Commit 22c11858 authored by Mike Snitzer's avatar Mike Snitzer

dm: introduce DM_TYPE_NVME_BIO_BASED

If dm_table_determine_type() establishes DM_TYPE_NVME_BIO_BASED then
all devices in the DM table do not support partial completions.  Also,
the table has a single immutable target that doesn't require DM core to
split bios.

This will enable adding NVMe optimizations to bio-based DM.
Signed-off-by: default avatarMike Snitzer <snitzer@redhat.com>
parent f3986374
...@@ -866,7 +866,8 @@ EXPORT_SYMBOL(dm_consume_args); ...@@ -866,7 +866,8 @@ EXPORT_SYMBOL(dm_consume_args);
static bool __table_type_bio_based(enum dm_queue_mode table_type) static bool __table_type_bio_based(enum dm_queue_mode table_type)
{ {
return (table_type == DM_TYPE_BIO_BASED || return (table_type == DM_TYPE_BIO_BASED ||
table_type == DM_TYPE_DAX_BIO_BASED); table_type == DM_TYPE_DAX_BIO_BASED ||
table_type == DM_TYPE_NVME_BIO_BASED);
} }
static bool __table_type_request_based(enum dm_queue_mode table_type) static bool __table_type_request_based(enum dm_queue_mode table_type)
...@@ -909,6 +910,8 @@ static bool dm_table_supports_dax(struct dm_table *t) ...@@ -909,6 +910,8 @@ static bool dm_table_supports_dax(struct dm_table *t)
return true; return true;
} }
static bool dm_table_does_not_support_partial_completion(struct dm_table *t);
static int dm_table_determine_type(struct dm_table *t) static int dm_table_determine_type(struct dm_table *t)
{ {
unsigned i; unsigned i;
...@@ -923,6 +926,14 @@ static int dm_table_determine_type(struct dm_table *t) ...@@ -923,6 +926,14 @@ static int dm_table_determine_type(struct dm_table *t)
/* target already set the table's type */ /* target already set the table's type */
if (t->type == DM_TYPE_BIO_BASED) if (t->type == DM_TYPE_BIO_BASED)
return 0; return 0;
else if (t->type == DM_TYPE_NVME_BIO_BASED) {
if (!dm_table_does_not_support_partial_completion(t)) {
DMERR("nvme bio-based is only possible with devices"
" that don't support partial completion");
return -EINVAL;
}
/* Fallthru, also verify all devices are blk-mq */
}
BUG_ON(t->type == DM_TYPE_DAX_BIO_BASED); BUG_ON(t->type == DM_TYPE_DAX_BIO_BASED);
goto verify_rq_based; goto verify_rq_based;
} }
...@@ -937,8 +948,8 @@ static int dm_table_determine_type(struct dm_table *t) ...@@ -937,8 +948,8 @@ static int dm_table_determine_type(struct dm_table *t)
bio_based = 1; bio_based = 1;
if (bio_based && request_based) { if (bio_based && request_based) {
DMWARN("Inconsistent table: different target types" DMERR("Inconsistent table: different target types"
" can't be mixed up"); " can't be mixed up");
return -EINVAL; return -EINVAL;
} }
} }
...@@ -959,8 +970,14 @@ static int dm_table_determine_type(struct dm_table *t) ...@@ -959,8 +970,14 @@ static int dm_table_determine_type(struct dm_table *t)
/* We must use this table as bio-based */ /* We must use this table as bio-based */
t->type = DM_TYPE_BIO_BASED; t->type = DM_TYPE_BIO_BASED;
if (dm_table_supports_dax(t) || if (dm_table_supports_dax(t) ||
(list_empty(devices) && live_md_type == DM_TYPE_DAX_BIO_BASED)) (list_empty(devices) && live_md_type == DM_TYPE_DAX_BIO_BASED)) {
t->type = DM_TYPE_DAX_BIO_BASED; t->type = DM_TYPE_DAX_BIO_BASED;
} else if ((dm_table_get_immutable_target(t) &&
dm_table_does_not_support_partial_completion(t)) ||
(list_empty(devices) && live_md_type == DM_TYPE_NVME_BIO_BASED)) {
t->type = DM_TYPE_NVME_BIO_BASED;
goto verify_rq_based;
}
return 0; return 0;
} }
...@@ -980,7 +997,8 @@ static int dm_table_determine_type(struct dm_table *t) ...@@ -980,7 +997,8 @@ static int dm_table_determine_type(struct dm_table *t)
* (e.g. request completion process for partial completion.) * (e.g. request completion process for partial completion.)
*/ */
if (t->num_targets > 1) { if (t->num_targets > 1) {
DMWARN("Request-based dm doesn't support multiple targets yet"); DMERR("%s DM doesn't support multiple targets",
t->type == DM_TYPE_NVME_BIO_BASED ? "nvme bio-based" : "request-based");
return -EINVAL; return -EINVAL;
} }
...@@ -997,6 +1015,15 @@ static int dm_table_determine_type(struct dm_table *t) ...@@ -997,6 +1015,15 @@ static int dm_table_determine_type(struct dm_table *t)
return 0; return 0;
} }
tgt = dm_table_get_immutable_target(t);
if (!tgt) {
DMERR("table load rejected: immutable target is required");
return -EINVAL;
} else if (tgt->max_io_len) {
DMERR("table load rejected: immutable target that splits IO is not supported");
return -EINVAL;
}
/* Non-request-stackable devices can't be used for request-based dm */ /* Non-request-stackable devices can't be used for request-based dm */
list_for_each_entry(dd, devices, list) { list_for_each_entry(dd, devices, list) {
struct request_queue *q = bdev_get_queue(dd->dm_dev->bdev); struct request_queue *q = bdev_get_queue(dd->dm_dev->bdev);
...@@ -1018,7 +1045,8 @@ static int dm_table_determine_type(struct dm_table *t) ...@@ -1018,7 +1045,8 @@ static int dm_table_determine_type(struct dm_table *t)
} }
t->all_blk_mq = mq_count > 0; t->all_blk_mq = mq_count > 0;
if (t->type == DM_TYPE_MQ_REQUEST_BASED && !t->all_blk_mq) { if (!t->all_blk_mq &&
(t->type == DM_TYPE_MQ_REQUEST_BASED || t->type == DM_TYPE_NVME_BIO_BASED)) {
DMERR("table load rejected: all devices are not blk-mq request-stackable"); DMERR("table load rejected: all devices are not blk-mq request-stackable");
return -EINVAL; return -EINVAL;
} }
...@@ -1708,6 +1736,20 @@ static bool dm_table_all_devices_attribute(struct dm_table *t, ...@@ -1708,6 +1736,20 @@ static bool dm_table_all_devices_attribute(struct dm_table *t,
return true; return true;
} }
static int device_no_partial_completion(struct dm_target *ti, struct dm_dev *dev,
sector_t start, sector_t len, void *data)
{
char b[BDEVNAME_SIZE];
/* For now, NVMe devices are the only devices of this class */
return (strncmp(bdevname(dev->bdev, b), "nvme", 3) == 0);
}
static bool dm_table_does_not_support_partial_completion(struct dm_table *t)
{
return dm_table_all_devices_attribute(t, device_no_partial_completion);
}
static int device_not_write_same_capable(struct dm_target *ti, struct dm_dev *dev, static int device_not_write_same_capable(struct dm_target *ti, struct dm_dev *dev,
sector_t start, sector_t len, void *data) sector_t start, sector_t len, void *data)
{ {
......
...@@ -2073,6 +2073,7 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t) ...@@ -2073,6 +2073,7 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t)
break; break;
case DM_TYPE_BIO_BASED: case DM_TYPE_BIO_BASED:
case DM_TYPE_DAX_BIO_BASED: case DM_TYPE_DAX_BIO_BASED:
case DM_TYPE_NVME_BIO_BASED:
dm_init_normal_md_queue(md); dm_init_normal_md_queue(md);
blk_queue_make_request(md->queue, dm_make_request); blk_queue_make_request(md->queue, dm_make_request);
break; break;
...@@ -2780,6 +2781,7 @@ struct dm_md_mempools *dm_alloc_md_mempools(struct mapped_device *md, enum dm_qu ...@@ -2780,6 +2781,7 @@ struct dm_md_mempools *dm_alloc_md_mempools(struct mapped_device *md, enum dm_qu
switch (type) { switch (type) {
case DM_TYPE_BIO_BASED: case DM_TYPE_BIO_BASED:
case DM_TYPE_DAX_BIO_BASED: case DM_TYPE_DAX_BIO_BASED:
case DM_TYPE_NVME_BIO_BASED:
pool_size = max(dm_get_reserved_bio_based_ios(), min_pool_size); pool_size = max(dm_get_reserved_bio_based_ios(), min_pool_size);
front_pad = roundup(per_io_data_size, __alignof__(struct dm_target_io)) + offsetof(struct dm_target_io, clone); front_pad = roundup(per_io_data_size, __alignof__(struct dm_target_io)) + offsetof(struct dm_target_io, clone);
io_front_pad = roundup(front_pad, __alignof__(struct dm_io)) + offsetof(struct dm_io, tio); io_front_pad = roundup(front_pad, __alignof__(struct dm_io)) + offsetof(struct dm_io, tio);
......
...@@ -28,6 +28,7 @@ enum dm_queue_mode { ...@@ -28,6 +28,7 @@ enum dm_queue_mode {
DM_TYPE_REQUEST_BASED = 2, DM_TYPE_REQUEST_BASED = 2,
DM_TYPE_MQ_REQUEST_BASED = 3, DM_TYPE_MQ_REQUEST_BASED = 3,
DM_TYPE_DAX_BIO_BASED = 4, DM_TYPE_DAX_BIO_BASED = 4,
DM_TYPE_NVME_BIO_BASED = 5,
}; };
typedef enum { STATUSTYPE_INFO, STATUSTYPE_TABLE } status_type_t; typedef enum { STATUSTYPE_INFO, STATUSTYPE_TABLE } status_type_t;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment