Optimize allocation throttling

Remove mc_lock use from metaslab_class_throttle_*().  The math there
is based on refcounts and so atomic, so the only race possible there
is between zfs_refcount_count() and zfs_refcount_add().  But in most
cases metaslab_class_throttle_reserve() is called with the allocator
lock held, which covers the race.  In cases where the lock is not
held, GANG_ALLOCATION() or METASLAB_MUST_RESERVE are set, and so we
do not use zfs_refcount_count().  And even if we assume some other
non-existing scenario, the worst that may happen from this race is
few more I/Os get to allocation earlier, that is not a problem.

Move locks and data of different allocators into different cache
lines to avoid false sharing.  Group spa_alloc_* arrays together
into single array of aligned struct spa_alloc spa_allocs.  Align
struct metaslab_class_allocator.

Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Don Brady <don.brady@delphix.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes #12314
This commit is contained in:
Alexander Motin
2021-07-21 08:40:36 -04:00
committed by GitHub
parent bc93935ef0
commit 1b50749ce9
6 changed files with 45 additions and 58 deletions
+4 -16
View File
@@ -5611,19 +5611,11 @@ metaslab_class_throttle_reserve(metaslab_class_t *mc, int slots, int allocator,
zio_t *zio, int flags)
{
metaslab_class_allocator_t *mca = &mc->mc_allocator[allocator];
uint64_t available_slots = 0;
boolean_t slot_reserved = B_FALSE;
uint64_t max = mca->mca_alloc_max_slots;
ASSERT(mc->mc_alloc_throttle_enabled);
mutex_enter(&mc->mc_lock);
uint64_t reserved_slots = zfs_refcount_count(&mca->mca_alloc_slots);
if (reserved_slots < max)
available_slots = max - reserved_slots;
if (slots <= available_slots || GANG_ALLOCATION(flags) ||
flags & METASLAB_MUST_RESERVE) {
if (GANG_ALLOCATION(flags) || (flags & METASLAB_MUST_RESERVE) ||
zfs_refcount_count(&mca->mca_alloc_slots) + slots <= max) {
/*
* We reserve the slots individually so that we can unreserve
* them individually when an I/O completes.
@@ -5631,11 +5623,9 @@ metaslab_class_throttle_reserve(metaslab_class_t *mc, int slots, int allocator,
for (int d = 0; d < slots; d++)
zfs_refcount_add(&mca->mca_alloc_slots, zio);
zio->io_flags |= ZIO_FLAG_IO_ALLOCATING;
slot_reserved = B_TRUE;
return (B_TRUE);
}
mutex_exit(&mc->mc_lock);
return (slot_reserved);
return (B_FALSE);
}
void
@@ -5645,10 +5635,8 @@ metaslab_class_throttle_unreserve(metaslab_class_t *mc, int slots,
metaslab_class_allocator_t *mca = &mc->mc_allocator[allocator];
ASSERT(mc->mc_alloc_throttle_enabled);
mutex_enter(&mc->mc_lock);
for (int d = 0; d < slots; d++)
zfs_refcount_remove(&mca->mca_alloc_slots, zio);
mutex_exit(&mc->mc_lock);
}
static int