path: root/fs/xfs/xfs_zone_gc.c

// 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. Additionally, the m_zonegc_low_space tunable
 * can be set to make sure a fraction of the unused blocks are available for
 * writing.
 */
bool
xfs_zoned_need_gc(
	struct xfs_mount	*mp)
{
	s64			available, free, threshold;
	s32			remainder;

	if (!xfs_group_marked(mp, XG_TYPE_RTG, XFS_RTG_RECLAIMABLE))
		return false;

	available = xfs_estimate_freecounter(mp, XC_FREE_RTAVAILABLE);

	if (available <
	    mp->m_groups[XG_TYPE_RTG].blocks *
	    (mp->m_max_open_zones - XFS_OPEN_GC_ZONES))
		return true;

	free = xfs_estimate_freecounter(mp, XC_FREE_RTEXTENTS);

	threshold = div_s64_rem(free, 100, &remainder);
	threshold = threshold * mp->m_zonegc_low_space +
		    remainder * div_s64(mp->m_zonegc_low_space, 100);

	if (available < threshold)
		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