARN_ONCE(state & NVM_CHK_ST_OPEN,
>>>> + "pblk: open chunk in new line: %d\n",
>>>> + line->id);
>>>> + }
>>>> +
>>>> + return blk_to_erase;
>>>> +}
>>>> +
>>>> static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>>>> {
>>>> struct pblk_line_meta *lm = &pblk->lm;
>>>> - int blk_in_line = atomic_read(&line->blk_in_line);
>>>> + int blk_to_erase;
>>>> line->map_bitmap = kzalloc(lm->sec_bitmap_len, GFP_ATOMIC);
>>>> if (!line->map_bitmap)
>>>> @@ -1110,7 +1184,21 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>>>> return -ENOMEM;
>>>> }
>>>> + /* Bad blocks do not need to be erased */
>>>> + bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
>>>> +
>>>> spin_lock(&line->lock);
>>>> +
>>>> + /* If we have not written to this line, we need to mark up free chunks
>>>> + * as already erased
>>>> + */
>>>> + if (line->state == PBLK_LINESTATE_NEW) {
>>>> + blk_to_erase = pblk_prepare_new_line(pblk, line);
>>>> + line->state = PBLK_LINESTATE_FREE;
>>>> + } else {
>>>> + blk_to_erase = atomic_read(&line->blk_in_line);
>>>> + }
>>>> +
>>>> if (line->state != PBLK_LINESTATE_FREE) {
>>>> kfree(line->map_bitmap);
>>>> kfree(line->invalid_bitmap);
>>>> @@ -1122,15 +1210,12 @@ static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
>>>> line->state = PBLK_LINESTATE_OPEN;
>>>> - atomic_set(&line->left_eblks, blk_in_line);
>>>> - atomic_set(&line->left_seblks, blk_in_line);
>>>> + atomic_set(&line->left_eblks, blk_to_erase);
>>>> + atomic_set(&line->left_seblks, blk_to_erase);
>>>> line->meta_distance = lm->meta_distance;
>>>> spin_unlock(&line->lock);
>>>> - /* Bad blocks do not need to be erased */
>>>> - bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);
>>>> -
>>>> kref_init(&line->ref);
>>>> return 0;
>>>> diff --git a/drivers/lightnvm/pblk-init.c b/drivers/lightnvm/pblk-init.c
>>>> index ec39800eea42..cf4f49d48aed 100644
>>>> --- a/drivers/lightnvm/pblk-init.c
>>>> +++ b/drivers/lightnvm/pblk-init.c
>>>> @@ -401,6 +401,7 @@ static void pblk_line_meta_free(struct pblk_line *line)
>>>> {
>>>> kfree(line->blk_bitmap);
>>>> kfree(line->erase_bitmap);
>>>> + kfree(line->chks);
>>>> }
>>>> static void pblk_lines_free(struct pblk *pblk)
>>>> @@ -440,54 +441,44 @@ static int pblk_bb_get_tbl(struct nvm_tgt_dev *dev, struct pblk_lun *rlun,
>>>> return 0;
>>>> }
>>>> -static void *pblk_bb_get_log(struct pblk *pblk)
>>>> +static void *pblk_bb_get_meta(struct pblk *pblk)
>>>> {
>>>> struct nvm_tgt_dev *dev = pblk->dev;
>>>> struct nvm_geo *geo = &dev->geo;
>>>> - u8 *log;
>>>> + u8 *meta;
>>>> int i, nr_blks, blk_per_lun;
>>>> int ret;
>>>> blk_per_lun = geo->c.num_chk * geo->c.pln_mode;
>>>> nr_blks = blk_per_lun * geo->all_luns;
>>>> - log = kmalloc(nr_blks, GFP_KERNEL);
>>>> - if (!log)
>>>> + meta = kmalloc(nr_blks, GFP_KERNEL);
>>>> + if (!meta)
>>>> return ERR_PTR(-ENOMEM);
>>>> for (i = 0; i < geo->all_luns; i++) {
>>>> struct pblk_lun *rlun = &pblk->luns[i];
>>>> - u8 *log_pos = log + i * blk_per_lun;
>>>> + u8 *meta_pos = meta + i * blk_per_lun;
>>>> - ret = pblk_bb_get_tbl(dev, rlun, log_pos, blk_per_lun);
>>>> + ret = pblk_bb_get_tbl(dev, rlun, meta_pos, blk_per_lun);
>>>> if (ret) {
>>>> - kfree(log);
>>>> + kfree(meta);
>>>> return ERR_PTR(-EIO);
>>>> }
>>>> }
>>>> - return log;
>>>> + return meta;
>>>> }
>>>> -static int pblk_bb_line(struct pblk *pblk, struct pblk_line *line,
>>>> - u8 *bb_log, int blk_per_line)
>>>> +static void *pblk_chunk_get_meta(struct pblk *pblk)
>>>> {
>>>> struct nvm_tgt_dev *dev = pblk->dev;
>>>> struct nvm_geo *geo = &dev->geo;
>>>> - int i, bb_cnt = 0;
>>>> - for (i = 0; i < blk_per_line; i++) {
>>>> - struct pblk_lun *rlun = &pblk->luns[i];
>>>> - u8 *lun_bb_log = bb_log + i * blk_per_line;
>>>> -
>>>> - if (lun_bb_log[line->id] == NVM_BLK_T_FREE)
>>>> - continue;
>>>> -
>>>> - set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
>>>> - bb_cnt++;
>>>> - }
>>>> -
>>>> - return bb_cnt;
>>>> + if (geo->c.version == NVM_OCSSD_SPEC_12)
>>>> + return pblk_bb_get_meta(pblk);
>>>> + else
>>>> + return pblk_chunk_get_info(pblk);
>>>
>>> This 1.2 or 2.0 would be nice to have inside the lightnvm core. A
>>> target should not care weather it is a 1.2 or 2.0 device.
>> It is not an easy thing to do since the bad block format and chunk
>> report return completely different formats. As I explained in a
>> different thread, doing this in core would force 1.2 to understand fake
>> chunk formats, which are not meaningful. It is better to leave the logic
>> at the target level, where we know what to do with the format.
>
> What if the core layer exported a generic "chunk meta" interface, and
> then filled it as is done in pblk. In the core, it will then fill out
> what it can fill out, and leave the rest of the fields untouched. The
> pblk code can then if () in certain places if for example wp attribute
> can,'t be used?
>
It is basically what we do in pblk now.
It is easy to abstract in core; I'm just concerned that hiding the chunk
metadata will end up making the pblk code more complex - in the end
doing a check on the version is meaningful for the reader.
> In the interest of time, and to get this moving, we can ignore it for now.
Ok. Let's do it this way now and we can see how to abstract it when we
start implementing features based on these fields. The first one will be
wear levelling, which I expect to have for next window.
>>>> }
>>>> static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>>>> @@ -516,6 +507,7 @@ static int pblk_luns_init(struct pblk *pblk, struct ppa_addr *luns)
>>>> rlun = &pblk->luns[i];
>>>> rlun->bppa = luns[lunid];
>>>> + rlun->chunk_bppa = luns[i];
>>>> sema_init(&rlun->wr_sem, 1);
>>>> }
>>>> @@ -695,8 +687,125 @@ static int pblk_lines_alloc_metadata(struct pblk *pblk)
>>>> return -ENOMEM;
>>>> }
>>>> -static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>>>> - void *chunk_log, long *nr_bad_blks)
>>>> +static int pblk_setup_line_meta_12(struct pblk *pblk, struct pblk_line *line,
>>>> + void *chunk_meta)
>>>> +{
>>>> + struct nvm_tgt_dev *dev = pblk->dev;
>>>> + struct nvm_geo *geo = &dev->geo;
>>>> + struct pblk_line_meta *lm = &pblk->lm;
>>>> + int i, chk_per_lun, nr_bad_chks = 0;
>>>> +
>>>> + chk_per_lun = geo->c.num_chk * geo->c.pln_mode;
>>>> +
>>>> + for (i = 0; i < lm->blk_per_line; i++) {
>>>> + struct pblk_chunk *chunk = &line->chks[i];
>>>> + struct pblk_lun *rlun = &pblk->luns[i];
>>>> + u8 *lun_bb_meta = chunk_meta + i * chk_per_lun;
>>>> +
>>>> + /*
>>>> + * In 1.2 spec. chunk state is not persisted by the device. Thus
>>>> + * some of the values are reset each time pblk is instantiated.
>>>> + */
>>>> + if (lun_bb_meta[line->id] == NVM_BLK_T_FREE)
>>>> + chunk->state = NVM_CHK_ST_HOST_USE;
>>>> + else
>>>> + chunk->state = NVM_CHK_ST_OFFLINE;
>>>> +
>>>> + chunk->type = NVM_CHK_TP_W_SEQ;
>>>> + chunk->wi = 0;
>>>> + chunk->slba = -1;
>>>> + chunk->cnlb = geo->c.clba;
>>>> + chunk->wp = 0;
>>>> +
>>>> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
>>>> + continue;
>>>> +
>>>> + set_bit(pblk_ppa_to_pos(geo, rlun->bppa), line->blk_bitmap);
>>>> + nr_bad_chks++;
>>>> + }
>>>> +
>>>> + return nr_bad_chks;
>>>> +}
>>>> +
>>>> +static int pblk_setup_line_meta_20(struct pblk *pblk, struct pblk_line *line,
>>>> + struct nvm_chk_meta *meta)
>>>> +{
>>>> + struct nvm_tgt_dev *dev = pblk->dev;
>>>> + struct nvm_geo *geo = &dev->geo;
>>>> + struct pblk_line_meta *lm = &pblk->lm;
>>>> + int i, nr_bad_chks = 0;
>>>> +
>>>> + for (i = 0; i < lm->blk_per_line; i++) {
>>>> + struct pblk_chunk *chunk = &line->chks[i];
>>>> + struct pblk_lun *rlun = &pblk->luns[i];
>>>> + struct nvm_chk_meta *chunk_meta;
>>>> + struct ppa_addr ppa;
>>>> +
>>>> + ppa = rlun->chunk_bppa;
>>>> + ppa.m.chk = line->id;
>>>> + chunk_meta = pblk_chunk_get_off(pblk, meta, ppa);
>>>> +
>>>> + chunk->state = chunk_meta->state;
>>>> + chunk->type = chunk_meta->type;
>>>> + chunk->wi = chunk_meta->wi;
>>>> + chunk->slba = chunk_meta->slba;
>>>> + chunk->cnlb = chunk_meta->cnlb;
>>>> + chunk->wp = chunk_meta->wp;
>>>> +
>>>> + if (!(chunk->state & NVM_CHK_ST_OFFLINE))
>>>> + continue;
>>>> +
>>>> + if (chunk->type & NVM_CHK_TP_SZ_SPEC) {
>>>> + WARN_ONCE(1, "pblk: custom-sized chunks unsupported\n");
>>>> + continue;
>>>> + }
>>>> +
>>>> + set_bit(pblk_ppa_to_pos(geo, rlun->chunk_bppa),
>>>> + line->blk_bitmap);
>>>> + nr_bad_chks++;
>>>> + }
>>>> +
>>>> + return nr_bad_chks;
>>>> +}
>>>> +
>>>> +static long pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>>>> + void *chunk_meta, int line_id)
>>>> +{
>>>> + struct nvm_tgt_dev *dev = pblk->dev;
>>>> + struct nvm_geo *geo = &dev->geo;
>>>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>>> + struct pblk_line_meta *lm = &pblk->lm;
>>>> + long nr_bad_chks, chk_in_line;
>>>> +
>>>> + line->pblk = pblk;
>>>> + line->id = line_id;
>>>> + line->type = PBLK_LINETYPE_FREE;
>>>> + line->state = PBLK_LINESTATE_NEW;
>>>> + line->gc_group = PBLK_LINEGC_NONE;
>>>> + line->vsc = &l_mg->vsc_list[line_id];
>>>> + spin_lock_init(&line->lock);
>>>> +
>>>> + if (geo->c.version == NVM_OCSSD_SPEC_12)
>>>> + nr_bad_chks = pblk_setup_line_meta_12(pblk, line, chunk_meta);
>>>> + else
>>>> + nr_bad_chks = pblk_setup_line_meta_20(pblk, line, chunk_meta);
>>>> +
>>>> + chk_in_line = lm->blk_per_line - nr_bad_chks;
>>>> + if (nr_bad_chks < 0 || nr_bad_chks > lm->blk_per_line ||
>>>> + chk_in_line < lm->min_blk_line) {
>>>> + line->state = PBLK_LINESTATE_BAD;
>>>> + list_add_tail(&line->list, &l_mg->bad_list);
>>>> + return 0;
>>>> + }
>>>> +
>>>> + atomic_set(&line->blk_in_line, chk_in_line);
>>>> + list_add_tail(&line->list, &l_mg->free_list);
>>>> + l_mg->nr_free_lines++;
>>>> +
>>>> + return chk_in_line;
>>>> +}
>>>> +
>>>> +static int pblk_alloc_line_meta(struct pblk *pblk, struct pblk_line *line)
>>>> {
>>>> struct pblk_line_meta *lm = &pblk->lm;
>>>> @@ -710,7 +819,13 @@ static int pblk_setup_line_meta(struct pblk *pblk, struct pblk_line *line,
>>>> return -ENOMEM;
>>>> }
>>>> - *nr_bad_blks = pblk_bb_line(pblk, line, chunk_log, lm->blk_per_line);
>>>> + line->chks = kmalloc(lm->blk_per_line * sizeof(struct pblk_chunk),
>>>> + GFP_KERNEL);
>>>> + if (!line->chks) {
>>>> + kfree(line->erase_bitmap);
>>>> + kfree(line->blk_bitmap);
>>>> + return -ENOMEM;
>>>> + }
>>>> return 0;
>>>> }
>>>> @@ -722,9 +837,9 @@ static int pblk_lines_init(struct pblk *pblk)
>>>> struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>>> struct pblk_line_meta *lm = &pblk->lm;
>>>> struct pblk_line *line;
>>>> - void *chunk_log;
>>>> + void *chunk_meta;
>>>> unsigned int smeta_len, emeta_len;
>>>> - long nr_bad_blks = 0, nr_free_blks = 0;
>>>> + long nr_free_chks = 0;
>>>> int bb_distance, max_write_ppas;
>>>> int i, ret;
>>>> @@ -743,6 +858,7 @@ static int pblk_lines_init(struct pblk *pblk)
>>>> l_mg->log_line = l_mg->data_line = NULL;
>>>> l_mg->l_seq_nr = l_mg->d_seq_nr = 0;
>>>> l_mg->nr_free_lines = 0;
>>>> + atomic_set(&l_mg->sysfs_line_state, -1);
>>>> bitmap_zero(&l_mg->meta_bitmap, PBLK_DATA_LINES);
>>>> lm->sec_per_line = geo->c.clba * geo->all_luns;
>>>> @@ -841,53 +957,31 @@ static int pblk_lines_init(struct pblk *pblk)
>>>> goto fail_free_bb_aux;
>>>> }
>>>> - chunk_log = pblk_bb_get_log(pblk);
>>>> - if (IS_ERR(chunk_log)) {
>>>> - pr_err("pblk: could not get bad block log (%lu)\n",
>>>> - PTR_ERR(chunk_log));
>>>> - ret = PTR_ERR(chunk_log);
>>>> + chunk_meta = pblk_chunk_get_meta(pblk);
>>>> + if (IS_ERR(chunk_meta)) {
>>>> + pr_err("pblk: could not get chunk log (%lu)\n",
>>>> + PTR_ERR(chunk_meta));
>>>> + ret = PTR_ERR(chunk_meta);
>>>> goto fail_free_lines;
>>>> }
>>>> for (i = 0; i < l_mg->nr_lines; i++) {
>>>> - int chk_in_line;
>>>> -
>>>> line = &pblk->lines[i];
>>>> - line->pblk = pblk;
>>>> - line->id = i;
>>>> - line->type = PBLK_LINETYPE_FREE;
>>>> - line->state = PBLK_LINESTATE_FREE;
>>>> - line->gc_group = PBLK_LINEGC_NONE;
>>>> - line->vsc = &l_mg->vsc_list[i];
>>>> - spin_lock_init(&line->lock);
>>>> -
>>>> - ret = pblk_setup_line_meta(pblk, line, chunk_log, &nr_bad_blks);
>>>> + ret = pblk_alloc_line_meta(pblk, line);
>>>> if (ret)
>>>> - goto fail_free_chunk_log;
>>>> + goto fail_free_chunk_meta;
>>>> - chk_in_line = lm->blk_per_line - nr_bad_blks;
>>>> - if (nr_bad_blks < 0 || nr_bad_blks > lm->blk_per_line ||
>>>> - chk_in_line < lm->min_blk_line) {
>>>> - line->state = PBLK_LINESTATE_BAD;
>>>> - list_add_tail(&line->list, &l_mg->bad_list);
>>>> - continue;
>>>> - }
>>>> -
>>>> - nr_free_blks += chk_in_line;
>>>> - atomic_set(&line->blk_in_line, chk_in_line);
>>>> -
>>>> - l_mg->nr_free_lines++;
>>>> - list_add_tail(&line->list, &l_mg->free_list);
>>>> + nr_free_chks += pblk_setup_line_meta(pblk, line, chunk_meta, i);
>>>> }
>>>> - pblk_set_provision(pblk, nr_free_blks);
>>>> + pblk_set_provision(pblk, nr_free_chks);
>>>> - kfree(chunk_log);
>>>> + kfree(chunk_meta);
>>>> return 0;
>>>> -fail_free_chunk_log:
>>>> - kfree(chunk_log);
>>>> +fail_free_chunk_meta:
>>>> + kfree(chunk_meta);
>>>> while (--i >= 0)
>>>> pblk_line_meta_free(&pblk->lines[i]);
>>>> fail_free_lines:
>>>> diff --git a/drivers/lightnvm/pblk-sysfs.c b/drivers/lightnvm/pblk-sysfs.c
>>>> index ccfb3abd2773..1ce5b956c622 100644
>>>> --- a/drivers/lightnvm/pblk-sysfs.c
>>>> +++ b/drivers/lightnvm/pblk-sysfs.c
>>>> @@ -142,6 +142,40 @@ static ssize_t pblk_sysfs_ppaf(struct pblk *pblk, char *page)
>>>> return sz;
>>>> }
>>>> +static ssize_t pblk_sysfs_line_state_show(struct pblk *pblk, char *page)
>>>> +{
>>>> + struct pblk_line_meta *lm = &pblk->lm;
>>>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>>> + struct pblk_line *line;
>>>> + int line_id = atomic_read(&l_mg->sysfs_line_state);
>>>> + ssize_t sz = 0;
>>>> + int i;
>>>> +
>>>> + if (line_id < 0 || line_id >= l_mg->nr_lines)
>>>> + return 0;
>>>> +
>>>> + sz = snprintf(page, PAGE_SIZE,
>>>> + "line\tchunk\tstate\ttype\twear-index\tslba\t\tcnlb\twp\n");
>>>> +
>>>> + line = &pblk->lines[line_id];
>>>> +
>>>> + for (i = 0; i < lm->blk_per_line; i++) {
>>>> + struct pblk_chunk *chunk = &line->chks[i];
>>>> +
>>>> + sz += snprintf(page + sz, PAGE_SIZE - sz,
>>>> + "%d\t%d\t%d\t%d\t%d\t\t%llu\t\t%llu\t%llu\n",
>>>> + line->id, i,
>>>> + chunk->state,
>>>> + chunk->type,
>>>> + chunk->wi,
>>>> + chunk->slba,
>>>> + chunk->cnlb,
>>>> + chunk->wp);
>>>> + }
>>>> +
>>>> + return sz;
>>>> +}
>>>> +
>>>> static ssize_t pblk_sysfs_lines(struct pblk *pblk, char *page)
>>>> {
>>>> struct nvm_tgt_dev *dev = pblk->dev;
>>>> @@ -398,6 +432,29 @@ static ssize_t pblk_sysfs_stats_debug(struct pblk *pblk, char *page)
>>>> }
>>>> #endif
>>>> +
>>>> +static ssize_t pblk_sysfs_line_state_store(struct pblk *pblk, const char *page,
>>>> + size_t len)
>>>> +{
>>>> + struct pblk_line_mgmt *l_mg = &pblk->l_mg;
>>>> + size_t c_len;
>>>> + int line_id;
>>>> +
>>>> + c_len = strcspn(page, "\n");
>>>> + if (c_len >= len)
>>>> + return -EINVAL;
>>>> +
>>>> + if (kstrtouint(page, 0, &line_id))
>>>> + return -EINVAL;
>>>> +
>>>> + if (line_id < 0 || line_id >= l_mg->nr_lines)
>>>> + return -EINVAL;
>>>> +
>>>> + atomic_set(&l_mg->sysfs_line_state, line_id);
>>>> +
>>>> + return len;
>>>> +}
>>>> +
>>>> static ssize_t pblk_sysfs_gc_force(struct pblk *pblk, const char *page,
>>>> size_t len)
>>>> {
>>>> @@ -529,6 +586,11 @@ static struct attribute sys_lines_info_attr = {
>>>> .mode = 0444,
>>>> };
>>>> +static struct attribute sys_line_state_attr = {
>>>> + .name = "line_state",
>>>> + .mode = 0644,
>>>> +};
>>>> +
>>>> static struct attribute sys_gc_force = {
>>>> .name = "gc_force",
>>>> .mode = 0200,
>>>> @@ -572,6 +634,7 @@ static struct attribute *pblk_attrs[] = {
>>>> &sys_stats_ppaf_attr,
>>>> &sys_lines_attr,
>>>> &sys_lines_info_attr,
>>>> + &sys_line_state_attr,
>>>> &sys_write_amp_mileage,
>>>> &sys_write_amp_trip,
>>>> &sys_padding_dist,
>>>> @@ -602,6 +665,8 @@ static ssize_t pblk_sysfs_show(struct kobject *kobj, struct attribute *attr,
>>>> return pblk_sysfs_lines(pblk, buf);
>>>> else if (strcmp(attr->name, "lines_info") == 0)
>>>> return pblk_sysfs_lines_info(pblk, buf);
>>>> + else if (strcmp(attr->name, "line_state") == 0)
>>>> + return pblk_sysfs_line_state_show(pblk, buf);
>>>> else if (strcmp(attr->name, "max_sec_per_write") == 0)
>>>> return pblk_sysfs_get_sec_per_write(pblk, buf);
>>>> else if (strcmp(attr->name, "write_amp_mileage") == 0)
>>>> @@ -628,6 +693,8 @@ static ssize_t pblk_sysfs_store(struct kobject *kobj, struct attribute *attr,
>>>> return pblk_sysfs_set_sec_per_write(pblk, buf, len);
>>>> else if (strcmp(attr->name, "write_amp_trip") == 0)
>>>> return pblk_sysfs_set_write_amp_trip(pblk, buf, len);
>>>> + else if (strcmp(attr->name, "line_state") == 0)
>>>> + return pblk_sysfs_line_state_store(pblk, buf, len);
>>>> else if (strcmp(attr->name, "padding_dist") == 0)
>>>> return pblk_sysfs_set_padding_dist(pblk, buf, len);
>>>> return 0;
>>>> diff --git a/drivers/lightnvm/pblk.h b/drivers/lightnvm/pblk.h
>>>> index 6e1fcd1a538a..bc31c67b725f 100644
>>>> --- a/drivers/lightnvm/pblk.h
>>>> +++ b/drivers/lightnvm/pblk.h
>>>> @@ -201,6 +201,8 @@ struct pblk_rb {
>>>> struct pblk_lun {
>>>> struct ppa_addr bppa;
>>>> + struct ppa_addr chunk_bppa;
>>>> +
>>>> struct semaphore wr_sem;
>>>> };
>>>> @@ -297,6 +299,7 @@ enum {
>>>> PBLK_LINETYPE_DATA = 2,
>>>> /* Line state */
>>>> + PBLK_LINESTATE_NEW = 9,
>>>> PBLK_LINESTATE_FREE = 10,
>>>> PBLK_LINESTATE_OPEN = 11,
>>>> PBLK_LINESTATE_CLOSED = 12,
>>>> @@ -412,6 +415,15 @@ struct pblk_smeta {
>>>> struct line_smeta *buf; /* smeta buffer in persistent format */
>>>> };
>>>> +struct pblk_chunk {
>>>> + int state;
>>>> + int type;
>>>> + int wi;
>>>> + u64 slba;
>>>> + u64 cnlb;
>>>> + u64 wp;
>>>> +};
>>>> +
>>>
>>> How come the code replicate the nvm_chk_meta data structure?
>> There is no need to maintain the chunk reserved bits in nvm_chk_meta and
>> waste memory in the process, as these are alive for the whole life of
>> the pblk instance. This structure only maintains the necessary bits.
>
> nvm_chk_meta is 32 bytes:
>
> struct nvm_chk_meta {
> u8 state; /* 0 1 */
> u8 type; /* 1 1 */
> u8 wli; /* 2 1 */
> u8 rsvd[5]; /* 3 5 */
> u64 slba; /* 8 8 */
> u64 cnlb; /* 16 8 */
> u64 wp; /* 24 8 */
>
> /* size: 32, cachelines: 1, members: 7 */
> /* last cacheline: 32 bytes */
> };
>
> pblk_chunk is 40 bytes:
>
> struct pblk_chunk {
> int state; /* 0 4 */
> int type; /* 4 4 */
> int wi; /* 8 4 */
>
> /* XXX 4 bytes hole, try to pack */
>
> u64 slba; /* 16 8 */
> u64 cnlb; /* 24 8 */
> u64 wp; /* 32 8 */
>
> /* size: 40, cachelines: 1, members: 6 */
> /* sum members: 36, holes: 1, sum holes: 4 */
> /* last cacheline: 40 bytes */
> };
>
> If assuming 64K chunks, pblk chunk uses 2560KB (and requires an extra
> traversal when the chunk meta information is returned to copy over the
> data to the new data structure), while nvm_chk_meta uses 2048KB.
> Granted, the nvm_chk_meta state/type/wli fields will be slightly
> slower to access due to the bit shifting, but I think it is worth it.
> One can measure and see if there is any performance difference.
>
Ok. This is not in the fast path now. We can look into it if it gets
there at some point.
I'll fix in the next version.
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 29+ messages in thread
end of thread, other threads:[~2018-02-27 19:50 UTC | newest]
Thread overview: 29+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-02-26 13:16 [PATCH V3 00/19] lightnvm: pblk: implement 2.0 support Javier González
2018-02-26 13:17 ` [PATCH 01/19] lightnvm: simplify geometry structure Javier González
2018-02-26 13:17 ` [PATCH 02/19] lightnvm: add controller capabilities to 2.0 Javier González
2018-02-26 13:17 ` [PATCH 03/19] lightnvm: add minor version to generic geometry Javier González
2018-02-26 13:17 ` [PATCH 04/19] lightnvm: add shorten OCSSD version in geo Javier González
2018-02-26 13:17 ` [PATCH 05/19] lightnvm: complete geo structure with maxoc* Javier González
2018-02-26 13:17 ` [PATCH 06/19] lightnvm: pblk: check for supported version Javier González
2018-02-26 13:17 ` [PATCH 07/19] lightnvm: complete 2.0 values in sysfs Javier González
2018-02-26 13:17 ` [PATCH 08/19] lightnvm: rename number of channels and luns Javier González
2018-02-26 13:17 ` [PATCH 09/19] lightnvm: rename sect_* to sec_* Javier González
2018-02-26 13:17 ` [PATCH 10/19] lightnvm: add support for 2.0 address format Javier González
2018-02-26 13:17 ` [PATCH 11/19] lightnvm: pblk: rename ppaf* to addrf* Javier González
2018-02-26 13:17 ` [PATCH 12/19] lightnvn: pblk: use generic address format Javier González
2018-02-26 13:17 ` [PATCH 13/19] lightnvm: make address conversions depend on generic device Javier González
2018-02-26 13:17 ` [PATCH 14/19] nvme: make nvme_get_log_ext available Javier González
2018-02-26 13:17 ` [PATCH 15/19] lightnvm: implement get log report chunk helpers Javier González
2018-02-26 13:17 ` [PATCH 16/19] lightnvm: define chunk states Javier González
2018-02-26 13:17 ` [PATCH 17/19] lightnvm: pblk: implement get log report chunk Javier González
2018-02-26 19:04 ` Matias Bjørling
2018-02-27 14:40 ` Javier González
2018-02-27 18:46 ` Matias Bjørling
2018-02-27 19:50 ` Javier González
2018-02-26 13:17 ` [PATCH 18/19] lightnvm: pblk: refactor init/exit sequences Javier González
2018-02-26 13:17 ` [PATCH 19/19] lightnvm: pblk: implement 2.0 support Javier González
2018-02-26 18:19 ` [PATCH V3 00/19] " Matias Bjørling
2018-02-26 18:21 ` Javier Gonzalez
2018-02-26 18:24 ` Matias Bjørling
2018-02-26 18:27 ` Javier Gonzalez
2018-02-26 18:33 ` Matias Bjørling
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).