linux/fs/bcachefs/alloc_background.c

// SPDX-License-Identifier: GPL-2.0
#include "bcachefs.h"
#include "alloc_background.h"
#include "alloc_foreground.h"
#include "btree_cache.h"
#include "btree_io.h"
#include "btree_key_cache.h"
#include "btree_update.h"
#include "btree_update_interior.h"
#include "btree_gc.h"
#include "buckets.h"
#include "buckets_waiting_for_journal.h"
#include "clock.h"
#include "debug.h"
#include "ec.h"
#include "error.h"
#include "recovery.h"
#include "trace.h"
#include "varint.h"

#include <linux/kthread.h>
#include <linux/math64.h>
#include <linux/random.h>
#include <linux/rculist.h>
#include <linux/rcupdate.h>
#include <linux/sched/task.h>
#include <linux/sort.h>

const char * const bch2_allocator_states[] = {
#define x(n)	#n,
	ALLOC_THREAD_STATES()
#undef x
	NULL
};

static const unsigned BCH_ALLOC_V1_FIELD_BYTES[] = {
#define x(name, bits) [BCH_ALLOC_FIELD_V1_##name] = bits / 8,
	BCH_ALLOC_FIELDS_V1()
#undef x
};

/* Persistent alloc info: */

static inline u64 alloc_field_v1_get(const struct bch_alloc *a,
				     const void **p, unsigned field)
{
	unsigned bytes = BCH_ALLOC_V1_FIELD_BYTES[field];
	u64 v;

	if (!(a->fields & (1 << field)))
		return 0;

	switch (bytes) {
	case 1:
		v = *((const u8 *) *p);
		break;
	case 2:
		v = le16_to_cpup(*p);
		break;
	case 4:
		v = le32_to_cpup(*p);
		break;
	case 8:
		v = le64_to_cpup(*p);
		break;
	default:
		BUG();
	}

	*p += bytes;
	return v;
}

static inline void alloc_field_v1_put(struct bkey_i_alloc *a, void **p,
				      unsigned field, u64 v)
{
	unsigned bytes = BCH_ALLOC_V1_FIELD_BYTES[field];

	if (!v)
		return;

	a->v.fields |= 1 << field;

	switch (bytes) {
	case 1:
		*((u8 *) *p) = v;
		break;
	case 2:
		*((__le16 *) *p) = cpu_to_le16(v);
		break;
	case 4:
		*((__le32 *) *p) = cpu_to_le32(v);
		break;
	case 8:
		*((__le64 *) *p) = cpu_to_le64(v);
		break;
	default:
		BUG();
	}

	*p += bytes;
}

static void bch2_alloc_unpack_v1(struct bkey_alloc_unpacked *out,
				 struct bkey_s_c k)
{
	const struct bch_alloc *in = bkey_s_c_to_alloc(k).v;
	const void *d = in->data;
	unsigned idx = 0;

	out->gen = in->gen;

#define x(_name, _bits) out->_name = alloc_field_v1_get(in, &d, idx++);
	BCH_ALLOC_FIELDS_V1()
#undef  x
}

static int bch2_alloc_unpack_v2(struct bkey_alloc_unpacked *out,
				struct bkey_s_c k)
{
	struct bkey_s_c_alloc_v2 a = bkey_s_c_to_alloc_v2(k);
	const u8 *in = a.v->data;
	const u8 *end = bkey_val_end(a);
	unsigned fieldnr = 0;
	int ret;
	u64 v;

	out->gen	= a.v->gen;
	out->oldest_gen	= a.v->oldest_gen;
	out->data_type	= a.v->data_type;

#define x(_name, _bits)							\
	if (fieldnr < a.v->nr_fields) {					\
		ret = bch2_varint_decode_fast(in, end, &v);		\
		if (ret < 0)						\
			return ret;					\
		in += ret;						\
	} else {							\
		v = 0;							\
	}								\
	out->_name = v;							\
	if (v != out->_name)						\
		return -1;						\
	fieldnr++;

	BCH_ALLOC_FIELDS_V2()
#undef  x
	return 0;
}

static int bch2_alloc_unpack_v3(struct bkey_alloc_unpacked *out,
				struct bkey_s_c k)
{
	struct bkey_s_c_alloc_v3 a = bkey_s_c_to_alloc_v3(k);
	const u8 *in = a.v->data;
	const u8 *end = bkey_val_end(a);
	unsigned fieldnr = 0;
	int ret;
	u64 v;

	out->gen	= a.v->gen;
	out->oldest_gen	= a.v->oldest_gen;
	out->data_type	= a.v->data_type;
	out->journal_seq = le64_to_cpu(a.v->journal_seq);

#define x(_name, _bits)							\
	if (fieldnr < a.v->nr_fields) {					\
		ret = bch2_varint_decode_fast(in, end, &v);		\
		if (ret < 0)						\
			return ret;					\
		in += ret;						\
	} else {							\
		v = 0;							\
	}								\
	out->_name = v;							\
	if (v != out->_name)						\
		return -1;						\
	fieldnr++;

	BCH_ALLOC_FIELDS_V2()
#undef  x
	return 0;
}

static void bch2_alloc_pack_v3(struct bkey_alloc_buf *dst,
			       const struct bkey_alloc_unpacked src)
{
	struct bkey_i_alloc_v3 *a = bkey_alloc_v3_init(&dst->k);
	unsigned nr_fields = 0, last_nonzero_fieldnr = 0;
	u8 *out = a->v.data;
	u8 *end = (void *) &dst[1];
	u8 *last_nonzero_field = out;
	unsigned bytes;

	a->k.p		= POS(src.dev, src.bucket);
	a->v.gen	= src.gen;
	a->v.oldest_gen	= src.oldest_gen;
	a->v.data_type	= src.data_type;
	a->v.journal_seq = cpu_to_le64(src.journal_seq);

#define x(_name, _bits)							\
	nr_fields++;							\
									\
	if (src._name) {						\
		out += bch2_varint_encode_fast(out, src._name);		\
									\
		last_nonzero_field = out;				\
		last_nonzero_fieldnr = nr_fields;			\
	} else {							\
		*out++ = 0;						\
	}

	BCH_ALLOC_FIELDS_V2()
#undef  x
	BUG_ON(out > end);

	out = last_nonzero_field;
	a->v.nr_fields = last_nonzero_fieldnr;

	bytes = (u8 *) out - (u8 *) &a->v;
	set_bkey_val_bytes(&a->k, bytes);
	memset_u64s_tail(&a->v, 0, bytes);
}

struct bkey_alloc_unpacked bch2_alloc_unpack(struct bkey_s_c k)
{
	struct bkey_alloc_unpacked ret = {
		.dev	= k.k->p.inode,
		.bucket	= k.k->p.offset,
		.gen	= 0,
	};

	switch (k.k->type) {
	case KEY_TYPE_alloc:
		bch2_alloc_unpack_v1(&ret, k);
		break;
	case KEY_TYPE_alloc_v2:
		bch2_alloc_unpack_v2(&ret, k);
		break;
	case KEY_TYPE_alloc_v3:
		bch2_alloc_unpack_v3(&ret, k);
		break;
	}

	return ret;
}

struct bkey_alloc_buf *bch2_alloc_pack(struct btree_trans *trans,
				       const struct bkey_alloc_unpacked src)
{
	struct bkey_alloc_buf *dst;

	dst = bch2_trans_kmalloc(trans, sizeof(struct bkey_alloc_buf));
	if (!IS_ERR(dst))
		bch2_alloc_pack_v3(dst, src);

	return dst;
}

int bch2_alloc_write(struct btree_trans *trans, struct btree_iter *iter,
		     struct bkey_alloc_unpacked *u, unsigned trigger_flags)
{
	struct bkey_alloc_buf *a = bch2_alloc_pack(trans, *u);

	/*
	 * Without BTREE_UPDATE_NO_KEY_CACHE_COHERENCY, we may end up updating
	 * the btree instead of the key cache - this can casue the allocator to
	 * self-deadlock, since updating the btree may require allocating new
	 * btree nodes:
	 */
	return PTR_ERR_OR_ZERO(a) ?:
		bch2_trans_update(trans, iter, &a->k, trigger_flags|
				 BTREE_UPDATE_NO_KEY_CACHE_COHERENCY);
}

static unsigned bch_alloc_v1_val_u64s(const struct bch_alloc *a)
{
	unsigned i, bytes = offsetof(struct bch_alloc, data);

	for (i = 0; i < ARRAY_SIZE(BCH_ALLOC_V1_FIELD_BYTES); i++)
		if (a->fields & (1 << i))
			bytes += BCH_ALLOC_V1_FIELD_BYTES[i];

	return DIV_ROUND_UP(bytes, sizeof(u64));
}

const char *bch2_alloc_v1_invalid(const struct bch_fs *c, struct bkey_s_c k)
{
	struct bkey_s_c_alloc a = bkey_s_c_to_alloc(k);

	if (k.k->p.inode >= c->sb.nr_devices ||
	    !c->devs[k.k->p.inode])
		return "invalid device";

	/* allow for unknown fields */
	if (bkey_val_u64s(a.k) < bch_alloc_v1_val_u64s(a.v))
		return "incorrect value size";

	return NULL;
}

const char *bch2_alloc_v2_invalid(const struct bch_fs *c, struct bkey_s_c k)
{
	struct bkey_alloc_unpacked u;

	if (k.k->p.inode >= c->sb.nr_devices ||
	    !c->devs[k.k->p.inode])
		return "invalid device";

	if (bch2_alloc_unpack_v2(&u, k))
		return "unpack error";

	return NULL;
}

const char *bch2_alloc_v3_invalid(const struct bch_fs *c, struct bkey_s_c k)
{
	struct bkey_alloc_unpacked u;

	if (k.k->p.inode >= c->sb.nr_devices ||
	    !c->devs[k.k->p.inode])
		return "invalid device";

	if (bch2_alloc_unpack_v3(&u, k))
		return "unpack error";

	return NULL;
}

void bch2_alloc_to_text(struct printbuf *out, struct bch_fs *c,
			   struct bkey_s_c k)
{
	struct bkey_alloc_unpacked u = bch2_alloc_unpack(k);

	pr_buf(out, "gen %u oldest_gen %u data_type %s journal_seq %llu",
	       u.gen, u.oldest_gen, bch2_data_types[u.data_type],
	       u.journal_seq);
#define x(_name, ...)	pr_buf(out, " " #_name " %llu", (u64) u._name);
	BCH_ALLOC_FIELDS_V2()
#undef  x
}

int bch2_alloc_read(struct bch_fs *c, bool gc, bool metadata_only)
{
	struct btree_trans trans;
	struct btree_iter iter;
	struct bkey_s_c k;
	struct bch_dev *ca;
	struct bucket *g;
	struct bkey_alloc_unpacked u;
	int ret;

	bch2_trans_init(&trans, c, 0, 0);

	for_each_btree_key(&trans, iter, BTREE_ID_alloc, POS_MIN,
			   BTREE_ITER_PREFETCH, k, ret) {
		ca = bch_dev_bkey_exists(c, k.k->p.inode);
		g = __bucket(ca, k.k->p.offset, gc);
		u = bch2_alloc_unpack(k);

		if (!gc)
			*bucket_gen(ca, k.k->p.offset) = u.gen;

		g->_mark.gen		= u.gen;
		g->io_time[READ]	= u.read_time;
		g->io_time[WRITE]	= u.write_time;
		g->oldest_gen		= !gc ? u.oldest_gen : u.gen;
		g->gen_valid		= 1;

		if (!gc ||
		    (metadata_only &&
		     (u.data_type == BCH_DATA_user ||
		      u.data_type == BCH_DATA_cached ||
		      u.data_type == BCH_DATA_parity))) {
			g->_mark.data_type	= u.data_type;
			g->_mark.dirty_sectors	= u.dirty_sectors;
			g->_mark.cached_sectors	= u.cached_sectors;
			g->_mark.stripe		= u.stripe != 0;
			g->stripe		= u.stripe;
			g->stripe_redundancy	= u.stripe_redundancy;
		}

	}
	bch2_trans_iter_exit(&trans, &iter);

	bch2_trans_exit(&trans);

	if (ret)
		bch_err(c, "error reading alloc info: %i", ret);

	return ret;
}

/* Bucket IO clocks: */

int bch2_bucket_io_time_reset(struct btree_trans *trans, unsigned dev,
			      size_t bucket_nr, int rw)
{
	struct bch_fs *c = trans->c;
	struct btree_iter iter;
	struct bkey_s_c k;
	struct bkey_alloc_unpacked u;
	u64 *time, now;
	int ret = 0;

	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc, POS(dev, bucket_nr),
			     BTREE_ITER_CACHED|
			     BTREE_ITER_INTENT);
	k = bch2_btree_iter_peek_slot(&iter);
	ret = bkey_err(k);
	if (ret)
		goto out;

	u = bch2_alloc_unpack(k);

	time = rw == READ ? &u.read_time : &u.write_time;
	now = atomic64_read(&c->io_clock[rw].now);
	if (*time == now)
		goto out;

	*time = now;

	ret   = bch2_alloc_write(trans, &iter, &u, 0) ?:
		bch2_trans_commit(trans, NULL, NULL, 0);
out:
	bch2_trans_iter_exit(trans, &iter);
	return ret;
}

/* Background allocator thread: */

/*
 * Scans for buckets to be invalidated, invalidates them, rewrites prios/gens
 * (marking them as invalidated on disk), then optionally issues discard
 * commands to the newly free buckets, then puts them on the various freelists.
 */

static bool bch2_can_invalidate_bucket(struct bch_dev *ca, size_t b,
				       struct bucket_mark m)
{
	u8 gc_gen;

	if (!is_available_bucket(m))
		return false;

	if (m.owned_by_allocator)
		return false;

	if (ca->buckets_nouse &&
	    test_bit(b, ca->buckets_nouse))
		return false;

	if (ca->new_fs_bucket_idx) {
		/*
		 * Device or filesystem is still being initialized, and we
		 * haven't fully marked superblocks & journal:
		 */
		if (is_superblock_bucket(ca, b))
			return false;

		if (b < ca->new_fs_bucket_idx)
			return false;
	}

	gc_gen = bucket_gc_gen(bucket(ca, b));

	ca->inc_gen_needs_gc		+= gc_gen >= BUCKET_GC_GEN_MAX / 2;
	ca->inc_gen_really_needs_gc	+= gc_gen >= BUCKET_GC_GEN_MAX;

	return gc_gen < BUCKET_GC_GEN_MAX;
}

/*
 * Determines what order we're going to reuse buckets, smallest bucket_key()
 * first.
 */

static unsigned bucket_sort_key(struct bucket *g, struct bucket_mark m,
				u64 now, u64 last_seq_ondisk)
{
	unsigned used = m.cached_sectors;

	if (used) {
		/*
		 * Prefer to keep buckets that have been read more recently, and
		 * buckets that have more data in them:
		 */
		u64 last_read = max_t(s64, 0, now - g->io_time[READ]);
		u32 last_read_scaled = max_t(u64, U32_MAX, div_u64(last_read, used));

		return -last_read_scaled;
	} else {
		/*
		 * Prefer to use buckets with smaller gc_gen so that we don't
		 * have to walk the btree and recalculate oldest_gen - but shift
		 * off the low bits so that buckets will still have equal sort
		 * keys when there's only a small difference, so that we can
		 * keep sequential buckets together:
		 */
		return bucket_gc_gen(g) >> 4;
	}
}

static inline int bucket_alloc_cmp(alloc_heap *h,
				   struct alloc_heap_entry l,
				   struct alloc_heap_entry r)
{
	return  cmp_int(l.key, r.key) ?:
		cmp_int(r.nr, l.nr) ?:
		cmp_int(l.bucket, r.bucket);
}

static inline int bucket_idx_cmp(const void *_l, const void *_r)
{
	const struct alloc_heap_entry *l = _l, *r = _r;

	return cmp_int(l->bucket, r->bucket);
}

static void find_reclaimable_buckets_lru(struct bch_fs *c, struct bch_dev *ca)
{
	struct bucket_array *buckets;
	struct alloc_heap_entry e = { 0 };
	u64 now, last_seq_ondisk;
	size_t b, i, nr = 0;

	down_read(&ca->bucket_lock);

	buckets = bucket_array(ca);
	ca->alloc_heap.used = 0;
	now = atomic64_read(&c->io_clock[READ].now);
	last_seq_ondisk = c->journal.flushed_seq_ondisk;

	/*
	 * Find buckets with lowest read priority, by building a maxheap sorted
	 * by read priority and repeatedly replacing the maximum element until
	 * all buckets have been visited.
	 */
	for (b = ca->mi.first_bucket; b < ca->mi.nbuckets; b++) {
		struct bucket *g = &buckets->b[b];
		struct bucket_mark m = READ_ONCE(g->mark);
		unsigned key = bucket_sort_key(g, m, now, last_seq_ondisk);

		cond_resched();

		if (!bch2_can_invalidate_bucket(ca, b, m))
			continue;

		if (!m.data_type &&
		    bch2_bucket_needs_journal_commit(&c->buckets_waiting_for_journal,
						     last_seq_ondisk,
						     ca->dev_idx, b)) {
			ca->buckets_waiting_on_journal++;
			continue;
		}

		if (e.nr && e.bucket + e.nr == b && e.key == key) {
			e.nr++;
		} else {
			if (e.nr)
				heap_add_or_replace(&ca->alloc_heap, e,
					-bucket_alloc_cmp, NULL);

			e = (struct alloc_heap_entry) {
				.bucket = b,
				.nr	= 1,
				.key	= key,
			};
		}
	}

	if (e.nr)
		heap_add_or_replace(&ca->alloc_heap, e,
				-bucket_alloc_cmp, NULL);

	for (i = 0; i < ca->alloc_heap.used; i++)
		nr += ca->alloc_heap.data[i].nr;

	while (nr - ca->alloc_heap.data[0].nr >= ALLOC_SCAN_BATCH(ca)) {
		nr -= ca->alloc_heap.data[0].nr;
		heap_pop(&ca->alloc_heap, e, -bucket_alloc_cmp, NULL);
	}

	up_read(&ca->bucket_lock);
}

static size_t find_reclaimable_buckets(struct bch_fs *c, struct bch_dev *ca)
{
	size_t i, nr = 0;

	ca->inc_gen_needs_gc			= 0;
	ca->inc_gen_really_needs_gc		= 0;
	ca->buckets_waiting_on_journal		= 0;

	find_reclaimable_buckets_lru(c, ca);

	heap_resort(&ca->alloc_heap, bucket_alloc_cmp, NULL);

	for (i = 0; i < ca->alloc_heap.used; i++)
		nr += ca->alloc_heap.data[i].nr;

	return nr;
}

static int bucket_invalidate_btree(struct btree_trans *trans,
				   struct bch_dev *ca, u64 b,
				   struct bkey_alloc_unpacked *u)
{
	struct bch_fs *c = trans->c;
	struct btree_iter iter;
	struct bkey_s_c k;
	int ret;

	bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc,
			     POS(ca->dev_idx, b),
			     BTREE_ITER_CACHED|
			     BTREE_ITER_INTENT);

	k = bch2_btree_iter_peek_slot(&iter);
	ret = bkey_err(k);
	if (ret)
		goto err;

	*u = bch2_alloc_unpack(k);
	u->gen++;
	u->data_type		= 0;
	u->dirty_sectors	= 0;
	u->cached_sectors	= 0;
	u->read_time		= atomic64_read(&c->io_clock[READ].now);
	u->write_time		= atomic64_read(&c->io_clock[WRITE].now);

	ret = bch2_alloc_write(trans, &iter, u,
			       BTREE_TRIGGER_BUCKET_INVALIDATE);
err:
	bch2_trans_iter_exit(trans, &iter);
	return ret;
}

static int bch2_invalidate_one_bucket(struct bch_fs *c, struct bch_dev *ca,
				      u64 *journal_seq, unsigned flags)
{
	struct bkey_alloc_unpacked u;
	size_t b;
	u64 commit_seq = 0;
	int ret = 0;

	/*
	 * If the read-only path is trying to shut down, we can't be generating
	 * new btree updates:
	 */
	if (test_bit(BCH_FS_ALLOCATOR_STOPPING, &c->flags))
		return 1;

	BUG_ON(!ca->alloc_heap.used ||
	       !ca->alloc_heap.data[0].nr);
	b = ca->alloc_heap.data[0].bucket;

	/* first, put on free_inc and mark as owned by allocator: */
	percpu_down_read(&c->mark_lock);

	bch2_mark_alloc_bucket(c, ca, b, true);

	spin_lock(&c->freelist_lock);
	verify_not_on_freelist(c, ca, b);
	BUG_ON(!fifo_push(&ca->free_inc, b));
	spin_unlock(&c->freelist_lock);

	percpu_up_read(&c->mark_lock);

	ret = bch2_trans_do(c, NULL, &commit_seq,
			    BTREE_INSERT_NOCHECK_RW|
			    BTREE_INSERT_NOFAIL|
			    BTREE_INSERT_JOURNAL_RESERVED|
			    flags,
			    bucket_invalidate_btree(&trans, ca, b, &u));

	if (!ret) {
		/* remove from alloc_heap: */
		struct alloc_heap_entry e, *top = ca->alloc_heap.data;

		top->bucket++;
		top->nr--;

		if (!top->nr)
			heap_pop(&ca->alloc_heap, e, bucket_alloc_cmp, NULL);

		/*
		 * If we invalidating cached data then we need to wait on the
		 * journal commit:
		 */
		if (u.data_type)
			*journal_seq = max(*journal_seq, commit_seq);

		/*
		 * We already waiting on u.alloc_seq when we filtered out
		 * buckets that need journal commit:
		 */
		BUG_ON(*journal_seq > u.journal_seq);
	} else {
		size_t b2;

		/* remove from free_inc: */
		percpu_down_read(&c->mark_lock);
		spin_lock(&c->freelist_lock);

		bch2_mark_alloc_bucket(c, ca, b, false);

		BUG_ON(!fifo_pop_back(&ca->free_inc, b2));
		BUG_ON(b != b2);

		spin_unlock(&c->freelist_lock);
		percpu_up_read(&c->mark_lock);
	}

	return ret < 0 ? ret : 0;
}

/*
 * Pull buckets off ca->alloc_heap, invalidate them, move them to ca->free_inc:
 */
static int bch2_invalidate_buckets(struct bch_fs *c, struct bch_dev *ca)
{
	u64 journal_seq = 0;
	int ret = 0;

	/* Only use nowait if we've already invalidated at least one bucket: */
	while (!ret &&
	       !fifo_full(&ca->free_inc) &&
	       ca->alloc_heap.used) {
		if (kthread_should_stop()) {
			ret = 1;
			break;
		}

		ret = bch2_invalidate_one_bucket(c, ca, &journal_seq,
				(!fifo_empty(&ca->free_inc)
				 ? BTREE_INSERT_NOWAIT : 0));
		/*
		 * We only want to batch up invalidates when they're going to
		 * require flushing the journal:
		 */
		if (!journal_seq)
			break;
	}

	/* If we used NOWAIT, don't return the error: */
	if (!fifo_empty(&ca->free_inc))
		ret = 0;
	if (ret < 0)
		bch_err(ca, "error invalidating buckets: %i", ret);
	if (ret)
		return ret;

	if (journal_seq)
		ret = bch2_journal_flush_seq(&c->journal, journal_seq);
	if (ret) {
		bch_err(ca, "journal error: %i", ret);
		return ret;
	}

	return 0;
}

static void alloc_thread_set_state(struct bch_dev *ca, unsigned new_state)
{
	if (ca->allocator_state != new_state) {
		ca->allocator_state = new_state;
		closure_wake_up(&ca->fs->freelist_wait);
	}
}

static int push_invalidated_bucket(struct bch_fs *c, struct bch_dev *ca, u64 b)
{
	unsigned i;
	int ret = 0;

	spin_lock(&c->freelist_lock);
	for (i = 0; i < RESERVE_NR; i++) {
		/*
		 * Don't strand buckets on the copygc freelist until
		 * after recovery is finished:
		 */
		if (i == RESERVE_MOVINGGC &&
		    !test_bit(BCH_FS_STARTED, &c->flags))
			continue;

		if (fifo_push(&ca->free[i], b)) {
			fifo_pop(&ca->free_inc, b);
			ret = 1;
			break;
		}
	}
	spin_unlock(&c->freelist_lock);

	ca->allocator_state = ret
		? ALLOCATOR_running
		: ALLOCATOR_blocked_full;
	closure_wake_up(&c->freelist_wait);
	return ret;
}

static void discard_one_bucket(struct bch_fs *c, struct bch_dev *ca, u64 b)
{
	if (ca->mi.discard &&
	    bdev_max_discard_sectors(ca->disk_sb.bdev))
		blkdev_issue_discard(ca->disk_sb.bdev, bucket_to_sector(ca, b),
				     ca->mi.bucket_size, GFP_NOFS);
}

static bool allocator_thread_running(struct bch_dev *ca)
{
	unsigned state = ca->mi.state == BCH_MEMBER_STATE_rw &&
		test_bit(BCH_FS_ALLOCATOR_RUNNING, &ca->fs->flags)
		? ALLOCATOR_running
		: ALLOCATOR_stopped;
	alloc_thread_set_state(ca, state);
	return state == ALLOCATOR_running;
}

static int buckets_available(struct bch_dev *ca, unsigned long gc_count)
{
	s64 available = dev_buckets_reclaimable(ca) -
		(gc_count == ca->fs->gc_count ? ca->inc_gen_really_needs_gc : 0);
	bool ret = available > 0;

	alloc_thread_set_state(ca, ret
			       ? ALLOCATOR_running
			       : ALLOCATOR_blocked);
	return ret;
}

/**
 * bch_allocator_thread - move buckets from free_inc to reserves
 *
 * The free_inc FIFO is populated by find_reclaimable_buckets(), and
 * the reserves are depleted by bucket allocation. When we run out
 * of free_inc, try to invalidate some buckets and write out
 * prios and gens.
 */
static int bch2_allocator_thread(void *arg)
{
	struct bch_dev *ca = arg;
	struct bch_fs *c = ca->fs;
	unsigned long gc_count = c->gc_count;
	size_t nr;
	int ret;

	set_freezable();

	while (1) {
		ret = kthread_wait_freezable(allocator_thread_running(ca));
		if (ret)
			goto stop;

		while (!ca->alloc_heap.used) {
			cond_resched();

			ret = kthread_wait_freezable(buckets_available(ca, gc_count));
			if (ret)
				goto stop;

			gc_count = c->gc_count;
			nr = find_reclaimable_buckets(c, ca);

			if (!nr && ca->buckets_waiting_on_journal) {
				ret = bch2_journal_flush(&c->journal);
				if (ret)
					goto stop;
			} else if (nr < (ca->mi.nbuckets >> 6) &&
				   ca->buckets_waiting_on_journal >= nr / 2) {
				bch2_journal_flush_async(&c->journal, NULL);
			}

			if ((ca->inc_gen_needs_gc >= ALLOC_SCAN_BATCH(ca) ||
			     ca->inc_gen_really_needs_gc) &&
			    c->gc_thread) {
				atomic_inc(&c->kick_gc);
				wake_up_process(c->gc_thread);
			}

			trace_alloc_scan(ca, nr, ca->inc_gen_needs_gc,
					 ca->inc_gen_really_needs_gc);
		}

		ret = bch2_invalidate_buckets(c, ca);
		if (ret)
			goto stop;

		while (!fifo_empty(&ca->free_inc)) {
			u64 b = fifo_peek(&ca->free_inc);

			discard_one_bucket(c, ca, b);

			ret = kthread_wait_freezable(push_invalidated_bucket(c, ca, b));
			if (ret)
				goto stop;
		}
	}
stop:
	alloc_thread_set_state(ca, ALLOCATOR_stopped);
	return 0;
}

/* Startup/shutdown (ro/rw): */

void bch2_recalc_capacity(struct bch_fs *c)
{
	struct bch_dev *ca;
	u64 capacity = 0, reserved_sectors = 0, gc_reserve;
	unsigned bucket_size_max = 0;
	unsigned long ra_pages = 0;
	unsigned i, j;

	lockdep_assert_held(&c->state_lock);

	for_each_online_member(ca, c, i) {
		struct backing_dev_info *bdi = ca->disk_sb.bdev->bd_disk->bdi;

		ra_pages += bdi->ra_pages;
	}

	bch2_set_ra_pages(c, ra_pages);

	for_each_rw_member(ca, c, i) {
		u64 dev_reserve = 0;

		/*
		 * We need to reserve buckets (from the number
		 * of currently available buckets) against
		 * foreground writes so that mainly copygc can
		 * make forward progress.
		 *
		 * We need enough to refill the various reserves
		 * from scratch - copygc will use its entire
		 * reserve all at once, then run against when
		 * its reserve is refilled (from the formerly
		 * available buckets).
		 *
		 * This reserve is just used when considering if
		 * allocations for foreground writes must wait -
		 * not -ENOSPC calculations.
		 */
		for (j = 0; j < RESERVE_NONE; j++)
			dev_reserve += ca->free[j].size;

		dev_reserve += 1;	/* btree write point */
		dev_reserve += 1;	/* copygc write point */
		dev_reserve += 1;	/* rebalance write point */

		dev_reserve *= ca->mi.bucket_size;

		capacity += bucket_to_sector(ca, ca->mi.nbuckets -
					     ca->mi.first_bucket);

		reserved_sectors += dev_reserve * 2;

		bucket_size_max = max_t(unsigned, bucket_size_max,
					ca->mi.bucket_size);
	}

	gc_reserve = c->opts.gc_reserve_bytes
		? c->opts.gc_reserve_bytes >> 9
		: div64_u64(capacity * c->opts.gc_reserve_percent, 100);

	reserved_sectors = max(gc_reserve, reserved_sectors);

	reserved_sectors = min(reserved_sectors, capacity);

	c->capacity = capacity - reserved_sectors;

	c->bucket_size_max = bucket_size_max;

	/* Wake up case someone was waiting for buckets */
	closure_wake_up(&c->freelist_wait);
}

static bool bch2_dev_has_open_write_point(struct bch_fs *c, struct bch_dev *ca)
{
	struct open_bucket *ob;
	bool ret = false;

	for (ob = c->open_buckets;
	     ob < c->open_buckets + ARRAY_SIZE(c->open_buckets);
	     ob++) {
		spin_lock(&ob->lock);
		if (ob->valid && !ob->on_partial_list &&
		    ob->dev == ca->dev_idx)
			ret = true;
		spin_unlock(&ob->lock);
	}

	return ret;
}

/* device goes ro: */
void bch2_dev_allocator_remove(struct bch_fs *c, struct bch_dev *ca)
{
	unsigned i;

	BUG_ON(ca->alloc_thread);

	/* First, remove device from allocation groups: */

	for (i = 0; i < ARRAY_SIZE(c->rw_devs); i++)
		clear_bit(ca->dev_idx, c->rw_devs[i].d);

	/*
	 * Capacity is calculated based off of devices in allocation groups:
	 */
	bch2_recalc_capacity(c);

	/* Next, close write points that point to this device... */
	for (i = 0; i < ARRAY_SIZE(c->write_points); i++)
		bch2_writepoint_stop(c, ca, &c->write_points[i]);

	bch2_writepoint_stop(c, ca, &c->copygc_write_point);
	bch2_writepoint_stop(c, ca, &c->rebalance_write_point);
	bch2_writepoint_stop(c, ca, &c->btree_write_point);

	mutex_lock(&c->btree_reserve_cache_lock);
	while (c->btree_reserve_cache_nr) {
		struct btree_alloc *a =
			&c->btree_reserve_cache[--c->btree_reserve_cache_nr];

		bch2_open_buckets_put(c, &a->ob);
	}
	mutex_unlock(&c->btree_reserve_cache_lock);

	while (1) {
		struct open_bucket *ob;

		spin_lock(&c->freelist_lock);
		if (!ca->open_buckets_partial_nr) {
			spin_unlock(&c->freelist_lock);
			break;
		}
		ob = c->open_buckets +
			ca->open_buckets_partial[--ca->open_buckets_partial_nr];
		ob->on_partial_list = false;
		spin_unlock(&c->freelist_lock);

		bch2_open_bucket_put(c, ob);
	}

	bch2_ec_stop_dev(c, ca);

	/*
	 * Wake up threads that were blocked on allocation, so they can notice
	 * the device can no longer be removed and the capacity has changed:
	 */
	closure_wake_up(&c->freelist_wait);

	/*
	 * journal_res_get() can block waiting for free space in the journal -
	 * it needs to notice there may not be devices to allocate from anymore:
	 */
	wake_up(&c->journal.wait);

	/* Now wait for any in flight writes: */

	closure_wait_event(&c->open_buckets_wait,
			   !bch2_dev_has_open_write_point(c, ca));
}

/* device goes rw: */
void bch2_dev_allocator_add(struct bch_fs *c, struct bch_dev *ca)
{
	unsigned i;

	for (i = 0; i < ARRAY_SIZE(c->rw_devs); i++)
		if (ca->mi.data_allowed & (1 << i))
			set_bit(ca->dev_idx, c->rw_devs[i].d);
}

void bch2_dev_allocator_quiesce(struct bch_fs *c, struct bch_dev *ca)
{
	if (ca->alloc_thread)
		closure_wait_event(&c->freelist_wait,
				   ca->allocator_state != ALLOCATOR_running);
}

/* stop allocator thread: */
void bch2_dev_allocator_stop(struct bch_dev *ca)
{
	struct task_struct *p;

	p = rcu_dereference_protected(ca->alloc_thread, 1);
	ca->alloc_thread = NULL;

	/*
	 * We need an rcu barrier between setting ca->alloc_thread = NULL and
	 * the thread shutting down to avoid bch2_wake_allocator() racing:
	 *
	 * XXX: it would be better to have the rcu barrier be asynchronous
	 * instead of blocking us here
	 */
	synchronize_rcu();

	if (p) {
		kthread_stop(p);
		put_task_struct(p);
	}
}

/* start allocator thread: */
int bch2_dev_allocator_start(struct bch_dev *ca)
{
	struct task_struct *p;

	/*
	 * allocator thread already started?
	 */
	if (ca->alloc_thread)
		return 0;

	p = kthread_create(bch2_allocator_thread, ca,
			   "bch-alloc/%s", ca->name);
	if (IS_ERR(p)) {
		bch_err(ca->fs, "error creating allocator thread: %li",
			PTR_ERR(p));
		return PTR_ERR(p);
	}

	get_task_struct(p);
	rcu_assign_pointer(ca->alloc_thread, p);
	wake_up_process(p);
	return 0;
}

void bch2_fs_allocator_background_init(struct bch_fs *c)
{
	spin_lock_init(&c->freelist_lock);
}