From: Hannes Reinecke <hare@suse.de>
To: Mike Snitzer <snitzer@redhat.com>
Cc: Damien LeMoal <damien.lemoal@wdc.com>, dm-devel@redhat.com
Subject: [PATCH 12/14] dm-zoned: allocate zone by device index
Date: Fri, 29 May 2020 19:39:05 +0200 [thread overview]
Message-ID: <20200529173907.40529-13-hare@suse.de> (raw)
In-Reply-To: <20200529173907.40529-1-hare@suse.de>
When allocating a zone we should pass in an indicator on which
device the zone should be allocated; this increases performance
for a multi-device setup as then reclaim can allocate zones on
the device for which reclaim is running.
Signed-off-by: Hannes Reinecke <hare@suse.de>
---
drivers/md/dm-zoned-metadata.c | 17 +++++++++++------
drivers/md/dm-zoned-reclaim.c | 3 ++-
drivers/md/dm-zoned.h | 3 ++-
3 files changed, 15 insertions(+), 8 deletions(-)
diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
index 221163ae5f68..7f46b2ea554c 100644
--- a/drivers/md/dm-zoned-metadata.c
+++ b/drivers/md/dm-zoned-metadata.c
@@ -2048,7 +2048,7 @@ struct dm_zone *dmz_get_chunk_mapping(struct dmz_metadata *zmd, unsigned int chu
goto out;
/* Allocate a random zone */
- dzone = dmz_alloc_zone(zmd, alloc_flags);
+ dzone = dmz_alloc_zone(zmd, 0, alloc_flags);
if (!dzone) {
if (dmz_dev_is_dying(zmd)) {
dzone = ERR_PTR(-EIO);
@@ -2154,7 +2154,7 @@ struct dm_zone *dmz_get_chunk_buffer(struct dmz_metadata *zmd,
goto out;
/* Allocate a random zone */
- bzone = dmz_alloc_zone(zmd, alloc_flags);
+ bzone = dmz_alloc_zone(zmd, 0, alloc_flags);
if (!bzone) {
if (dmz_dev_is_dying(zmd)) {
bzone = ERR_PTR(-EIO);
@@ -2185,11 +2185,12 @@ struct dm_zone *dmz_get_chunk_buffer(struct dmz_metadata *zmd,
* Get an unmapped (free) zone.
* This must be called with the mapping lock held.
*/
-struct dm_zone *dmz_alloc_zone(struct dmz_metadata *zmd, unsigned long flags)
+struct dm_zone *dmz_alloc_zone(struct dmz_metadata *zmd, unsigned int dev_idx,
+ unsigned long flags)
{
struct list_head *list;
struct dm_zone *zone;
- unsigned int dev_idx = 0;
+ int i = 0;
again:
if (flags & DMZ_ALLOC_CACHE)
@@ -2205,8 +2206,12 @@ struct dm_zone *dmz_alloc_zone(struct dmz_metadata *zmd, unsigned long flags)
*/
if (!(flags & DMZ_ALLOC_RECLAIM))
return NULL;
- if (dev_idx < zmd->nr_devs) {
- dev_idx++;
+ /*
+ * Try to allocate from other devices
+ */
+ if (i < zmd->nr_devs) {
+ dev_idx = (dev_idx + 1) % zmd->nr_devs;
+ i++;
goto again;
}
diff --git a/drivers/md/dm-zoned-reclaim.c b/drivers/md/dm-zoned-reclaim.c
index 18edf1b9bf52..5a04e34d17a9 100644
--- a/drivers/md/dm-zoned-reclaim.c
+++ b/drivers/md/dm-zoned-reclaim.c
@@ -288,7 +288,8 @@ static int dmz_reclaim_rnd_data(struct dmz_reclaim *zrc, struct dm_zone *dzone)
/* Get a free random or sequential zone */
dmz_lock_map(zmd);
again:
- szone = dmz_alloc_zone(zmd, alloc_flags | DMZ_ALLOC_RECLAIM);
+ szone = dmz_alloc_zone(zmd, zrc->dev_idx,
+ alloc_flags | DMZ_ALLOC_RECLAIM);
if (!szone && alloc_flags == DMZ_ALLOC_SEQ && dmz_nr_cache_zones(zmd)) {
alloc_flags = DMZ_ALLOC_RND;
goto again;
diff --git a/drivers/md/dm-zoned.h b/drivers/md/dm-zoned.h
index f2a760f62db5..ec020bb1caf7 100644
--- a/drivers/md/dm-zoned.h
+++ b/drivers/md/dm-zoned.h
@@ -214,7 +214,8 @@ bool dmz_dev_is_dying(struct dmz_metadata *zmd);
#define DMZ_ALLOC_SEQ 0x04
#define DMZ_ALLOC_RECLAIM 0x10
-struct dm_zone *dmz_alloc_zone(struct dmz_metadata *zmd, unsigned long flags);
+struct dm_zone *dmz_alloc_zone(struct dmz_metadata *zmd,
+ unsigned int dev_idx, unsigned long flags);
void dmz_free_zone(struct dmz_metadata *zmd, struct dm_zone *zone);
void dmz_map_zone(struct dmz_metadata *zmd, struct dm_zone *zone,
--
2.16.4
next prev parent reply other threads:[~2020-05-29 17:39 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-29 17:38 [PATCHv3 00/14] dm-zoned: multiple drive support Hannes Reinecke
2020-05-29 17:38 ` [PATCH 01/14] dm-zoned: add debugging message for reading superblocks Hannes Reinecke
2020-05-29 17:38 ` [PATCH 02/14] dm-zoned: secondary superblock must reside on the same devices than primary superblock Hannes Reinecke
2020-05-29 17:38 ` [PATCH 03/14] dm-zoned: improve logging messages for reclaim Hannes Reinecke
2020-05-29 17:38 ` [PATCH 04/14] dm-zoned: add a 'reserved' zone flag Hannes Reinecke
2020-05-29 17:38 ` [PATCH 05/14] dm-zoned: convert to xarray Hannes Reinecke
2020-05-29 17:38 ` [PATCH 06/14] dm-zoned: temporary superblock for tertiary devices Hannes Reinecke
2020-05-31 8:58 ` Damien Le Moal
2020-05-29 17:39 ` [PATCH 07/14] dm-zoned: add device pointer to struct dm_zone Hannes Reinecke
2020-05-29 17:39 ` [PATCH 08/14] dm-zoned: add metadata pointer to struct dmz_dev Hannes Reinecke
2020-05-29 17:39 ` [PATCH 09/14] dm-zoned: per-device reclaim Hannes Reinecke
2020-05-31 9:19 ` Damien Le Moal
2020-05-29 17:39 ` [PATCH 10/14] dm-zoned: move random and sequential zones into struct dmz_dev Hannes Reinecke
2020-05-31 9:06 ` Damien Le Moal
2020-05-29 17:39 ` [PATCH 11/14] dm-zoned: support arbitrary number of devices Hannes Reinecke
2020-05-31 9:10 ` Damien Le Moal
2020-05-31 13:06 ` Hannes Reinecke
2020-05-31 23:54 ` Damien Le Moal
2020-06-02 6:42 ` Hannes Reinecke
2020-05-29 17:39 ` Hannes Reinecke [this message]
2020-05-31 9:12 ` [PATCH 12/14] dm-zoned: allocate zone by device index Damien Le Moal
2020-05-29 17:39 ` [PATCH 13/14] dm-zoned: select reclaim zone based on " Hannes Reinecke
2020-05-29 17:39 ` [PATCH 14/14] dm-zoned: prefer full zones for reclaim Hannes Reinecke
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200529173907.40529-13-hare@suse.de \
--to=hare@suse.de \
--cc=damien.lemoal@wdc.com \
--cc=dm-devel@redhat.com \
--cc=snitzer@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox