diff options
author | avg <avg@FreeBSD.org> | 2014-01-16 15:57:39 +0000 |
---|---|---|
committer | avg <avg@FreeBSD.org> | 2014-01-16 15:57:39 +0000 |
commit | 26096ba436a54f225f8e56bb965a0e2214256c71 (patch) | |
tree | a24d61eb5c3c255647d4c4bb29075b63b1da94fc /sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c | |
parent | 81894fb568ed816e9e9f3a43f778e69751857339 (diff) | |
download | FreeBSD-src-26096ba436a54f225f8e56bb965a0e2214256c71.zip FreeBSD-src-26096ba436a54f225f8e56bb965a0e2214256c71.tar.gz |
MFC r258632,258704: MFV r255255: 4045 zfs write throttle & i/o scheduler
performance work
Sponsored by: HybridCluster [merge]
Diffstat (limited to 'sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c')
-rw-r--r-- | sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c | 757 |
1 files changed, 547 insertions, 210 deletions
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c index 0a687b2..11f05a0 100644 --- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c +++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_queue.c @@ -24,35 +24,137 @@ */ /* - * Copyright (c) 2012 by Delphix. All rights reserved. + * Copyright (c) 2013 by Delphix. All rights reserved. */ #include <sys/zfs_context.h> #include <sys/vdev_impl.h> +#include <sys/spa_impl.h> #include <sys/zio.h> #include <sys/avl.h> +#include <sys/dsl_pool.h> /* - * These tunables are for performance analysis. + * ZFS I/O Scheduler + * --------------- + * + * ZFS issues I/O operations to leaf vdevs to satisfy and complete zios. The + * I/O scheduler determines when and in what order those operations are + * issued. The I/O scheduler divides operations into five I/O classes + * prioritized in the following order: sync read, sync write, async read, + * async write, and scrub/resilver. Each queue defines the minimum and + * maximum number of concurrent operations that may be issued to the device. + * In addition, the device has an aggregate maximum. Note that the sum of the + * per-queue minimums must not exceed the aggregate maximum, and if the + * aggregate maximum is equal to or greater than the sum of the per-queue + * maximums, the per-queue minimum has no effect. + * + * For many physical devices, throughput increases with the number of + * concurrent operations, but latency typically suffers. Further, physical + * devices typically have a limit at which more concurrent operations have no + * effect on throughput or can actually cause it to decrease. + * + * The scheduler selects the next operation to issue by first looking for an + * I/O class whose minimum has not been satisfied. Once all are satisfied and + * the aggregate maximum has not been hit, the scheduler looks for classes + * whose maximum has not been satisfied. Iteration through the I/O classes is + * done in the order specified above. No further operations are issued if the + * aggregate maximum number of concurrent operations has been hit or if there + * are no operations queued for an I/O class that has not hit its maximum. + * Every time an i/o is queued or an operation completes, the I/O scheduler + * looks for new operations to issue. + * + * All I/O classes have a fixed maximum number of outstanding operations + * except for the async write class. Asynchronous writes represent the data + * that is committed to stable storage during the syncing stage for + * transaction groups (see txg.c). Transaction groups enter the syncing state + * periodically so the number of queued async writes will quickly burst up and + * then bleed down to zero. Rather than servicing them as quickly as possible, + * the I/O scheduler changes the maximum number of active async write i/os + * according to the amount of dirty data in the pool (see dsl_pool.c). Since + * both throughput and latency typically increase with the number of + * concurrent operations issued to physical devices, reducing the burstiness + * in the number of concurrent operations also stabilizes the response time of + * operations from other -- and in particular synchronous -- queues. In broad + * strokes, the I/O scheduler will issue more concurrent operations from the + * async write queue as there's more dirty data in the pool. + * + * Async Writes + * + * The number of concurrent operations issued for the async write I/O class + * follows a piece-wise linear function defined by a few adjustable points. + * + * | o---------| <-- zfs_vdev_async_write_max_active + * ^ | /^ | + * | | / | | + * active | / | | + * I/O | / | | + * count | / | | + * | / | | + * |------------o | | <-- zfs_vdev_async_write_min_active + * 0|____________^______|_________| + * 0% | | 100% of zfs_dirty_data_max + * | | + * | `-- zfs_vdev_async_write_active_max_dirty_percent + * `--------- zfs_vdev_async_write_active_min_dirty_percent + * + * Until the amount of dirty data exceeds a minimum percentage of the dirty + * data allowed in the pool, the I/O scheduler will limit the number of + * concurrent operations to the minimum. As that threshold is crossed, the + * number of concurrent operations issued increases linearly to the maximum at + * the specified maximum percentage of the dirty data allowed in the pool. + * + * Ideally, the amount of dirty data on a busy pool will stay in the sloped + * part of the function between zfs_vdev_async_write_active_min_dirty_percent + * and zfs_vdev_async_write_active_max_dirty_percent. If it exceeds the + * maximum percentage, this indicates that the rate of incoming data is + * greater than the rate that the backend storage can handle. In this case, we + * must further throttle incoming writes (see dmu_tx_delay() for details). */ -/* The maximum number of I/Os concurrently pending to each device. */ -int zfs_vdev_max_pending = 10; - /* - * The initial number of I/Os pending to each device, before it starts ramping - * up to zfs_vdev_max_pending. + * The maximum number of i/os active to each device. Ideally, this will be >= + * the sum of each queue's max_active. It must be at least the sum of each + * queue's min_active. */ -int zfs_vdev_min_pending = 4; +uint32_t zfs_vdev_max_active = 1000; /* - * The deadlines are grouped into buckets based on zfs_vdev_time_shift: - * deadline = pri + gethrtime() >> time_shift) + * Per-queue limits on the number of i/os active to each device. If the + * sum of the queue's max_active is < zfs_vdev_max_active, then the + * min_active comes into play. We will send min_active from each queue, + * and then select from queues in the order defined by zio_priority_t. + * + * In general, smaller max_active's will lead to lower latency of synchronous + * operations. Larger max_active's may lead to higher overall throughput, + * depending on underlying storage. + * + * The ratio of the queues' max_actives determines the balance of performance + * between reads, writes, and scrubs. E.g., increasing + * zfs_vdev_scrub_max_active will cause the scrub or resilver to complete + * more quickly, but reads and writes to have higher latency and lower + * throughput. */ -int zfs_vdev_time_shift = 29; /* each bucket is 0.537 seconds */ +uint32_t zfs_vdev_sync_read_min_active = 10; +uint32_t zfs_vdev_sync_read_max_active = 10; +uint32_t zfs_vdev_sync_write_min_active = 10; +uint32_t zfs_vdev_sync_write_max_active = 10; +uint32_t zfs_vdev_async_read_min_active = 1; +uint32_t zfs_vdev_async_read_max_active = 3; +uint32_t zfs_vdev_async_write_min_active = 1; +uint32_t zfs_vdev_async_write_max_active = 10; +uint32_t zfs_vdev_scrub_min_active = 1; +uint32_t zfs_vdev_scrub_max_active = 2; -/* exponential I/O issue ramp-up rate */ -int zfs_vdev_ramp_rate = 2; +/* + * When the pool has less than zfs_vdev_async_write_active_min_dirty_percent + * dirty data, use zfs_vdev_async_write_min_active. When it has more than + * zfs_vdev_async_write_active_max_dirty_percent, use + * zfs_vdev_async_write_max_active. The value is linearly interpolated + * between min and max. + */ +int zfs_vdev_async_write_active_min_dirty_percent = 30; +int zfs_vdev_async_write_active_max_dirty_percent = 60; /* * To reduce IOPs, we aggregate small adjacent I/Os into one large I/O. @@ -64,20 +166,42 @@ int zfs_vdev_aggregation_limit = SPA_MAXBLOCKSIZE; int zfs_vdev_read_gap_limit = 32 << 10; int zfs_vdev_write_gap_limit = 4 << 10; +#ifdef __FreeBSD__ SYSCTL_DECL(_vfs_zfs_vdev); -TUNABLE_INT("vfs.zfs.vdev.max_pending", &zfs_vdev_max_pending); -SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, max_pending, CTLFLAG_RW, - &zfs_vdev_max_pending, 0, "Maximum I/O requests pending on each device"); -TUNABLE_INT("vfs.zfs.vdev.min_pending", &zfs_vdev_min_pending); -SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, min_pending, CTLFLAG_RW, - &zfs_vdev_min_pending, 0, - "Initial number of I/O requests pending to each device"); -TUNABLE_INT("vfs.zfs.vdev.time_shift", &zfs_vdev_time_shift); -SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, time_shift, CTLFLAG_RW, - &zfs_vdev_time_shift, 0, "Used for calculating I/O request deadline"); -TUNABLE_INT("vfs.zfs.vdev.ramp_rate", &zfs_vdev_ramp_rate); -SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, ramp_rate, CTLFLAG_RW, - &zfs_vdev_ramp_rate, 0, "Exponential I/O issue ramp-up rate"); +TUNABLE_INT("vfs.zfs.vdev.max_active", &zfs_vdev_max_active); +SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, max_active, CTLFLAG_RW, + &zfs_vdev_max_active, 0, + "The maximum number of i/os of all types active for each device."); + +#define ZFS_VDEV_QUEUE_KNOB_MIN(name) \ +TUNABLE_INT("vfs.zfs.vdev." #name "_min_active", \ + &zfs_vdev_ ## name ## _min_active); \ +SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, name ## _min_active, CTLFLAG_RW, \ + &zfs_vdev_ ## name ## _min_active, 0, \ + "Initial number of I/O requests of type " #name \ + " active for each device"); + +#define ZFS_VDEV_QUEUE_KNOB_MAX(name) \ +TUNABLE_INT("vfs.zfs.vdev." #name "_max_active", \ + &zfs_vdev_ ## name ## _max_active); \ +SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, name ## _max_active, CTLFLAG_RW, \ + &zfs_vdev_ ## name ## _max_active, 0, \ + "Maximum number of I/O requests of type " #name \ + " active for each device"); + +ZFS_VDEV_QUEUE_KNOB_MIN(sync_read); +ZFS_VDEV_QUEUE_KNOB_MAX(sync_read); +ZFS_VDEV_QUEUE_KNOB_MIN(sync_write); +ZFS_VDEV_QUEUE_KNOB_MAX(sync_write); +ZFS_VDEV_QUEUE_KNOB_MIN(async_read); +ZFS_VDEV_QUEUE_KNOB_MAX(async_read); +ZFS_VDEV_QUEUE_KNOB_MIN(async_write); +ZFS_VDEV_QUEUE_KNOB_MAX(async_write); +ZFS_VDEV_QUEUE_KNOB_MIN(scrub); +ZFS_VDEV_QUEUE_KNOB_MAX(scrub); + +#undef ZFS_VDEV_QUEUE_KNOB + TUNABLE_INT("vfs.zfs.vdev.aggregation_limit", &zfs_vdev_aggregation_limit); SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, aggregation_limit, CTLFLAG_RW, &zfs_vdev_aggregation_limit, 0, @@ -90,21 +214,14 @@ TUNABLE_INT("vfs.zfs.vdev.write_gap_limit", &zfs_vdev_write_gap_limit); SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, write_gap_limit, CTLFLAG_RW, &zfs_vdev_write_gap_limit, 0, "Acceptable gap between two writes being aggregated"); +#endif -/* - * Virtual device vector for disk I/O scheduling. - */ int -vdev_queue_deadline_compare(const void *x1, const void *x2) +vdev_queue_offset_compare(const void *x1, const void *x2) { const zio_t *z1 = x1; const zio_t *z2 = x2; - if (z1->io_deadline < z2->io_deadline) - return (-1); - if (z1->io_deadline > z2->io_deadline) - return (1); - if (z1->io_offset < z2->io_offset) return (-1); if (z1->io_offset > z2->io_offset) @@ -119,14 +236,14 @@ vdev_queue_deadline_compare(const void *x1, const void *x2) } int -vdev_queue_offset_compare(const void *x1, const void *x2) +vdev_queue_timestamp_compare(const void *x1, const void *x2) { const zio_t *z1 = x1; const zio_t *z2 = x2; - if (z1->io_offset < z2->io_offset) + if (z1->io_timestamp < z2->io_timestamp) return (-1); - if (z1->io_offset > z2->io_offset) + if (z1->io_timestamp > z2->io_timestamp) return (1); if (z1 < z2) @@ -143,18 +260,25 @@ vdev_queue_init(vdev_t *vd) vdev_queue_t *vq = &vd->vdev_queue; mutex_init(&vq->vq_lock, NULL, MUTEX_DEFAULT, NULL); + vq->vq_vdev = vd; - avl_create(&vq->vq_deadline_tree, vdev_queue_deadline_compare, - sizeof (zio_t), offsetof(struct zio, io_deadline_node)); - - avl_create(&vq->vq_read_tree, vdev_queue_offset_compare, - sizeof (zio_t), offsetof(struct zio, io_offset_node)); + avl_create(&vq->vq_active_tree, vdev_queue_offset_compare, + sizeof (zio_t), offsetof(struct zio, io_queue_node)); - avl_create(&vq->vq_write_tree, vdev_queue_offset_compare, - sizeof (zio_t), offsetof(struct zio, io_offset_node)); - - avl_create(&vq->vq_pending_tree, vdev_queue_offset_compare, - sizeof (zio_t), offsetof(struct zio, io_offset_node)); + for (zio_priority_t p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++) { + /* + * The synchronous i/o queues are FIFO rather than LBA ordered. + * This provides more consistent latency for these i/os, and + * they tend to not be tightly clustered anyway so there is + * little to no throughput loss. + */ + boolean_t fifo = (p == ZIO_PRIORITY_SYNC_READ || + p == ZIO_PRIORITY_SYNC_WRITE); + avl_create(&vq->vq_class[p].vqc_queued_tree, + fifo ? vdev_queue_timestamp_compare : + vdev_queue_offset_compare, + sizeof (zio_t), offsetof(struct zio, io_queue_node)); + } } void @@ -162,10 +286,9 @@ vdev_queue_fini(vdev_t *vd) { vdev_queue_t *vq = &vd->vdev_queue; - avl_destroy(&vq->vq_deadline_tree); - avl_destroy(&vq->vq_read_tree); - avl_destroy(&vq->vq_write_tree); - avl_destroy(&vq->vq_pending_tree); + for (zio_priority_t p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++) + avl_destroy(&vq->vq_class[p].vqc_queued_tree); + avl_destroy(&vq->vq_active_tree); mutex_destroy(&vq->vq_lock); } @@ -173,30 +296,204 @@ vdev_queue_fini(vdev_t *vd) static void vdev_queue_io_add(vdev_queue_t *vq, zio_t *zio) { - avl_add(&vq->vq_deadline_tree, zio); - avl_add(zio->io_vdev_tree, zio); + spa_t *spa = zio->io_spa; + ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE); + avl_add(&vq->vq_class[zio->io_priority].vqc_queued_tree, zio); + +#ifdef illumos + mutex_enter(&spa->spa_iokstat_lock); + spa->spa_queue_stats[zio->io_priority].spa_queued++; + if (spa->spa_iokstat != NULL) + kstat_waitq_enter(spa->spa_iokstat->ks_data); + mutex_exit(&spa->spa_iokstat_lock); +#endif } static void vdev_queue_io_remove(vdev_queue_t *vq, zio_t *zio) { - avl_remove(&vq->vq_deadline_tree, zio); - avl_remove(zio->io_vdev_tree, zio); + spa_t *spa = zio->io_spa; + ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE); + avl_remove(&vq->vq_class[zio->io_priority].vqc_queued_tree, zio); + +#ifdef illumos + mutex_enter(&spa->spa_iokstat_lock); + ASSERT3U(spa->spa_queue_stats[zio->io_priority].spa_queued, >, 0); + spa->spa_queue_stats[zio->io_priority].spa_queued--; + if (spa->spa_iokstat != NULL) + kstat_waitq_exit(spa->spa_iokstat->ks_data); + mutex_exit(&spa->spa_iokstat_lock); +#endif } static void -vdev_queue_agg_io_done(zio_t *aio) +vdev_queue_pending_add(vdev_queue_t *vq, zio_t *zio) +{ + spa_t *spa = zio->io_spa; + ASSERT(MUTEX_HELD(&vq->vq_lock)); + ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE); + vq->vq_class[zio->io_priority].vqc_active++; + avl_add(&vq->vq_active_tree, zio); + +#ifdef illumos + mutex_enter(&spa->spa_iokstat_lock); + spa->spa_queue_stats[zio->io_priority].spa_active++; + if (spa->spa_iokstat != NULL) + kstat_runq_enter(spa->spa_iokstat->ks_data); + mutex_exit(&spa->spa_iokstat_lock); +#endif +} + +static void +vdev_queue_pending_remove(vdev_queue_t *vq, zio_t *zio) { - zio_t *pio; + spa_t *spa = zio->io_spa; + ASSERT(MUTEX_HELD(&vq->vq_lock)); + ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE); + vq->vq_class[zio->io_priority].vqc_active--; + avl_remove(&vq->vq_active_tree, zio); + +#ifdef illumos + mutex_enter(&spa->spa_iokstat_lock); + ASSERT3U(spa->spa_queue_stats[zio->io_priority].spa_active, >, 0); + spa->spa_queue_stats[zio->io_priority].spa_active--; + if (spa->spa_iokstat != NULL) { + kstat_io_t *ksio = spa->spa_iokstat->ks_data; + + kstat_runq_exit(spa->spa_iokstat->ks_data); + if (zio->io_type == ZIO_TYPE_READ) { + ksio->reads++; + ksio->nread += zio->io_size; + } else if (zio->io_type == ZIO_TYPE_WRITE) { + ksio->writes++; + ksio->nwritten += zio->io_size; + } + } + mutex_exit(&spa->spa_iokstat_lock); +#endif +} - while ((pio = zio_walk_parents(aio)) != NULL) - if (aio->io_type == ZIO_TYPE_READ) +static void +vdev_queue_agg_io_done(zio_t *aio) +{ + if (aio->io_type == ZIO_TYPE_READ) { + zio_t *pio; + while ((pio = zio_walk_parents(aio)) != NULL) { bcopy((char *)aio->io_data + (pio->io_offset - aio->io_offset), pio->io_data, pio->io_size); + } + } zio_buf_free(aio->io_data, aio->io_size); } +static int +vdev_queue_class_min_active(zio_priority_t p) +{ + switch (p) { + case ZIO_PRIORITY_SYNC_READ: + return (zfs_vdev_sync_read_min_active); + case ZIO_PRIORITY_SYNC_WRITE: + return (zfs_vdev_sync_write_min_active); + case ZIO_PRIORITY_ASYNC_READ: + return (zfs_vdev_async_read_min_active); + case ZIO_PRIORITY_ASYNC_WRITE: + return (zfs_vdev_async_write_min_active); + case ZIO_PRIORITY_SCRUB: + return (zfs_vdev_scrub_min_active); + default: + panic("invalid priority %u", p); + return (0); + } +} + +static int +vdev_queue_max_async_writes(uint64_t dirty) +{ + int writes; + uint64_t min_bytes = zfs_dirty_data_max * + zfs_vdev_async_write_active_min_dirty_percent / 100; + uint64_t max_bytes = zfs_dirty_data_max * + zfs_vdev_async_write_active_max_dirty_percent / 100; + + if (dirty < min_bytes) + return (zfs_vdev_async_write_min_active); + if (dirty > max_bytes) + return (zfs_vdev_async_write_max_active); + + /* + * linear interpolation: + * slope = (max_writes - min_writes) / (max_bytes - min_bytes) + * move right by min_bytes + * move up by min_writes + */ + writes = (dirty - min_bytes) * + (zfs_vdev_async_write_max_active - + zfs_vdev_async_write_min_active) / + (max_bytes - min_bytes) + + zfs_vdev_async_write_min_active; + ASSERT3U(writes, >=, zfs_vdev_async_write_min_active); + ASSERT3U(writes, <=, zfs_vdev_async_write_max_active); + return (writes); +} + +static int +vdev_queue_class_max_active(spa_t *spa, zio_priority_t p) +{ + switch (p) { + case ZIO_PRIORITY_SYNC_READ: + return (zfs_vdev_sync_read_max_active); + case ZIO_PRIORITY_SYNC_WRITE: + return (zfs_vdev_sync_write_max_active); + case ZIO_PRIORITY_ASYNC_READ: + return (zfs_vdev_async_read_max_active); + case ZIO_PRIORITY_ASYNC_WRITE: + return (vdev_queue_max_async_writes( + spa->spa_dsl_pool->dp_dirty_total)); + case ZIO_PRIORITY_SCRUB: + return (zfs_vdev_scrub_max_active); + default: + panic("invalid priority %u", p); + return (0); + } +} + +/* + * Return the i/o class to issue from, or ZIO_PRIORITY_MAX_QUEUEABLE if + * there is no eligible class. + */ +static zio_priority_t +vdev_queue_class_to_issue(vdev_queue_t *vq) +{ + spa_t *spa = vq->vq_vdev->vdev_spa; + zio_priority_t p; + + if (avl_numnodes(&vq->vq_active_tree) >= zfs_vdev_max_active) + return (ZIO_PRIORITY_NUM_QUEUEABLE); + + /* find a queue that has not reached its minimum # outstanding i/os */ + for (p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++) { + if (avl_numnodes(&vq->vq_class[p].vqc_queued_tree) > 0 && + vq->vq_class[p].vqc_active < + vdev_queue_class_min_active(p)) + return (p); + } + + /* + * If we haven't found a queue, look for one that hasn't reached its + * maximum # outstanding i/os. + */ + for (p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++) { + if (avl_numnodes(&vq->vq_class[p].vqc_queued_tree) > 0 && + vq->vq_class[p].vqc_active < + vdev_queue_class_max_active(spa, p)) + return (p); + } + + /* No eligible queued i/os */ + return (ZIO_PRIORITY_NUM_QUEUEABLE); +} + /* * Compute the range spanned by two i/os, which is the endpoint of the last * (lio->io_offset + lio->io_size) minus start of the first (fio->io_offset). @@ -207,154 +504,192 @@ vdev_queue_agg_io_done(zio_t *aio) #define IO_GAP(fio, lio) (-IO_SPAN(lio, fio)) static zio_t * -vdev_queue_io_to_issue(vdev_queue_t *vq, uint64_t pending_limit) +vdev_queue_aggregate(vdev_queue_t *vq, zio_t *zio) { - zio_t *fio, *lio, *aio, *dio, *nio, *mio; - avl_tree_t *t; - int flags; - uint64_t maxspan = zfs_vdev_aggregation_limit; - uint64_t maxgap; - int stretch; - -again: - ASSERT(MUTEX_HELD(&vq->vq_lock)); + zio_t *first, *last, *aio, *dio, *mandatory, *nio; + uint64_t maxgap = 0; + uint64_t size; + boolean_t stretch = B_FALSE; + vdev_queue_class_t *vqc = &vq->vq_class[zio->io_priority]; + avl_tree_t *t = &vqc->vqc_queued_tree; + enum zio_flag flags = zio->io_flags & ZIO_FLAG_AGG_INHERIT; + + if (zio->io_flags & ZIO_FLAG_DONT_AGGREGATE) + return (NULL); - if (avl_numnodes(&vq->vq_pending_tree) >= pending_limit || - avl_numnodes(&vq->vq_deadline_tree) == 0) + /* + * The synchronous i/o queues are not sorted by LBA, so we can't + * find adjacent i/os. These i/os tend to not be tightly clustered, + * or too large to aggregate, so this has little impact on performance. + */ + if (zio->io_priority == ZIO_PRIORITY_SYNC_READ || + zio->io_priority == ZIO_PRIORITY_SYNC_WRITE) return (NULL); - fio = lio = avl_first(&vq->vq_deadline_tree); + first = last = zio; - t = fio->io_vdev_tree; - flags = fio->io_flags & ZIO_FLAG_AGG_INHERIT; - maxgap = (t == &vq->vq_read_tree) ? zfs_vdev_read_gap_limit : 0; + if (zio->io_type == ZIO_TYPE_READ) + maxgap = zfs_vdev_read_gap_limit; - if (!(flags & ZIO_FLAG_DONT_AGGREGATE)) { - /* - * We can aggregate I/Os that are sufficiently adjacent and of - * the same flavor, as expressed by the AGG_INHERIT flags. - * The latter requirement is necessary so that certain - * attributes of the I/O, such as whether it's a normal I/O - * or a scrub/resilver, can be preserved in the aggregate. - * We can include optional I/Os, but don't allow them - * to begin a range as they add no benefit in that situation. - */ + /* + * We can aggregate I/Os that are sufficiently adjacent and of + * the same flavor, as expressed by the AGG_INHERIT flags. + * The latter requirement is necessary so that certain + * attributes of the I/O, such as whether it's a normal I/O + * or a scrub/resilver, can be preserved in the aggregate. + * We can include optional I/Os, but don't allow them + * to begin a range as they add no benefit in that situation. + */ - /* - * We keep track of the last non-optional I/O. - */ - mio = (fio->io_flags & ZIO_FLAG_OPTIONAL) ? NULL : fio; + /* + * We keep track of the last non-optional I/O. + */ + mandatory = (first->io_flags & ZIO_FLAG_OPTIONAL) ? NULL : first; - /* - * Walk backwards through sufficiently contiguous I/Os - * recording the last non-option I/O. - */ - while ((dio = AVL_PREV(t, fio)) != NULL && - (dio->io_flags & ZIO_FLAG_AGG_INHERIT) == flags && - IO_SPAN(dio, lio) <= maxspan && - IO_GAP(dio, fio) <= maxgap) { - fio = dio; - if (mio == NULL && !(fio->io_flags & ZIO_FLAG_OPTIONAL)) - mio = fio; - } + /* + * Walk backwards through sufficiently contiguous I/Os + * recording the last non-option I/O. + */ + while ((dio = AVL_PREV(t, first)) != NULL && + (dio->io_flags & ZIO_FLAG_AGG_INHERIT) == flags && + IO_SPAN(dio, last) <= zfs_vdev_aggregation_limit && + IO_GAP(dio, first) <= maxgap) { + first = dio; + if (mandatory == NULL && !(first->io_flags & ZIO_FLAG_OPTIONAL)) + mandatory = first; + } - /* - * Skip any initial optional I/Os. - */ - while ((fio->io_flags & ZIO_FLAG_OPTIONAL) && fio != lio) { - fio = AVL_NEXT(t, fio); - ASSERT(fio != NULL); - } + /* + * Skip any initial optional I/Os. + */ + while ((first->io_flags & ZIO_FLAG_OPTIONAL) && first != last) { + first = AVL_NEXT(t, first); + ASSERT(first != NULL); + } - /* - * Walk forward through sufficiently contiguous I/Os. - */ - while ((dio = AVL_NEXT(t, lio)) != NULL && - (dio->io_flags & ZIO_FLAG_AGG_INHERIT) == flags && - IO_SPAN(fio, dio) <= maxspan && - IO_GAP(lio, dio) <= maxgap) { - lio = dio; - if (!(lio->io_flags & ZIO_FLAG_OPTIONAL)) - mio = lio; - } + /* + * Walk forward through sufficiently contiguous I/Os. + */ + while ((dio = AVL_NEXT(t, last)) != NULL && + (dio->io_flags & ZIO_FLAG_AGG_INHERIT) == flags && + IO_SPAN(first, dio) <= zfs_vdev_aggregation_limit && + IO_GAP(last, dio) <= maxgap) { + last = dio; + if (!(last->io_flags & ZIO_FLAG_OPTIONAL)) + mandatory = last; + } - /* - * Now that we've established the range of the I/O aggregation - * we must decide what to do with trailing optional I/Os. - * For reads, there's nothing to do. While we are unable to - * aggregate further, it's possible that a trailing optional - * I/O would allow the underlying device to aggregate with - * subsequent I/Os. We must therefore determine if the next - * non-optional I/O is close enough to make aggregation - * worthwhile. - */ - stretch = B_FALSE; - if (t != &vq->vq_read_tree && mio != NULL) { - nio = lio; - while ((dio = AVL_NEXT(t, nio)) != NULL && - IO_GAP(nio, dio) == 0 && - IO_GAP(mio, dio) <= zfs_vdev_write_gap_limit) { - nio = dio; - if (!(nio->io_flags & ZIO_FLAG_OPTIONAL)) { - stretch = B_TRUE; - break; - } + /* + * Now that we've established the range of the I/O aggregation + * we must decide what to do with trailing optional I/Os. + * For reads, there's nothing to do. While we are unable to + * aggregate further, it's possible that a trailing optional + * I/O would allow the underlying device to aggregate with + * subsequent I/Os. We must therefore determine if the next + * non-optional I/O is close enough to make aggregation + * worthwhile. + */ + if (zio->io_type == ZIO_TYPE_WRITE && mandatory != NULL) { + zio_t *nio = last; + while ((dio = AVL_NEXT(t, nio)) != NULL && + IO_GAP(nio, dio) == 0 && + IO_GAP(mandatory, dio) <= zfs_vdev_write_gap_limit) { + nio = dio; + if (!(nio->io_flags & ZIO_FLAG_OPTIONAL)) { + stretch = B_TRUE; + break; } } + } - if (stretch) { - /* This may be a no-op. */ - VERIFY((dio = AVL_NEXT(t, lio)) != NULL); - dio->io_flags &= ~ZIO_FLAG_OPTIONAL; - } else { - while (lio != mio && lio != fio) { - ASSERT(lio->io_flags & ZIO_FLAG_OPTIONAL); - lio = AVL_PREV(t, lio); - ASSERT(lio != NULL); - } + if (stretch) { + /* This may be a no-op. */ + dio = AVL_NEXT(t, last); + dio->io_flags &= ~ZIO_FLAG_OPTIONAL; + } else { + while (last != mandatory && last != first) { + ASSERT(last->io_flags & ZIO_FLAG_OPTIONAL); + last = AVL_PREV(t, last); + ASSERT(last != NULL); } } - if (fio != lio) { - uint64_t size = IO_SPAN(fio, lio); - ASSERT(size <= zfs_vdev_aggregation_limit); - - aio = zio_vdev_delegated_io(fio->io_vd, fio->io_offset, - zio_buf_alloc(size), size, fio->io_type, ZIO_PRIORITY_AGG, - flags | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE, - vdev_queue_agg_io_done, NULL); - aio->io_timestamp = fio->io_timestamp; - - nio = fio; - do { - dio = nio; - nio = AVL_NEXT(t, dio); - ASSERT(dio->io_type == aio->io_type); - ASSERT(dio->io_vdev_tree == t); - - if (dio->io_flags & ZIO_FLAG_NODATA) { - ASSERT(dio->io_type == ZIO_TYPE_WRITE); - bzero((char *)aio->io_data + (dio->io_offset - - aio->io_offset), dio->io_size); - } else if (dio->io_type == ZIO_TYPE_WRITE) { - bcopy(dio->io_data, (char *)aio->io_data + - (dio->io_offset - aio->io_offset), - dio->io_size); - } + if (first == last) + return (NULL); + + size = IO_SPAN(first, last); + ASSERT3U(size, <=, zfs_vdev_aggregation_limit); + + aio = zio_vdev_delegated_io(first->io_vd, first->io_offset, + zio_buf_alloc(size), size, first->io_type, zio->io_priority, + flags | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE, + vdev_queue_agg_io_done, NULL); + aio->io_timestamp = first->io_timestamp; + + nio = first; + do { + dio = nio; + nio = AVL_NEXT(t, dio); + ASSERT3U(dio->io_type, ==, aio->io_type); + + if (dio->io_flags & ZIO_FLAG_NODATA) { + ASSERT3U(dio->io_type, ==, ZIO_TYPE_WRITE); + bzero((char *)aio->io_data + (dio->io_offset - + aio->io_offset), dio->io_size); + } else if (dio->io_type == ZIO_TYPE_WRITE) { + bcopy(dio->io_data, (char *)aio->io_data + + (dio->io_offset - aio->io_offset), + dio->io_size); + } - zio_add_child(dio, aio); - vdev_queue_io_remove(vq, dio); - zio_vdev_io_bypass(dio); - zio_execute(dio); - } while (dio != lio); + zio_add_child(dio, aio); + vdev_queue_io_remove(vq, dio); + zio_vdev_io_bypass(dio); + zio_execute(dio); + } while (dio != last); + + return (aio); +} + +static zio_t * +vdev_queue_io_to_issue(vdev_queue_t *vq) +{ + zio_t *zio, *aio; + zio_priority_t p; + avl_index_t idx; + vdev_queue_class_t *vqc; + zio_t search; + +again: + ASSERT(MUTEX_HELD(&vq->vq_lock)); - avl_add(&vq->vq_pending_tree, aio); + p = vdev_queue_class_to_issue(vq); - return (aio); + if (p == ZIO_PRIORITY_NUM_QUEUEABLE) { + /* No eligible queued i/os */ + return (NULL); } - ASSERT(fio->io_vdev_tree == t); - vdev_queue_io_remove(vq, fio); + /* + * For LBA-ordered queues (async / scrub), issue the i/o which follows + * the most recently issued i/o in LBA (offset) order. + * + * For FIFO queues (sync), issue the i/o with the lowest timestamp. + */ + vqc = &vq->vq_class[p]; + search.io_timestamp = 0; + search.io_offset = vq->vq_last_offset + 1; + VERIFY3P(avl_find(&vqc->vqc_queued_tree, &search, &idx), ==, NULL); + zio = avl_nearest(&vqc->vqc_queued_tree, idx, AVL_AFTER); + if (zio == NULL) + zio = avl_first(&vqc->vqc_queued_tree); + ASSERT3U(zio->io_priority, ==, p); + + aio = vdev_queue_aggregate(vq, zio); + if (aio != NULL) + zio = aio; + else + vdev_queue_io_remove(vq, zio); /* * If the I/O is or was optional and therefore has no data, we need to @@ -362,17 +697,18 @@ again: * deadlock that we could encounter since this I/O will complete * immediately. */ - if (fio->io_flags & ZIO_FLAG_NODATA) { + if (zio->io_flags & ZIO_FLAG_NODATA) { mutex_exit(&vq->vq_lock); - zio_vdev_io_bypass(fio); - zio_execute(fio); + zio_vdev_io_bypass(zio); + zio_execute(zio); mutex_enter(&vq->vq_lock); goto again; } - avl_add(&vq->vq_pending_tree, fio); + vdev_queue_pending_add(vq, zio); + vq->vq_last_offset = zio->io_offset; - return (fio); + return (zio); } zio_t * @@ -381,28 +717,31 @@ vdev_queue_io(zio_t *zio) vdev_queue_t *vq = &zio->io_vd->vdev_queue; zio_t *nio; - ASSERT(zio->io_type == ZIO_TYPE_READ || zio->io_type == ZIO_TYPE_WRITE); - if (zio->io_flags & ZIO_FLAG_DONT_QUEUE) return (zio); - zio->io_flags |= ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE; + /* + * Children i/os inherent their parent's priority, which might + * not match the child's i/o type. Fix it up here. + */ + if (zio->io_type == ZIO_TYPE_READ) { + if (zio->io_priority != ZIO_PRIORITY_SYNC_READ && + zio->io_priority != ZIO_PRIORITY_ASYNC_READ && + zio->io_priority != ZIO_PRIORITY_SCRUB) + zio->io_priority = ZIO_PRIORITY_ASYNC_READ; + } else { + ASSERT(zio->io_type == ZIO_TYPE_WRITE); + if (zio->io_priority != ZIO_PRIORITY_SYNC_WRITE && + zio->io_priority != ZIO_PRIORITY_ASYNC_WRITE) + zio->io_priority = ZIO_PRIORITY_ASYNC_WRITE; + } - if (zio->io_type == ZIO_TYPE_READ) - zio->io_vdev_tree = &vq->vq_read_tree; - else - zio->io_vdev_tree = &vq->vq_write_tree; + zio->io_flags |= ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE; mutex_enter(&vq->vq_lock); - zio->io_timestamp = gethrtime(); - zio->io_deadline = (zio->io_timestamp >> zfs_vdev_time_shift) + - zio->io_priority; - vdev_queue_io_add(vq, zio); - - nio = vdev_queue_io_to_issue(vq, zfs_vdev_min_pending); - + nio = vdev_queue_io_to_issue(vq); mutex_exit(&vq->vq_lock); if (nio == NULL) @@ -420,20 +759,18 @@ void vdev_queue_io_done(zio_t *zio) { vdev_queue_t *vq = &zio->io_vd->vdev_queue; + zio_t *nio; if (zio_injection_enabled) delay(SEC_TO_TICK(zio_handle_io_delay(zio))); mutex_enter(&vq->vq_lock); - avl_remove(&vq->vq_pending_tree, zio); + vdev_queue_pending_remove(vq, zio); vq->vq_io_complete_ts = gethrtime(); - for (int i = 0; i < zfs_vdev_ramp_rate; i++) { - zio_t *nio = vdev_queue_io_to_issue(vq, zfs_vdev_max_pending); - if (nio == NULL) - break; + while ((nio = vdev_queue_io_to_issue(vq)) != NULL) { mutex_exit(&vq->vq_lock); if (nio->io_done == vdev_queue_agg_io_done) { zio_nowait(nio); |