Merge tag 'xfs-zoned-allocator-2025-03-03' of git://git.infradead.org/users/hch/xfs into xfs-6.15-zoned_devices

xfs: add support for zoned devices

Add support for the new zoned space allocator and thus for zoned devices:

    https://zonedstorage.io/docs/introduction/zoned-storage

to XFS. This has been developed for and tested on both SMR hard drives,
which are the oldest and most common class of zoned devices:

   https://zonedstorage.io/docs/introduction/smr

and ZNS SSDs:

   https://zonedstorage.io/docs/introduction/zns

It has not been tested with zoned UFS devices, as their current capacity
points and performance characteristics aren't too interesting for XFS
use cases (but never say never).

Sequential write only zones are only supported for data using a new
allocator for the RT device, which maps each zone to a rtgroup which
is written sequentially.  All metadata and (for now) the log require
using randomly writable space. This means a realtime device is required
to support zoned storage, but for the common case of SMR hard drives
that contain random writable zones and sequential write required zones
on the same block device, the concept of an internal RT device is added
which means using XFS on a SMR HDD is as simple as:

$ mkfs.xfs /dev/sda
$ mount /dev/sda /mnt

When using NVMe ZNS SSDs that do not support conventional zones, the
traditional multi-device RT configuration is required.  E.g. for an
SSD with a conventional namespace 1 and a zoned namespace 2:

$ mkfs.xfs /dev/nvme0n1 -o rtdev=/dev/nvme0n2
$ mount -o rtdev=/dev/nvme0n2 /dev/nvme0n1 /mnt

The zoned allocator can also be used on conventional block devices, or
on conventional zones (e.g. when using an SMR HDD as the external RT
device).  For example using zoned XFS on normal SSDs shows very nice
performance advantages and write amplification reduction for intelligent
workloads like RocksDB.

Some work is still in progress or planned, but should not affect the
integration with the rest of XFS or the on-disk format:

 - support for quotas
 - support for reflinks

Note that the I/O path already supports reflink, but garbage collection
isn't refcount aware yet and would unshare shared blocks, thus rendering
the feature useless.

1 files changed, 1165 insertions, 0 deletions

diff --git a/fs/xfs/xfs_zone_gc.c b/fs/xfs/xfs_zone_gc.c
new file mode 100644
index 000000000000..c5136ea9bb1d
--- /dev/null
+++ b/fs/xfs/xfs_zone_gc.c
@@ -0,0 +1,1165 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2023-2025 Christoph Hellwig.
+ * Copyright (c) 2024-2025, Western Digital Corporation or its affiliates.
+ */
+#include "xfs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_trans.h"
+#include "xfs_icache.h"
+#include "xfs_rmap.h"
+#include "xfs_rtbitmap.h"
+#include "xfs_rtrmap_btree.h"
+#include "xfs_zone_alloc.h"
+#include "xfs_zone_priv.h"
+#include "xfs_zones.h"
+#include "xfs_trace.h"
+
+/*
+ * Implement Garbage Collection (GC) of partially used zoned.
+ *
+ * To support the purely sequential writes in each zone, zoned XFS needs to be
+ * able to move data remaining in a zone out of it to reset the zone to prepare
+ * for writing to it again.
+ *
+ * This is done by the GC thread implemented in this file.  To support that a
+ * number of zones (XFS_GC_ZONES) is reserved from the user visible capacity to
+ * write the garbage collected data into.
+ *
+ * Whenever the available space is below the chosen threshold, the GC thread
+ * looks for potential non-empty but not fully used zones that are worth
+ * reclaiming.  Once found the rmap for the victim zone is queried, and after
+ * a bit of sorting to reduce fragmentation, the still live extents are read
+ * into memory and written to the GC target zone, and the bmap btree of the
+ * files is updated to point to the new location.  To avoid taking the IOLOCK
+ * and MMAPLOCK for the entire GC process and thus affecting the latency of
+ * user reads and writes to the files, the GC writes are speculative and the
+ * I/O completion checks that no other writes happened for the affected regions
+ * before remapping.
+ *
+ * Once a zone does not contain any valid data, be that through GC or user
+ * block removal, it is queued for for a zone reset.  The reset operation
+ * carefully ensures that the RT device cache is flushed and all transactions
+ * referencing the rmap have been committed to disk.
+ */
+
+/*
+ * Size of each GC scratch pad.  This is also the upper bound for each
+ * GC I/O, which helps to keep latency down.
+ */
+#define XFS_GC_CHUNK_SIZE	SZ_1M
+
+/*
+ * Scratchpad data to read GCed data into.
+ *
+ * The offset member tracks where the next allocation starts, and freed tracks
+ * the amount of space that is not used anymore.
+ */
+#define XFS_ZONE_GC_NR_SCRATCH	2
+struct xfs_zone_scratch {
+	struct folio			*folio;
+	unsigned int			offset;
+	unsigned int			freed;
+};
+
+/*
+ * Chunk that is read and written for each GC operation.
+ *
+ * Note that for writes to actual zoned devices, the chunk can be split when
+ * reaching the hardware limit.
+ */
+struct xfs_gc_bio {
+	struct xfs_zone_gc_data		*data;
+
+	/*
+	 * Entry into the reading/writing/resetting list.  Only accessed from
+	 * the GC thread, so no locking needed.
+	 */
+	struct list_head		entry;
+
+	/*
+	 * State of this gc_bio.  Done means the current I/O completed.
+	 * Set from the bio end I/O handler, read from the GC thread.
+	 */
+	enum {
+		XFS_GC_BIO_NEW,
+		XFS_GC_BIO_DONE,
+	} state;
+
+	/*
+	 * Pointer to the inode and byte range in the inode that this
+	 * GC chunk is operating on.
+	 */
+	struct xfs_inode		*ip;
+	loff_t				offset;
+	unsigned int			len;
+
+	/*
+	 * Existing startblock (in the zone to be freed) and newly assigned
+	 * daddr in the zone GCed into.
+	 */
+	xfs_fsblock_t			old_startblock;
+	xfs_daddr_t			new_daddr;
+	struct xfs_zone_scratch		*scratch;
+
+	/* Are we writing to a sequential write required zone? */
+	bool				is_seq;
+
+	/* Open Zone being written to */
+	struct xfs_open_zone		*oz;
+
+	/* Bio used for reads and writes, including the bvec used by it */
+	struct bio_vec			bv;
+	struct bio			bio;	/* must be last */
+};
+
+#define XFS_ZONE_GC_RECS		1024
+
+/* iterator, needs to be reinitialized for each victim zone */
+struct xfs_zone_gc_iter {
+	struct xfs_rtgroup		*victim_rtg;
+	unsigned int			rec_count;
+	unsigned int			rec_idx;
+	xfs_agblock_t			next_startblock;
+	struct xfs_rmap_irec		*recs;
+};
+
+/*
+ * Per-mount GC state.
+ */
+struct xfs_zone_gc_data {
+	struct xfs_mount		*mp;
+
+	/* bioset used to allocate the gc_bios */
+	struct bio_set			bio_set;
+
+	/*
+	 * Scratchpad used, and index to indicated which one is used.
+	 */
+	struct xfs_zone_scratch		scratch[XFS_ZONE_GC_NR_SCRATCH];
+	unsigned int			scratch_idx;
+
+	/*
+	 * List of bios currently being read, written and reset.
+	 * These lists are only accessed by the GC thread itself, and must only
+	 * be processed in order.
+	 */
+	struct list_head		reading;
+	struct list_head		writing;
+	struct list_head		resetting;
+
+	/*
+	 * Iterator for the victim zone.
+	 */
+	struct xfs_zone_gc_iter		iter;
+};
+
+/*
+ * We aim to keep enough zones free in stock to fully use the open zone limit
+ * for data placement purposes.
+ */
+bool
+xfs_zoned_need_gc(
+	struct xfs_mount	*mp)
+{
+	if (!xfs_group_marked(mp, XG_TYPE_RTG, XFS_RTG_RECLAIMABLE))
+		return false;
+	if (xfs_estimate_freecounter(mp, XC_FREE_RTAVAILABLE) <
+	    mp->m_groups[XG_TYPE_RTG].blocks *
+	    (mp->m_max_open_zones - XFS_OPEN_GC_ZONES))
+		return true;
+	return false;
+}
+
+static struct xfs_zone_gc_data *
+xfs_zone_gc_data_alloc(
+	struct xfs_mount	*mp)
+{
+	struct xfs_zone_gc_data	*data;
+	int			i;
+
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return NULL;
+	data->iter.recs = kcalloc(XFS_ZONE_GC_RECS, sizeof(*data->iter.recs),
+			GFP_KERNEL);
+	if (!data->iter.recs)
+		goto out_free_data;
+
+	/*
+	 * We actually only need a single bio_vec.  It would be nice to have
+	 * a flag that only allocates the inline bvecs and not the separate
+	 * bvec pool.
+	 */
+	if (bioset_init(&data->bio_set, 16, offsetof(struct xfs_gc_bio, bio),
+			BIOSET_NEED_BVECS))
+		goto out_free_recs;
+	for (i = 0; i < XFS_ZONE_GC_NR_SCRATCH; i++) {
+		data->scratch[i].folio =
+			folio_alloc(GFP_KERNEL, get_order(XFS_GC_CHUNK_SIZE));
+		if (!data->scratch[i].folio)
+			goto out_free_scratch;
+	}
+	INIT_LIST_HEAD(&data->reading);
+	INIT_LIST_HEAD(&data->writing);
+	INIT_LIST_HEAD(&data->resetting);
+	data->mp = mp;
+	return data;
+
+out_free_scratch:
+	while (--i >= 0)
+		folio_put(data->scratch[i].folio);
+	bioset_exit(&data->bio_set);
+out_free_recs:
+	kfree(data->iter.recs);
+out_free_data:
+	kfree(data);
+	return NULL;
+}
+
+static void
+xfs_zone_gc_data_free(
+	struct xfs_zone_gc_data	*data)
+{
+	int			i;
+
+	for (i = 0; i < XFS_ZONE_GC_NR_SCRATCH; i++)
+		folio_put(data->scratch[i].folio);
+	bioset_exit(&data->bio_set);
+	kfree(data->iter.recs);
+	kfree(data);
+}
+
+static void
+xfs_zone_gc_iter_init(
+	struct xfs_zone_gc_iter	*iter,
+	struct xfs_rtgroup	*victim_rtg)
+
+{
+	iter->next_startblock = 0;
+	iter->rec_count = 0;
+	iter->rec_idx = 0;
+	iter->victim_rtg = victim_rtg;
+}
+
+/*
+ * Query the rmap of the victim zone to gather the records to evacuate.
+ */
+static int
+xfs_zone_gc_query_cb(
+	struct xfs_btree_cur	*cur,
+	const struct xfs_rmap_irec *irec,
+	void			*private)
+{
+	struct xfs_zone_gc_iter	*iter = private;
+
+	ASSERT(!XFS_RMAP_NON_INODE_OWNER(irec->rm_owner));
+	ASSERT(!xfs_is_sb_inum(cur->bc_mp, irec->rm_owner));
+	ASSERT(!(irec->rm_flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK)));
+
+	iter->recs[iter->rec_count] = *irec;
+	if (++iter->rec_count == XFS_ZONE_GC_RECS) {
+		iter->next_startblock =
+			irec->rm_startblock + irec->rm_blockcount;
+		return 1;
+	}
+	return 0;
+}
+
+#define cmp_int(l, r)		((l > r) - (l < r))
+
+static int
+xfs_zone_gc_rmap_rec_cmp(
+	const void			*a,
+	const void			*b)
+{
+	const struct xfs_rmap_irec	*reca = a;
+	const struct xfs_rmap_irec	*recb = b;
+	int				diff;
+
+	diff = cmp_int(reca->rm_owner, recb->rm_owner);
+	if (diff)
+		return diff;
+	return cmp_int(reca->rm_offset, recb->rm_offset);
+}
+
+static int
+xfs_zone_gc_query(
+	struct xfs_mount	*mp,
+	struct xfs_zone_gc_iter	*iter)
+{
+	struct xfs_rtgroup	*rtg = iter->victim_rtg;
+	struct xfs_rmap_irec	ri_low = { };
+	struct xfs_rmap_irec	ri_high;
+	struct xfs_btree_cur	*cur;
+	struct xfs_trans	*tp;
+	int			error;
+
+	ASSERT(iter->next_startblock <= rtg_blocks(rtg));
+	if (iter->next_startblock == rtg_blocks(rtg))
+		goto done;
+
+	ASSERT(iter->next_startblock < rtg_blocks(rtg));
+	ri_low.rm_startblock = iter->next_startblock;
+	memset(&ri_high, 0xFF, sizeof(ri_high));
+
+	iter->rec_idx = 0;
+	iter->rec_count = 0;
+
+	error = xfs_trans_alloc_empty(mp, &tp);
+	if (error)
+		return error;
+
+	xfs_rtgroup_lock(rtg, XFS_RTGLOCK_RMAP);
+	cur = xfs_rtrmapbt_init_cursor(tp, rtg);
+	error = xfs_rmap_query_range(cur, &ri_low, &ri_high,
+			xfs_zone_gc_query_cb, iter);
+	xfs_rtgroup_unlock(rtg, XFS_RTGLOCK_RMAP);
+	xfs_btree_del_cursor(cur, error < 0 ? error : 0);
+	xfs_trans_cancel(tp);
+
+	if (error < 0)
+		return error;
+
+	/*
+	 * Sort the rmap records by inode number and increasing offset to
+	 * defragment the mappings.
+	 *
+	 * This could be further enhanced by an even bigger look ahead window,
+	 * but that's better left until we have better detection of changes to
+	 * inode mapping to avoid the potential of GCing already dead data.
+	 */
+	sort(iter->recs, iter->rec_count, sizeof(iter->recs[0]),
+			xfs_zone_gc_rmap_rec_cmp, NULL);
+
+	if (error == 0) {
+		/*
+		 * We finished iterating through the zone.
+		 */
+		iter->next_startblock = rtg_blocks(rtg);
+		if (iter->rec_count == 0)
+			goto done;
+	}
+
+	return 0;
+done:
+	xfs_rtgroup_rele(iter->victim_rtg);
+	iter->victim_rtg = NULL;
+	return 0;
+}
+
+static bool
+xfs_zone_gc_iter_next(
+	struct xfs_mount	*mp,
+	struct xfs_zone_gc_iter	*iter,
+	struct xfs_rmap_irec	*chunk_rec,
+	struct xfs_inode	**ipp)
+{
+	struct xfs_rmap_irec	*irec;
+	int			error;
+
+	if (!iter->victim_rtg)
+		return false;
+
+retry:
+	if (iter->rec_idx == iter->rec_count) {
+		error = xfs_zone_gc_query(mp, iter);
+		if (error)
+			goto fail;
+		if (!iter->victim_rtg)
+			return false;
+	}
+
+	irec = &iter->recs[iter->rec_idx];
+	error = xfs_iget(mp, NULL, irec->rm_owner,
+			XFS_IGET_UNTRUSTED | XFS_IGET_DONTCACHE, 0, ipp);
+	if (error) {
+		/*
+		 * If the inode was already deleted, skip over it.
+		 */
+		if (error == -ENOENT) {
+			iter->rec_idx++;
+			goto retry;
+		}
+		goto fail;
+	}
+
+	if (!S_ISREG(VFS_I(*ipp)->i_mode) || !XFS_IS_REALTIME_INODE(*ipp)) {
+		iter->rec_idx++;
+		xfs_irele(*ipp);
+		goto retry;
+	}
+
+	*chunk_rec = *irec;
+	return true;
+
+fail:
+	xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+	return false;
+}
+
+static void
+xfs_zone_gc_iter_advance(
+	struct xfs_zone_gc_iter	*iter,
+	xfs_extlen_t		count_fsb)
+{
+	struct xfs_rmap_irec	*irec = &iter->recs[iter->rec_idx];
+
+	irec->rm_offset += count_fsb;
+	irec->rm_startblock += count_fsb;
+	irec->rm_blockcount -= count_fsb;
+	if (!irec->rm_blockcount)
+		iter->rec_idx++;
+}
+
+static struct xfs_rtgroup *
+xfs_zone_gc_pick_victim_from(
+	struct xfs_mount	*mp,
+	uint32_t		bucket)
+{
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+	uint32_t		victim_used = U32_MAX;
+	struct xfs_rtgroup	*victim_rtg = NULL;
+	uint32_t		bit;
+
+	if (!zi->zi_used_bucket_entries[bucket])
+		return NULL;
+
+	for_each_set_bit(bit, zi->zi_used_bucket_bitmap[bucket],
+			mp->m_sb.sb_rgcount) {
+		struct xfs_rtgroup *rtg = xfs_rtgroup_grab(mp, bit);
+
+		if (!rtg)
+			continue;
+
+		/* skip zones that are just waiting for a reset */
+		if (rtg_rmap(rtg)->i_used_blocks == 0 ||
+		    rtg_rmap(rtg)->i_used_blocks >= victim_used) {
+			xfs_rtgroup_rele(rtg);
+			continue;
+		}
+
+		if (victim_rtg)
+			xfs_rtgroup_rele(victim_rtg);
+		victim_rtg = rtg;
+		victim_used = rtg_rmap(rtg)->i_used_blocks;
+
+		/*
+		 * Any zone that is less than 1 percent used is fair game for
+		 * instant reclaim. All of these zones are in the last
+		 * bucket, so avoid the expensive division for the zones
+		 * in the other buckets.
+		 */
+		if (bucket == 0 &&
+		    rtg_rmap(rtg)->i_used_blocks < rtg_blocks(rtg) / 100)
+			break;
+	}
+
+	return victim_rtg;
+}
+
+/*
+ * Iterate through all zones marked as reclaimable and find a candidate to
+ * reclaim.
+ */
+static bool
+xfs_zone_gc_select_victim(
+	struct xfs_zone_gc_data	*data)
+{
+	struct xfs_zone_gc_iter	*iter = &data->iter;
+	struct xfs_mount	*mp = data->mp;
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+	struct xfs_rtgroup	*victim_rtg = NULL;
+	unsigned int		bucket;
+
+	if (xfs_is_shutdown(mp))
+		return false;
+
+	if (iter->victim_rtg)
+		return true;
+
+	/*
+	 * Don't start new work if we are asked to stop or park.
+	 */
+	if (kthread_should_stop() || kthread_should_park())
+		return false;
+
+	if (!xfs_zoned_need_gc(mp))
+		return false;
+
+	spin_lock(&zi->zi_used_buckets_lock);
+	for (bucket = 0; bucket < XFS_ZONE_USED_BUCKETS; bucket++) {
+		victim_rtg = xfs_zone_gc_pick_victim_from(mp, bucket);
+		if (victim_rtg)
+			break;
+	}
+	spin_unlock(&zi->zi_used_buckets_lock);
+
+	if (!victim_rtg)
+		return false;
+
+	trace_xfs_zone_gc_select_victim(victim_rtg, bucket);
+	xfs_zone_gc_iter_init(iter, victim_rtg);
+	return true;
+}
+
+static struct xfs_open_zone *
+xfs_zone_gc_steal_open(
+	struct xfs_zone_info	*zi)
+{
+	struct xfs_open_zone	*oz, *found = NULL;
+
+	spin_lock(&zi->zi_open_zones_lock);
+	list_for_each_entry(oz, &zi->zi_open_zones, oz_entry) {
+		if (!found ||
+		    oz->oz_write_pointer < found->oz_write_pointer)
+			found = oz;
+	}
+
+	if (found) {
+		found->oz_is_gc = true;
+		list_del_init(&found->oz_entry);
+		zi->zi_nr_open_zones--;
+	}
+
+	spin_unlock(&zi->zi_open_zones_lock);
+	return found;
+}
+
+static struct xfs_open_zone *
+xfs_zone_gc_select_target(
+	struct xfs_mount	*mp)
+{
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+	struct xfs_open_zone	*oz = zi->zi_open_gc_zone;
+
+	/*
+	 * We need to wait for pending writes to finish.
+	 */
+	if (oz && oz->oz_written < rtg_blocks(oz->oz_rtg))
+		return NULL;
+
+	ASSERT(zi->zi_nr_open_zones <=
+		mp->m_max_open_zones - XFS_OPEN_GC_ZONES);
+	oz = xfs_open_zone(mp, WRITE_LIFE_NOT_SET, true);
+	if (oz)
+		trace_xfs_zone_gc_target_opened(oz->oz_rtg);
+	spin_lock(&zi->zi_open_zones_lock);
+	zi->zi_open_gc_zone = oz;
+	spin_unlock(&zi->zi_open_zones_lock);
+	return oz;
+}
+
+/*
+ * Ensure we have a valid open zone to write the GC data to.
+ *
+ * If the current target zone has space keep writing to it, else first wait for
+ * all pending writes and then pick a new one.
+ */
+static struct xfs_open_zone *
+xfs_zone_gc_ensure_target(
+	struct xfs_mount	*mp)
+{
+	struct xfs_open_zone	*oz = mp->m_zone_info->zi_open_gc_zone;
+
+	if (!oz || oz->oz_write_pointer == rtg_blocks(oz->oz_rtg))
+		return xfs_zone_gc_select_target(mp);
+	return oz;
+}
+
+static unsigned int
+xfs_zone_gc_scratch_available(
+	struct xfs_zone_gc_data	*data)
+{
+	return XFS_GC_CHUNK_SIZE - data->scratch[data->scratch_idx].offset;
+}
+
+static bool
+xfs_zone_gc_space_available(
+	struct xfs_zone_gc_data	*data)
+{
+	struct xfs_open_zone	*oz;
+
+	oz = xfs_zone_gc_ensure_target(data->mp);
+	if (!oz)
+		return false;
+	return oz->oz_write_pointer < rtg_blocks(oz->oz_rtg) &&
+		xfs_zone_gc_scratch_available(data);
+}
+
+static void
+xfs_zone_gc_end_io(
+	struct bio		*bio)
+{
+	struct xfs_gc_bio	*chunk =
+		container_of(bio, struct xfs_gc_bio, bio);
+	struct xfs_zone_gc_data	*data = chunk->data;
+
+	WRITE_ONCE(chunk->state, XFS_GC_BIO_DONE);
+	wake_up_process(data->mp->m_zone_info->zi_gc_thread);
+}
+
+static struct xfs_open_zone *
+xfs_zone_gc_alloc_blocks(
+	struct xfs_zone_gc_data	*data,
+	xfs_extlen_t		*count_fsb,
+	xfs_daddr_t		*daddr,
+	bool			*is_seq)
+{
+	struct xfs_mount	*mp = data->mp;
+	struct xfs_open_zone	*oz;
+
+	oz = xfs_zone_gc_ensure_target(mp);
+	if (!oz)
+		return NULL;
+
+	*count_fsb = min(*count_fsb,
+		XFS_B_TO_FSB(mp, xfs_zone_gc_scratch_available(data)));
+
+	/*
+	 * Directly allocate GC blocks from the reserved pool.
+	 *
+	 * If we'd take them from the normal pool we could be stealing blocks
+	 * from a regular writer, which would then have to wait for GC and
+	 * deadlock.
+	 */
+	spin_lock(&mp->m_sb_lock);
+	*count_fsb = min(*count_fsb,
+			rtg_blocks(oz->oz_rtg) - oz->oz_write_pointer);
+	*count_fsb = min3(*count_fsb,
+			mp->m_free[XC_FREE_RTEXTENTS].res_avail,
+			mp->m_free[XC_FREE_RTAVAILABLE].res_avail);
+	mp->m_free[XC_FREE_RTEXTENTS].res_avail -= *count_fsb;
+	mp->m_free[XC_FREE_RTAVAILABLE].res_avail -= *count_fsb;
+	spin_unlock(&mp->m_sb_lock);
+
+	if (!*count_fsb)
+		return NULL;
+
+	*daddr = xfs_gbno_to_daddr(&oz->oz_rtg->rtg_group, 0);
+	*is_seq = bdev_zone_is_seq(mp->m_rtdev_targp->bt_bdev, *daddr);
+	if (!*is_seq)
+		*daddr += XFS_FSB_TO_BB(mp, oz->oz_write_pointer);
+	oz->oz_write_pointer += *count_fsb;
+	atomic_inc(&oz->oz_ref);
+	return oz;
+}
+
+static bool
+xfs_zone_gc_start_chunk(
+	struct xfs_zone_gc_data	*data)
+{
+	struct xfs_zone_gc_iter	*iter = &data->iter;
+	struct xfs_mount	*mp = data->mp;
+	struct block_device	*bdev = mp->m_rtdev_targp->bt_bdev;
+	struct xfs_open_zone	*oz;
+	struct xfs_rmap_irec	irec;
+	struct xfs_gc_bio	*chunk;
+	struct xfs_inode	*ip;
+	struct bio		*bio;
+	xfs_daddr_t		daddr;
+	bool			is_seq;
+
+	if (xfs_is_shutdown(mp))
+		return false;
+
+	if (!xfs_zone_gc_iter_next(mp, iter, &irec, &ip))
+		return false;
+	oz = xfs_zone_gc_alloc_blocks(data, &irec.rm_blockcount, &daddr,
+			&is_seq);
+	if (!oz) {
+		xfs_irele(ip);
+		return false;
+	}
+
+	bio = bio_alloc_bioset(bdev, 1, REQ_OP_READ, GFP_NOFS, &data->bio_set);
+
+	chunk = container_of(bio, struct xfs_gc_bio, bio);
+	chunk->ip = ip;
+	chunk->offset = XFS_FSB_TO_B(mp, irec.rm_offset);
+	chunk->len = XFS_FSB_TO_B(mp, irec.rm_blockcount);
+	chunk->old_startblock =
+		xfs_rgbno_to_rtb(iter->victim_rtg, irec.rm_startblock);
+	chunk->new_daddr = daddr;
+	chunk->is_seq = is_seq;
+	chunk->scratch = &data->scratch[data->scratch_idx];
+	chunk->data = data;
+	chunk->oz = oz;
+
+	bio->bi_iter.bi_sector = xfs_rtb_to_daddr(mp, chunk->old_startblock);
+	bio->bi_end_io = xfs_zone_gc_end_io;
+	bio_add_folio_nofail(bio, chunk->scratch->folio, chunk->len,
+			chunk->scratch->offset);
+	chunk->scratch->offset += chunk->len;
+	if (chunk->scratch->offset == XFS_GC_CHUNK_SIZE) {
+		data->scratch_idx =
+			(data->scratch_idx + 1) % XFS_ZONE_GC_NR_SCRATCH;
+	}
+	WRITE_ONCE(chunk->state, XFS_GC_BIO_NEW);
+	list_add_tail(&chunk->entry, &data->reading);
+	xfs_zone_gc_iter_advance(iter, irec.rm_blockcount);
+
+	submit_bio(bio);
+	return true;
+}
+
+static void
+xfs_zone_gc_free_chunk(
+	struct xfs_gc_bio	*chunk)
+{
+	list_del(&chunk->entry);
+	xfs_open_zone_put(chunk->oz);
+	xfs_irele(chunk->ip);
+	bio_put(&chunk->bio);
+}
+
+static void
+xfs_zone_gc_submit_write(
+	struct xfs_zone_gc_data	*data,
+	struct xfs_gc_bio	*chunk)
+{
+	if (chunk->is_seq) {
+		chunk->bio.bi_opf &= ~REQ_OP_WRITE;
+		chunk->bio.bi_opf |= REQ_OP_ZONE_APPEND;
+	}
+	chunk->bio.bi_iter.bi_sector = chunk->new_daddr;
+	chunk->bio.bi_end_io = xfs_zone_gc_end_io;
+	submit_bio(&chunk->bio);
+}
+
+static struct xfs_gc_bio *
+xfs_zone_gc_split_write(
+	struct xfs_zone_gc_data	*data,
+	struct xfs_gc_bio	*chunk)
+{
+	struct queue_limits	*lim =
+		&bdev_get_queue(chunk->bio.bi_bdev)->limits;
+	struct xfs_gc_bio	*split_chunk;
+	int			split_sectors;
+	unsigned int		split_len;
+	struct bio		*split;
+	unsigned int		nsegs;
+
+	if (!chunk->is_seq)
+		return NULL;
+
+	split_sectors = bio_split_rw_at(&chunk->bio, lim, &nsegs,
+			lim->max_zone_append_sectors << SECTOR_SHIFT);
+	if (!split_sectors)
+		return NULL;
+
+	/* ensure the split chunk is still block size aligned */
+	split_sectors = ALIGN_DOWN(split_sectors << SECTOR_SHIFT,
+			data->mp->m_sb.sb_blocksize) >> SECTOR_SHIFT;
+	split_len = split_sectors << SECTOR_SHIFT;
+
+	split = bio_split(&chunk->bio, split_sectors, GFP_NOFS, &data->bio_set);
+	split_chunk = container_of(split, struct xfs_gc_bio, bio);
+	split_chunk->data = data;
+	ihold(VFS_I(chunk->ip));
+	split_chunk->ip = chunk->ip;
+	split_chunk->is_seq = chunk->is_seq;
+	split_chunk->scratch = chunk->scratch;
+	split_chunk->offset = chunk->offset;
+	split_chunk->len = split_len;
+	split_chunk->old_startblock = chunk->old_startblock;
+	split_chunk->new_daddr = chunk->new_daddr;
+	split_chunk->oz = chunk->oz;
+	atomic_inc(&chunk->oz->oz_ref);
+
+	chunk->offset += split_len;
+	chunk->len -= split_len;
+	chunk->old_startblock += XFS_B_TO_FSB(data->mp, split_len);
+
+	/* add right before the original chunk */
+	WRITE_ONCE(split_chunk->state, XFS_GC_BIO_NEW);
+	list_add_tail(&split_chunk->entry, &chunk->entry);
+	return split_chunk;
+}
+
+static void
+xfs_zone_gc_write_chunk(
+	struct xfs_gc_bio	*chunk)
+{
+	struct xfs_zone_gc_data	*data = chunk->data;
+	struct xfs_mount	*mp = chunk->ip->i_mount;
+	unsigned int		folio_offset = chunk->bio.bi_io_vec->bv_offset;
+	struct xfs_gc_bio	*split_chunk;
+
+	if (chunk->bio.bi_status)
+		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+	if (xfs_is_shutdown(mp)) {
+		xfs_zone_gc_free_chunk(chunk);
+		return;
+	}
+
+	WRITE_ONCE(chunk->state, XFS_GC_BIO_NEW);
+	list_move_tail(&chunk->entry, &data->writing);
+
+	bio_reset(&chunk->bio, mp->m_rtdev_targp->bt_bdev, REQ_OP_WRITE);
+	bio_add_folio_nofail(&chunk->bio, chunk->scratch->folio, chunk->len,
+			folio_offset);
+
+	while ((split_chunk = xfs_zone_gc_split_write(data, chunk)))
+		xfs_zone_gc_submit_write(data, split_chunk);
+	xfs_zone_gc_submit_write(data, chunk);
+}
+
+static void
+xfs_zone_gc_finish_chunk(
+	struct xfs_gc_bio	*chunk)
+{
+	uint			iolock = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
+	struct xfs_inode	*ip = chunk->ip;
+	struct xfs_mount	*mp = ip->i_mount;
+	int			error;
+
+	if (chunk->bio.bi_status)
+		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+	if (xfs_is_shutdown(mp)) {
+		xfs_zone_gc_free_chunk(chunk);
+		return;
+	}
+
+	chunk->scratch->freed += chunk->len;
+	if (chunk->scratch->freed == chunk->scratch->offset) {
+		chunk->scratch->offset = 0;
+		chunk->scratch->freed = 0;
+	}
+
+	/*
+	 * Cycle through the iolock and wait for direct I/O and layouts to
+	 * ensure no one is reading from the old mapping before it goes away.
+	 *
+	 * Note that xfs_zoned_end_io() below checks that no other writer raced
+	 * with us to update the mapping by checking that the old startblock
+	 * didn't change.
+	 */
+	xfs_ilock(ip, iolock);
+	error = xfs_break_layouts(VFS_I(ip), &iolock, BREAK_UNMAP);
+	if (!error)
+		inode_dio_wait(VFS_I(ip));
+	xfs_iunlock(ip, iolock);
+	if (error)
+		goto free;
+
+	if (chunk->is_seq)
+		chunk->new_daddr = chunk->bio.bi_iter.bi_sector;
+	error = xfs_zoned_end_io(ip, chunk->offset, chunk->len,
+			chunk->new_daddr, chunk->oz, chunk->old_startblock);
+free:
+	if (error)
+		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+	xfs_zone_gc_free_chunk(chunk);
+}
+
+static void
+xfs_zone_gc_finish_reset(
+	struct xfs_gc_bio	*chunk)
+{
+	struct xfs_rtgroup	*rtg = chunk->bio.bi_private;
+	struct xfs_mount	*mp = rtg_mount(rtg);
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+
+	if (chunk->bio.bi_status) {
+		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+		goto out;
+	}
+
+	xfs_group_set_mark(&rtg->rtg_group, XFS_RTG_FREE);
+	atomic_inc(&zi->zi_nr_free_zones);
+
+	xfs_zoned_add_available(mp, rtg_blocks(rtg));
+
+	wake_up_all(&zi->zi_zone_wait);
+out:
+	list_del(&chunk->entry);
+	bio_put(&chunk->bio);
+}
+
+static bool
+xfs_zone_gc_prepare_reset(
+	struct bio		*bio,
+	struct xfs_rtgroup	*rtg)
+{
+	trace_xfs_zone_reset(rtg);
+
+	ASSERT(rtg_rmap(rtg)->i_used_blocks == 0);
+	bio->bi_iter.bi_sector = xfs_gbno_to_daddr(&rtg->rtg_group, 0);
+	if (!bdev_zone_is_seq(bio->bi_bdev, bio->bi_iter.bi_sector)) {
+		if (!bdev_max_discard_sectors(bio->bi_bdev))
+			return false;
+		bio->bi_opf = REQ_OP_DISCARD | REQ_SYNC;
+		bio->bi_iter.bi_size =
+			XFS_FSB_TO_B(rtg_mount(rtg), rtg_blocks(rtg));
+	}
+
+	return true;
+}
+
+int
+xfs_zone_gc_reset_sync(
+	struct xfs_rtgroup	*rtg)
+{
+	int			error = 0;
+	struct bio		bio;
+
+	bio_init(&bio, rtg_mount(rtg)->m_rtdev_targp->bt_bdev, NULL, 0,
+			REQ_OP_ZONE_RESET);
+	if (xfs_zone_gc_prepare_reset(&bio, rtg))
+		error = submit_bio_wait(&bio);
+	bio_uninit(&bio);
+
+	return error;
+}
+
+static void
+xfs_zone_gc_reset_zones(
+	struct xfs_zone_gc_data	*data,
+	struct xfs_group	*reset_list)
+{
+	struct xfs_group	*next = reset_list;
+
+	if (blkdev_issue_flush(data->mp->m_rtdev_targp->bt_bdev) < 0) {
+		xfs_force_shutdown(data->mp, SHUTDOWN_META_IO_ERROR);
+		return;
+	}
+
+	do {
+		struct xfs_rtgroup	*rtg = to_rtg(next);
+		struct xfs_gc_bio	*chunk;
+		struct bio		*bio;
+
+		xfs_log_force_inode(rtg_rmap(rtg));
+
+		next = rtg_group(rtg)->xg_next_reset;
+		rtg_group(rtg)->xg_next_reset = NULL;
+
+		bio = bio_alloc_bioset(rtg_mount(rtg)->m_rtdev_targp->bt_bdev,
+				0, REQ_OP_ZONE_RESET, GFP_NOFS, &data->bio_set);
+		bio->bi_private = rtg;
+		bio->bi_end_io = xfs_zone_gc_end_io;
+
+		chunk = container_of(bio, struct xfs_gc_bio, bio);
+		chunk->data = data;
+		WRITE_ONCE(chunk->state, XFS_GC_BIO_NEW);
+		list_add_tail(&chunk->entry, &data->resetting);
+
+		/*
+		 * Also use the bio to drive the state machine when neither
+		 * zone reset nor discard is supported to keep things simple.
+		 */
+		if (xfs_zone_gc_prepare_reset(bio, rtg))
+			submit_bio(bio);
+		else
+			bio_endio(bio);
+	} while (next);
+}
+
+/*
+ * Handle the work to read and write data for GC and to reset the zones,
+ * including handling all completions.
+ *
+ * Note that the order of the chunks is preserved so that we don't undo the
+ * optimal order established by xfs_zone_gc_query().
+ */
+static bool
+xfs_zone_gc_handle_work(
+	struct xfs_zone_gc_data	*data)
+{
+	struct xfs_zone_info	*zi = data->mp->m_zone_info;
+	struct xfs_gc_bio	*chunk, *next;
+	struct xfs_group	*reset_list;
+	struct blk_plug		plug;
+
+	spin_lock(&zi->zi_reset_list_lock);
+	reset_list = zi->zi_reset_list;
+	zi->zi_reset_list = NULL;
+	spin_unlock(&zi->zi_reset_list_lock);
+
+	if (!xfs_zone_gc_select_victim(data) ||
+	    !xfs_zone_gc_space_available(data)) {
+		if (list_empty(&data->reading) &&
+		    list_empty(&data->writing) &&
+		    list_empty(&data->resetting) &&
+		    !reset_list)
+			return false;
+	}
+
+	__set_current_state(TASK_RUNNING);
+	try_to_freeze();
+
+	if (reset_list)
+		xfs_zone_gc_reset_zones(data, reset_list);
+
+	list_for_each_entry_safe(chunk, next, &data->resetting, entry) {
+		if (READ_ONCE(chunk->state) != XFS_GC_BIO_DONE)
+			break;
+		xfs_zone_gc_finish_reset(chunk);
+	}
+
+	list_for_each_entry_safe(chunk, next, &data->writing, entry) {
+		if (READ_ONCE(chunk->state) != XFS_GC_BIO_DONE)
+			break;
+		xfs_zone_gc_finish_chunk(chunk);
+	}
+
+	blk_start_plug(&plug);
+	list_for_each_entry_safe(chunk, next, &data->reading, entry) {
+		if (READ_ONCE(chunk->state) != XFS_GC_BIO_DONE)
+			break;
+		xfs_zone_gc_write_chunk(chunk);
+	}
+	blk_finish_plug(&plug);
+
+	blk_start_plug(&plug);
+	while (xfs_zone_gc_start_chunk(data))
+		;
+	blk_finish_plug(&plug);
+	return true;
+}
+
+/*
+ * Note that the current GC algorithm would break reflinks and thus duplicate
+ * data that was shared by multiple owners before.  Because of that reflinks
+ * are currently not supported on zoned file systems and can't be created or
+ * mounted.
+ */
+static int
+xfs_zoned_gcd(
+	void			*private)
+{
+	struct xfs_zone_gc_data	*data = private;
+	struct xfs_mount	*mp = data->mp;
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+	unsigned int		nofs_flag;
+
+	nofs_flag = memalloc_nofs_save();
+	set_freezable();
+
+	for (;;) {
+		set_current_state(TASK_INTERRUPTIBLE | TASK_FREEZABLE);
+		xfs_set_zonegc_running(mp);
+		if (xfs_zone_gc_handle_work(data))
+			continue;
+
+		if (list_empty(&data->reading) &&
+		    list_empty(&data->writing) &&
+		    list_empty(&data->resetting) &&
+		    !zi->zi_reset_list) {
+			xfs_clear_zonegc_running(mp);
+			xfs_zoned_resv_wake_all(mp);
+
+			if (kthread_should_stop()) {
+				__set_current_state(TASK_RUNNING);
+				break;
+			}
+
+			if (kthread_should_park()) {
+				__set_current_state(TASK_RUNNING);
+				kthread_parkme();
+				continue;
+			}
+		}
+
+		schedule();
+	}
+	xfs_clear_zonegc_running(mp);
+
+	if (data->iter.victim_rtg)
+		xfs_rtgroup_rele(data->iter.victim_rtg);
+
+	memalloc_nofs_restore(nofs_flag);
+	xfs_zone_gc_data_free(data);
+	return 0;
+}
+
+void
+xfs_zone_gc_start(
+	struct xfs_mount	*mp)
+{
+	if (xfs_has_zoned(mp))
+		kthread_unpark(mp->m_zone_info->zi_gc_thread);
+}
+
+void
+xfs_zone_gc_stop(
+	struct xfs_mount	*mp)
+{
+	if (xfs_has_zoned(mp))
+		kthread_park(mp->m_zone_info->zi_gc_thread);
+}
+
+int
+xfs_zone_gc_mount(
+	struct xfs_mount	*mp)
+{
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+	struct xfs_zone_gc_data	*data;
+	struct xfs_open_zone	*oz;
+	int			error;
+
+	/*
+	 * If there are no free zones available for GC, pick the open zone with
+	 * the least used space to GC into.  This should only happen after an
+	 * unclean shutdown near ENOSPC while GC was ongoing.
+	 *
+	 * We also need to do this for the first gc zone allocation if we
+	 * unmounted while at the open limit.
+	 */
+	if (!xfs_group_marked(mp, XG_TYPE_RTG, XFS_RTG_FREE) ||
+	    zi->zi_nr_open_zones == mp->m_max_open_zones)
+		oz = xfs_zone_gc_steal_open(zi);
+	else
+		oz = xfs_open_zone(mp, WRITE_LIFE_NOT_SET, true);
+	if (!oz) {
+		xfs_warn(mp, "unable to allocate a zone for gc");
+		error = -EIO;
+		goto out;
+	}
+
+	trace_xfs_zone_gc_target_opened(oz->oz_rtg);
+	zi->zi_open_gc_zone = oz;
+
+	data = xfs_zone_gc_data_alloc(mp);
+	if (!data) {
+		error = -ENOMEM;
+		goto out_put_gc_zone;
+	}
+
+	mp->m_zone_info->zi_gc_thread = kthread_create(xfs_zoned_gcd, data,
+			"xfs-zone-gc/%s", mp->m_super->s_id);
+	if (IS_ERR(mp->m_zone_info->zi_gc_thread)) {
+		xfs_warn(mp, "unable to create zone gc thread");
+		error = PTR_ERR(mp->m_zone_info->zi_gc_thread);
+		goto out_free_gc_data;
+	}
+
+	/* xfs_zone_gc_start will unpark for rw mounts */
+	kthread_park(mp->m_zone_info->zi_gc_thread);
+	return 0;
+
+out_free_gc_data:
+	kfree(data);
+out_put_gc_zone:
+	xfs_open_zone_put(zi->zi_open_gc_zone);
+out:
+	return error;
+}
+
+void
+xfs_zone_gc_unmount(
+	struct xfs_mount	*mp)
+{
+	struct xfs_zone_info	*zi = mp->m_zone_info;
+
+	kthread_stop(zi->zi_gc_thread);
+	if (zi->zi_open_gc_zone)
+		xfs_open_zone_put(zi->zi_open_gc_zone);
+}