diff options
| author | Alexander Motin <[email protected]> | 2024-11-10 17:29:25 -0500 |
|---|---|---|
| committer | Brian Behlendorf <[email protected]> | 2024-11-15 15:04:11 -0800 |
| commit | fd6e8c1d2a5c8d95d509b691b81c66597e354981 (patch) | |
| tree | c0e097909e2d2bc381ee2c6ab0ab824c9248ce14 /module/zfs/spa_misc.c | |
| parent | 309ce6303fa7e455e9adee0674a29685bdbbcc8d (diff) | |
BRT: Rework structures and locks to be per-vdev
While block cloning operation from the beginning was made per-vdev,
before this change most of its data were protected by two pool-
wide locks. It created lots of lock contention in many workload.
This change makes most of block cloning data structures per-vdev,
which allows to lock them separately. The only pool-wide lock now
it spa_brt_lock, protecting array of per-vdev pointers and in most
cases taken as reader. Also this splits per-vdev locks into three
different ones: bv_pending_lock protects the AVL-tree of pending
operations in open context, bv_mos_entries_lock protects BRT ZAP
object from while being prefetched, and bv_lock protects the rest
of per-vdev context during TXG commit process. There should be
no functional difference aside of some optimizations.
Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Pawel Jakub Dawidek <[email protected]>
Reviewed-by: Brian Atkinson <[email protected]>
Signed-off-by: Alexander Motin <[email protected]>
Sponsored by: iXsystems, Inc.
Closes #16740
Diffstat (limited to 'module/zfs/spa_misc.c')
| -rw-r--r-- | module/zfs/spa_misc.c | 17 |
1 files changed, 6 insertions, 11 deletions
diff --git a/module/zfs/spa_misc.c b/module/zfs/spa_misc.c index f486513fc..32542e7ce 100644 --- a/module/zfs/spa_misc.c +++ b/module/zfs/spa_misc.c @@ -1870,13 +1870,7 @@ spa_get_slop_space(spa_t *spa) if (spa->spa_dedup_dspace == ~0ULL) spa_update_dspace(spa); - /* - * spa_get_dspace() includes the space only logically "used" by - * deduplicated data, so since it's not useful to reserve more - * space with more deduplicated data, we subtract that out here. - */ - space = - spa_get_dspace(spa) - spa->spa_dedup_dspace - brt_get_dspace(spa); + space = spa->spa_rdspace; slop = MIN(space >> spa_slop_shift, spa_max_slop); /* @@ -1912,8 +1906,7 @@ spa_get_checkpoint_space(spa_t *spa) void spa_update_dspace(spa_t *spa) { - spa->spa_dspace = metaslab_class_get_dspace(spa_normal_class(spa)) + - ddt_get_dedup_dspace(spa) + brt_get_dspace(spa); + spa->spa_rdspace = metaslab_class_get_dspace(spa_normal_class(spa)); if (spa->spa_nonallocating_dspace > 0) { /* * Subtract the space provided by all non-allocating vdevs that @@ -1933,9 +1926,11 @@ spa_update_dspace(spa_t *spa) * doesn't matter that the data we are moving may be * allocated twice (on the old device and the new device). */ - ASSERT3U(spa->spa_dspace, >=, spa->spa_nonallocating_dspace); - spa->spa_dspace -= spa->spa_nonallocating_dspace; + ASSERT3U(spa->spa_rdspace, >=, spa->spa_nonallocating_dspace); + spa->spa_rdspace -= spa->spa_nonallocating_dspace; } + spa->spa_dspace = spa->spa_rdspace + ddt_get_dedup_dspace(spa) + + brt_get_dspace(spa); } /* |