X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libbcachefs%2Fbtree_update_leaf.c;h=47623f3aa610e7d9dee7d9ed653dfc0bf1c09a92;hb=d34e731082d8fcd710c2af6377a3b7fa927c8451;hp=536ac9215f570c17c2b566da36a2b05fd14d0568;hpb=7f69c4161c31b8f43723a9ccad1a9a358f4e2e70;p=bcachefs-tools-debian

diff --git a/libbcachefs/btree_update_leaf.c b/libbcachefs/btree_update_leaf.c
index 536ac92..47623f3 100644
--- a/libbcachefs/btree_update_leaf.c
+++ b/libbcachefs/btree_update_leaf.c
@@ -6,269 +6,208 @@
 #include "btree_gc.h"
 #include "btree_io.h"
 #include "btree_iter.h"
+#include "btree_key_cache.h"
 #include "btree_locking.h"
 #include "buckets.h"
 #include "debug.h"
 #include "error.h"
-#include "extents.h"
+#include "extent_update.h"
 #include "journal.h"
 #include "journal_reclaim.h"
 #include "keylist.h"
+#include "recovery.h"
+#include "subvolume.h"
 #include "replicas.h"
 
+#include <linux/prefetch.h>
 #include <linux/sort.h>
 #include <trace/events/bcachefs.h>
 
-static inline bool same_leaf_as_prev(struct btree_trans *trans,
-				     unsigned sorted_idx)
+static int __must_check
+bch2_trans_update_by_path(struct btree_trans *, struct btree_path *,
+			  struct bkey_i *, enum btree_update_flags);
+
+static inline int btree_insert_entry_cmp(const struct btree_insert_entry *l,
+					 const struct btree_insert_entry *r)
+{
+	return   cmp_int(l->btree_id,	r->btree_id) ?:
+		 -cmp_int(l->level,	r->level) ?:
+		 bpos_cmp(l->k->k.p,	r->k->k.p);
+}
+
+static inline struct btree_path_level *insert_l(struct btree_insert_entry *i)
 {
-	struct btree_insert_entry *i = trans->updates +
-		trans->updates_sorted[sorted_idx];
-	struct btree_insert_entry *prev = sorted_idx
-		? trans->updates + trans->updates_sorted[sorted_idx - 1]
-		: NULL;
+	return i->path->l + i->level;
+}
 
-	return prev &&
-		i->iter->l[0].b == prev->iter->l[0].b;
+static inline bool same_leaf_as_prev(struct btree_trans *trans,
+				     struct btree_insert_entry *i)
+{
+	return i != trans->updates &&
+		insert_l(&i[0])->b == insert_l(&i[-1])->b;
 }
 
-#define trans_for_each_update_sorted(_trans, _i, _iter)			\
-	for (_iter = 0;							\
-	     _iter < _trans->nr_updates &&				\
-	     (_i = _trans->updates + _trans->updates_sorted[_iter], 1);	\
-	     _iter++)
+static inline bool same_leaf_as_next(struct btree_trans *trans,
+				     struct btree_insert_entry *i)
+{
+	return i + 1 < trans->updates + trans->nr_updates &&
+		insert_l(&i[0])->b == insert_l(&i[1])->b;
+}
 
-inline void bch2_btree_node_lock_for_insert(struct bch_fs *c, struct btree *b,
-					    struct btree_iter *iter)
+static inline void bch2_btree_node_prep_for_write(struct btree_trans *trans,
+						  struct btree_path *path,
+						  struct btree *b)
 {
-	bch2_btree_node_lock_write(b, iter);
+	struct bch_fs *c = trans->c;
+
+	if (path->cached)
+		return;
 
 	if (unlikely(btree_node_just_written(b)) &&
 	    bch2_btree_post_write_cleanup(c, b))
-		bch2_btree_iter_reinit_node(iter, b);
+		bch2_trans_node_reinit_iter(trans, b);
 
 	/*
 	 * If the last bset has been written, or if it's gotten too big - start
 	 * a new bset to insert into:
 	 */
 	if (want_new_bset(c, b))
-		bch2_btree_init_next(c, b, iter);
-}
-
-static void btree_trans_lock_write(struct btree_trans *trans, bool lock)
-{
-	struct bch_fs *c = trans->c;
-	struct btree_insert_entry *i;
-	unsigned iter;
-
-	trans_for_each_update_sorted(trans, i, iter) {
-		if (same_leaf_as_prev(trans, iter))
-			continue;
-
-		if (lock)
-			bch2_btree_node_lock_for_insert(c, i->iter->l[0].b, i->iter);
-		else
-			bch2_btree_node_unlock_write(i->iter->l[0].b, i->iter);
-	}
+		bch2_btree_init_next(trans, b);
 }
 
-static inline void btree_trans_sort_updates(struct btree_trans *trans)
+void bch2_btree_node_lock_for_insert(struct btree_trans *trans,
+				     struct btree_path *path,
+				     struct btree *b)
 {
-	struct btree_insert_entry *l, *r;
-	unsigned nr = 0, pos;
-
-	trans_for_each_update(trans, l) {
-		for (pos = 0; pos < nr; pos++) {
-			r = trans->updates + trans->updates_sorted[pos];
-
-			if (btree_iter_cmp(l->iter, r->iter) <= 0)
-				break;
-		}
-
-		memmove(&trans->updates_sorted[pos + 1],
-			&trans->updates_sorted[pos],
-			(nr - pos) * sizeof(trans->updates_sorted[0]));
-
-		trans->updates_sorted[pos] = l - trans->updates;
-		nr++;
-	}
-
-	BUG_ON(nr != trans->nr_updates);
+	bch2_btree_node_lock_write(trans, path, b);
+	bch2_btree_node_prep_for_write(trans, path, b);
 }
 
 /* Inserting into a given leaf node (last stage of insert): */
 
 /* Handle overwrites and do insert, for non extents: */
-bool bch2_btree_bset_insert_key(struct btree_iter *iter,
+bool bch2_btree_bset_insert_key(struct btree_trans *trans,
+				struct btree_path *path,
 				struct btree *b,
 				struct btree_node_iter *node_iter,
 				struct bkey_i *insert)
 {
-	const struct bkey_format *f = &b->format;
 	struct bkey_packed *k;
-	unsigned clobber_u64s;
+	unsigned clobber_u64s = 0, new_u64s = 0;
 
 	EBUG_ON(btree_node_just_written(b));
 	EBUG_ON(bset_written(b, btree_bset_last(b)));
 	EBUG_ON(bkey_deleted(&insert->k) && bkey_val_u64s(&insert->k));
-	EBUG_ON(bkey_cmp(bkey_start_pos(&insert->k), b->data->min_key) < 0 ||
-		bkey_cmp(insert->k.p, b->data->max_key) > 0);
+	EBUG_ON(bpos_cmp(insert->k.p, b->data->min_key) < 0);
+	EBUG_ON(bpos_cmp(insert->k.p, b->data->max_key) > 0);
+	EBUG_ON(insert->k.u64s >
+		bch_btree_keys_u64s_remaining(trans->c, b));
 
 	k = bch2_btree_node_iter_peek_all(node_iter, b);
-	if (k && !bkey_cmp_packed(b, k, &insert->k)) {
-		BUG_ON(bkey_whiteout(k));
-
-		if (!bkey_written(b, k) &&
-		    bkey_val_u64s(&insert->k) == bkeyp_val_u64s(f, k) &&
-		    !bkey_whiteout(&insert->k)) {
-			k->type = insert->k.type;
-			memcpy_u64s(bkeyp_val(f, k), &insert->v,
-				    bkey_val_u64s(&insert->k));
-			return true;
-		}
+	if (k && bkey_cmp_left_packed(b, k, &insert->k.p))
+		k = NULL;
 
-		insert->k.needs_whiteout = k->needs_whiteout;
+	/* @k is the key being overwritten/deleted, if any: */
+	EBUG_ON(k && bkey_deleted(k));
 
+	/* Deleting, but not found? nothing to do: */
+	if (bkey_deleted(&insert->k) && !k)
+		return false;
+
+	if (bkey_deleted(&insert->k)) {
+		/* Deleting: */
 		btree_account_key_drop(b, k);
+		k->type = KEY_TYPE_deleted;
+
+		if (k->needs_whiteout)
+			push_whiteout(trans->c, b, insert->k.p);
+		k->needs_whiteout = false;
 
 		if (k >= btree_bset_last(b)->start) {
 			clobber_u64s = k->u64s;
-
-			/*
-			 * If we're deleting, and the key we're deleting doesn't
-			 * need a whiteout (it wasn't overwriting a key that had
-			 * been written to disk) - just delete it:
-			 */
-			if (bkey_whiteout(&insert->k) && !k->needs_whiteout) {
-				bch2_bset_delete(b, k, clobber_u64s);
-				bch2_btree_node_iter_fix(iter, b, node_iter,
-							 k, clobber_u64s, 0);
-				return true;
-			}
-
-			goto overwrite;
+			bch2_bset_delete(b, k, clobber_u64s);
+			goto fix_iter;
+		} else {
+			bch2_btree_path_fix_key_modified(trans, b, k);
 		}
 
+		return true;
+	}
+
+	if (k) {
+		/* Overwriting: */
+		btree_account_key_drop(b, k);
 		k->type = KEY_TYPE_deleted;
-		bch2_btree_node_iter_fix(iter, b, node_iter, k,
-					 k->u64s, k->u64s);
 
-		if (bkey_whiteout(&insert->k)) {
-			reserve_whiteout(b, k);
-			return true;
+		insert->k.needs_whiteout = k->needs_whiteout;
+		k->needs_whiteout = false;
+
+		if (k >= btree_bset_last(b)->start) {
+			clobber_u64s = k->u64s;
+			goto overwrite;
 		} else {
-			k->needs_whiteout = false;
+			bch2_btree_path_fix_key_modified(trans, b, k);
 		}
-	} else {
-		/*
-		 * Deleting, but the key to delete wasn't found - nothing to do:
-		 */
-		if (bkey_whiteout(&insert->k))
-			return false;
-
-		insert->k.needs_whiteout = false;
 	}
 
 	k = bch2_btree_node_iter_bset_pos(node_iter, b, bset_tree_last(b));
-	clobber_u64s = 0;
 overwrite:
 	bch2_bset_insert(b, node_iter, k, insert, clobber_u64s);
-	bch2_btree_node_iter_fix(iter, b, node_iter, k,
-				 clobber_u64s, k->u64s);
+	new_u64s = k->u64s;
+fix_iter:
+	if (clobber_u64s != new_u64s)
+		bch2_btree_node_iter_fix(trans, path, b, node_iter, k,
+					 clobber_u64s, new_u64s);
 	return true;
 }
 
-static void __btree_node_flush(struct journal *j, struct journal_entry_pin *pin,
+static int __btree_node_flush(struct journal *j, struct journal_entry_pin *pin,
 			       unsigned i, u64 seq)
 {
 	struct bch_fs *c = container_of(j, struct bch_fs, journal);
 	struct btree_write *w = container_of(pin, struct btree_write, journal);
 	struct btree *b = container_of(w, struct btree, writes[i]);
+	unsigned long old, new, v;
+	unsigned idx = w - b->writes;
+
+	six_lock_read(&b->c.lock, NULL, NULL);
+	v = READ_ONCE(b->flags);
+
+	do {
+		old = new = v;
+
+		if (!(old & (1 << BTREE_NODE_dirty)) ||
+		    !!(old & (1 << BTREE_NODE_write_idx)) != idx ||
+		    w->journal.seq != seq)
+			break;
+
+		new |= 1 << BTREE_NODE_need_write;
+	} while ((v = cmpxchg(&b->flags, old, new)) != old);
 
-	btree_node_lock_type(c, b, SIX_LOCK_read);
-	bch2_btree_node_write_cond(c, b,
-		(btree_current_write(b) == w && w->journal.seq == seq));
-	six_unlock_read(&b->lock);
+	btree_node_write_if_need(c, b, SIX_LOCK_read);
+	six_unlock_read(&b->c.lock);
+	return 0;
 }
 
-static void btree_node_flush0(struct journal *j, struct journal_entry_pin *pin, u64 seq)
+static int btree_node_flush0(struct journal *j, struct journal_entry_pin *pin, u64 seq)
 {
 	return __btree_node_flush(j, pin, 0, seq);
 }
 
-static void btree_node_flush1(struct journal *j, struct journal_entry_pin *pin, u64 seq)
+static int btree_node_flush1(struct journal *j, struct journal_entry_pin *pin, u64 seq)
 {
 	return __btree_node_flush(j, pin, 1, seq);
 }
 
-static inline void __btree_journal_key(struct btree_trans *trans,
-				       enum btree_id btree_id,
-				       struct bkey_i *insert)
-{
-	struct journal *j = &trans->c->journal;
-	u64 seq = trans->journal_res.seq;
-	bool needs_whiteout = insert->k.needs_whiteout;
-
-	/* ick */
-	insert->k.needs_whiteout = false;
-	bch2_journal_add_keys(j, &trans->journal_res,
-			      btree_id, insert);
-	insert->k.needs_whiteout = needs_whiteout;
-
-	bch2_journal_set_has_inode(j, &trans->journal_res,
-				   insert->k.p.inode);
-
-	if (trans->journal_seq)
-		*trans->journal_seq = seq;
-}
-
-void bch2_btree_journal_key(struct btree_trans *trans,
-			   struct btree_iter *iter,
-			   struct bkey_i *insert)
+inline void bch2_btree_add_journal_pin(struct bch_fs *c,
+				       struct btree *b, u64 seq)
 {
-	struct bch_fs *c = trans->c;
-	struct journal *j = &c->journal;
-	struct btree *b = iter->l[0].b;
 	struct btree_write *w = btree_current_write(b);
 
-	EBUG_ON(iter->level || b->level);
-	EBUG_ON(trans->journal_res.ref !=
-		!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY));
-
-	if (likely(!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY))) {
-		__btree_journal_key(trans, iter->btree_id, insert);
-		btree_bset_last(b)->journal_seq =
-			cpu_to_le64(trans->journal_res.seq);
-	}
-
-	if (unlikely(!journal_pin_active(&w->journal))) {
-		u64 seq = likely(!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY))
-			? trans->journal_res.seq
-			: j->replay_journal_seq;
-
-		bch2_journal_pin_add(j, seq, &w->journal,
-				     btree_node_write_idx(b) == 0
-				     ? btree_node_flush0
-				     : btree_node_flush1);
-	}
-
-	if (unlikely(!btree_node_dirty(b)))
-		set_btree_node_dirty(b);
-}
-
-static void bch2_insert_fixup_key(struct btree_trans *trans,
-				  struct btree_insert_entry *insert)
-{
-	struct btree_iter *iter = insert->iter;
-	struct btree_iter_level *l = &iter->l[0];
-
-	EBUG_ON(iter->level);
-	EBUG_ON(insert->k->k.u64s >
-		bch_btree_keys_u64s_remaining(trans->c, l->b));
-
-	if (bch2_btree_bset_insert_key(iter, l->b, &l->iter,
-				       insert->k))
-		bch2_btree_journal_key(trans, iter, insert->k);
+	bch2_journal_pin_add(&c->journal, seq, &w->journal,
+			     btree_node_write_idx(b) == 0
+			     ? btree_node_flush0
+			     : btree_node_flush1);
 }
 
 /**
@@ -278,19 +217,27 @@ static void btree_insert_key_leaf(struct btree_trans *trans,
 				  struct btree_insert_entry *insert)
 {
 	struct bch_fs *c = trans->c;
-	struct btree_iter *iter = insert->iter;
-	struct btree *b = iter->l[0].b;
-	int old_u64s = le16_to_cpu(btree_bset_last(b)->u64s);
+	struct btree *b = insert_l(insert)->b;
+	struct bset_tree *t = bset_tree_last(b);
+	struct bset *i = bset(b, t);
+	int old_u64s = bset_u64s(t);
 	int old_live_u64s = b->nr.live_u64s;
 	int live_u64s_added, u64s_added;
 
-	if (!btree_node_is_extents(b))
-		bch2_insert_fixup_key(trans, insert);
-	else
-		bch2_insert_fixup_extent(trans, insert);
+	if (unlikely(!bch2_btree_bset_insert_key(trans, insert->path, b,
+					&insert_l(insert)->iter, insert->k)))
+		return;
+
+	i->journal_seq = cpu_to_le64(max(trans->journal_res.seq,
+					 le64_to_cpu(i->journal_seq)));
+
+	bch2_btree_add_journal_pin(c, b, trans->journal_res.seq);
+
+	if (unlikely(!btree_node_dirty(b)))
+		set_btree_node_dirty_acct(c, b);
 
 	live_u64s_added = (int) b->nr.live_u64s - old_live_u64s;
-	u64s_added = (int) le16_to_cpu(btree_bset_last(b)->u64s) - old_u64s;
+	u64s_added = (int) bset_u64s(t) - old_u64s;
 
 	if (b->sib_u64s[0] != U16_MAX && live_u64s_added < 0)
 		b->sib_u64s[0] = max(0, (int) b->sib_u64s[0] + live_u64s_added);
@@ -299,50 +246,34 @@ static void btree_insert_key_leaf(struct btree_trans *trans,
 
 	if (u64s_added > live_u64s_added &&
 	    bch2_maybe_compact_whiteouts(c, b))
-		bch2_btree_iter_reinit_node(iter, b);
-
-	trace_btree_insert_key(c, b, insert->k);
+		bch2_trans_node_reinit_iter(trans, b);
 }
 
+/* Cached btree updates: */
+
 /* Normal update interface: */
 
 static inline void btree_insert_entry_checks(struct btree_trans *trans,
 					     struct btree_insert_entry *i)
 {
-	struct bch_fs *c = trans->c;
-
-	BUG_ON(i->iter->level);
-	BUG_ON(bkey_cmp(bkey_start_pos(&i->k->k), i->iter->pos));
-	EBUG_ON((i->iter->flags & BTREE_ITER_IS_EXTENTS) &&
-		bkey_cmp(i->k->k.p, i->iter->l[0].b->key.k.p) > 0);
-	EBUG_ON((i->iter->flags & BTREE_ITER_IS_EXTENTS) &&
-		!(trans->flags & BTREE_INSERT_ATOMIC));
-
-	BUG_ON(debug_check_bkeys(c) &&
-	       !bkey_deleted(&i->k->k) &&
-	       bch2_bkey_invalid(c, bkey_i_to_s_c(i->k), i->iter->btree_id));
+	BUG_ON(bpos_cmp(i->k->k.p, i->path->pos));
+	BUG_ON(i->cached	!= i->path->cached);
+	BUG_ON(i->level		!= i->path->level);
+	BUG_ON(i->btree_id	!= i->path->btree_id);
+	EBUG_ON(!i->level &&
+		!(i->flags & BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) &&
+		test_bit(JOURNAL_REPLAY_DONE, &trans->c->journal.flags) &&
+		i->k->k.p.snapshot &&
+		bch2_snapshot_internal_node(trans->c, i->k->k.p.snapshot));
 }
 
-static int bch2_trans_journal_preres_get(struct btree_trans *trans)
+static noinline int
+bch2_trans_journal_preres_get_cold(struct btree_trans *trans, unsigned u64s,
+				   unsigned long trace_ip)
 {
 	struct bch_fs *c = trans->c;
-	struct btree_insert_entry *i;
-	unsigned u64s = 0;
 	int ret;
 
-	trans_for_each_update(trans, i)
-		if (0)
-			u64s += jset_u64s(i->k->k.u64s);
-
-	if (!u64s)
-		return 0;
-
-	ret = bch2_journal_preres_get(&c->journal,
-			&trans->journal_preres, u64s,
-			JOURNAL_RES_GET_NONBLOCK);
-	if (ret != -EAGAIN)
-		return ret;
-
 	bch2_trans_unlock(trans);
 
 	ret = bch2_journal_preres_get(&c->journal,
@@ -351,15 +282,15 @@ static int bch2_trans_journal_preres_get(struct btree_trans *trans)
 		return ret;
 
 	if (!bch2_trans_relock(trans)) {
-		trace_trans_restart_journal_preres_get(trans->ip);
+		trace_trans_restart_journal_preres_get(trans->fn, trace_ip);
 		return -EINTR;
 	}
 
 	return 0;
 }
 
-static int bch2_trans_journal_res_get(struct btree_trans *trans,
-				      unsigned flags)
+static inline int bch2_trans_journal_res_get(struct btree_trans *trans,
+					     unsigned flags)
 {
 	struct bch_fs *c = trans->c;
 	int ret;
@@ -373,474 +304,1322 @@ static int bch2_trans_journal_res_get(struct btree_trans *trans,
 	return ret == -EAGAIN ? BTREE_INSERT_NEED_JOURNAL_RES : ret;
 }
 
-static enum btree_insert_ret
-btree_key_can_insert(struct btree_trans *trans,
-		     struct btree_insert_entry *insert,
-		     unsigned *u64s)
+#define JSET_ENTRY_LOG_U64s		4
+
+static noinline void journal_transaction_name(struct btree_trans *trans)
 {
 	struct bch_fs *c = trans->c;
-	struct btree *b = insert->iter->l[0].b;
-	static enum btree_insert_ret ret;
-
-	if (unlikely(btree_node_fake(b)))
-		return BTREE_INSERT_BTREE_NODE_FULL;
+	struct jset_entry *entry = journal_res_entry(&c->journal, &trans->journal_res);
+	struct jset_entry_log *l = container_of(entry, struct jset_entry_log, entry);
+	unsigned u64s = JSET_ENTRY_LOG_U64s - 1;
+	unsigned b, buflen = u64s * sizeof(u64);
+
+	l->entry.u64s		= cpu_to_le16(u64s);
+	l->entry.btree_id	= 0;
+	l->entry.level		= 0;
+	l->entry.type		= BCH_JSET_ENTRY_log;
+	l->entry.pad[0]		= 0;
+	l->entry.pad[1]		= 0;
+	l->entry.pad[2]		= 0;
+	b = min_t(unsigned, strlen(trans->fn), buflen);
+	memcpy(l->d, trans->fn, b);
+	while (b < buflen)
+		l->d[b++] = '\0';
+
+	trans->journal_res.offset	+= JSET_ENTRY_LOG_U64s;
+	trans->journal_res.u64s		-= JSET_ENTRY_LOG_U64s;
+}
 
-	ret = !btree_node_is_extents(b)
-		? BTREE_INSERT_OK
-		: bch2_extent_can_insert(trans, insert, u64s);
-	if (ret)
-		return ret;
+static inline enum btree_insert_ret
+btree_key_can_insert(struct btree_trans *trans,
+		     struct btree *b,
+		     unsigned u64s)
+{
+	struct bch_fs *c = trans->c;
 
-	if (*u64s > bch_btree_keys_u64s_remaining(c, b))
+	if (!bch2_btree_node_insert_fits(c, b, u64s))
 		return BTREE_INSERT_BTREE_NODE_FULL;
 
 	return BTREE_INSERT_OK;
 }
 
-static int btree_trans_check_can_insert(struct btree_trans *trans,
-					struct btree_insert_entry **stopped_at)
+static enum btree_insert_ret
+btree_key_can_insert_cached(struct btree_trans *trans,
+			    struct btree_path *path,
+			    unsigned u64s)
 {
-	struct btree_insert_entry *i;
-	unsigned iter, u64s = 0;
-	int ret;
+	struct bch_fs *c = trans->c;
+	struct bkey_cached *ck = (void *) path->l[0].b;
+	unsigned new_u64s;
+	struct bkey_i *new_k;
 
-	trans_for_each_update_sorted(trans, i, iter) {
-		/* Multiple inserts might go to same leaf: */
-		if (!same_leaf_as_prev(trans, iter))
-			u64s = 0;
+	EBUG_ON(path->level);
 
-		u64s += i->k->k.u64s;
-		ret = btree_key_can_insert(trans, i, &u64s);
-		if (ret) {
-			*stopped_at = i;
-			return ret;
-		}
+	if (!test_bit(BKEY_CACHED_DIRTY, &ck->flags) &&
+	    bch2_btree_key_cache_must_wait(c) &&
+	    !(trans->flags & BTREE_INSERT_JOURNAL_RECLAIM))
+		return BTREE_INSERT_NEED_JOURNAL_RECLAIM;
+
+	/*
+	 * bch2_varint_decode can read past the end of the buffer by at most 7
+	 * bytes (it won't be used):
+	 */
+	u64s += 1;
+
+	if (u64s <= ck->u64s)
+		return BTREE_INSERT_OK;
+
+	new_u64s	= roundup_pow_of_two(u64s);
+	new_k		= krealloc(ck->k, new_u64s * sizeof(u64), GFP_NOFS);
+	if (!new_k) {
+		bch_err(c, "error allocating memory for key cache key, btree %s u64s %u",
+			bch2_btree_ids[path->btree_id], new_u64s);
+		return -ENOMEM;
 	}
 
-	return 0;
+	ck->u64s	= new_u64s;
+	ck->k		= new_k;
+	/*
+	 * Keys returned by peek() are no longer valid pointers, so we need a
+	 * transaction restart:
+	 */
+	trace_trans_restart_key_cache_key_realloced(trans->fn, _RET_IP_,
+					     path->btree_id, &path->pos);
+	/*
+	 * Not using btree_trans_restart() because we can't unlock here, we have
+	 * write locks held:
+	 */
+	trans->restarted = true;
+	return -EINTR;
 }
 
 static inline void do_btree_insert_one(struct btree_trans *trans,
-				       struct btree_insert_entry *insert)
+				       struct btree_insert_entry *i)
 {
-	btree_insert_key_leaf(trans, insert);
-}
+	struct bch_fs *c = trans->c;
+	struct journal *j = &c->journal;
 
-static inline bool update_triggers_transactional(struct btree_trans *trans,
-						 struct btree_insert_entry *i)
-{
-	return likely(!(trans->flags & BTREE_INSERT_MARK_INMEM)) &&
-		(i->iter->btree_id == BTREE_ID_EXTENTS ||
-		 i->iter->btree_id == BTREE_ID_INODES ||
-		 i->iter->btree_id == BTREE_ID_REFLINK);
-}
+	EBUG_ON(trans->journal_res.ref !=
+		!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY));
 
-static inline bool update_has_triggers(struct btree_trans *trans,
-				       struct btree_insert_entry *i)
-{
-	return likely(!(trans->flags & BTREE_INSERT_NOMARK)) &&
-		btree_node_type_needs_gc(i->iter->btree_id);
+	i->k->k.needs_whiteout = false;
+
+	if (!i->cached)
+		btree_insert_key_leaf(trans, i);
+	else
+		bch2_btree_insert_key_cached(trans, i->path, i->k);
+
+	if (likely(!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY))) {
+		bch2_journal_add_keys(j, &trans->journal_res,
+				      i->btree_id,
+				      i->level,
+				      i->k);
+
+		if (trans->journal_seq)
+			*trans->journal_seq = trans->journal_res.seq;
+	}
 }
 
-/*
- * Get journal reservation, take write locks, and attempt to do btree update(s):
- */
-static inline int do_btree_insert_at(struct btree_trans *trans,
-				     struct btree_insert_entry **stopped_at)
+/* Triggers: */
+
+static int run_one_mem_trigger(struct btree_trans *trans,
+			       struct btree_insert_entry *i,
+			       unsigned flags)
 {
-	struct bch_fs *c = trans->c;
-	struct bch_fs_usage *fs_usage = NULL;
-	struct btree_insert_entry *i;
-	unsigned mark_flags = trans->flags & BTREE_INSERT_BUCKET_INVALIDATE
-		? BCH_BUCKET_MARK_BUCKET_INVALIDATE
-		: 0;
+	struct bkey_s_c old = { &i->old_k, i->old_v };
+	struct bkey_i *new = i->k;
 	int ret;
 
-	trans_for_each_update(trans, i)
-		BUG_ON(i->iter->uptodate >= BTREE_ITER_NEED_RELOCK);
+	if (unlikely(flags & BTREE_TRIGGER_NORUN))
+		return 0;
 
-	/*
-	 * note: running triggers will append more updates to the list of
-	 * updates as we're walking it:
-	 */
-	trans_for_each_update(trans, i)
-		if (update_has_triggers(trans, i) &&
-		    update_triggers_transactional(trans, i)) {
-			ret = bch2_trans_mark_update(trans, i->iter, i->k);
-			if (ret == -EINTR)
-				trace_trans_restart_mark(trans->ip);
-			if (ret)
-				goto out_clear_replicas;
-		}
+	if (!btree_node_type_needs_gc(i->btree_id))
+		return 0;
 
-	if (IS_ENABLED(CONFIG_BCACHEFS_DEBUG))
-		trans_for_each_update(trans, i)
-			btree_insert_entry_checks(trans, i);
-	bch2_btree_trans_verify_locks(trans);
+	if (bch2_bkey_ops[old.k->type].atomic_trigger ==
+	    bch2_bkey_ops[i->k->k.type].atomic_trigger &&
+	    ((1U << old.k->type) & BTREE_TRIGGER_WANTS_OLD_AND_NEW)) {
+		ret   = bch2_mark_key(trans, old, bkey_i_to_s_c(new),
+				BTREE_TRIGGER_INSERT|BTREE_TRIGGER_OVERWRITE|flags);
+	} else {
+		struct bkey		_deleted = KEY(0, 0, 0);
+		struct bkey_s_c		deleted = (struct bkey_s_c) { &_deleted, NULL };
 
-	/*
-	 * No more updates can be added - sort updates so we can take write
-	 * locks in the correct order:
-	 */
-	btree_trans_sort_updates(trans);
+		_deleted.p = i->path->pos;
+
+		ret   = bch2_mark_key(trans, deleted, bkey_i_to_s_c(new),
+				BTREE_TRIGGER_INSERT|flags) ?:
+			bch2_mark_key(trans, old, deleted,
+				BTREE_TRIGGER_OVERWRITE|flags);
+	}
 
-	btree_trans_lock_write(trans, true);
+	return ret;
+}
 
-	if (race_fault()) {
-		ret = -EINTR;
-		trace_trans_restart_fault_inject(trans->ip);
-		goto out;
+static int run_one_trans_trigger(struct btree_trans *trans, struct btree_insert_entry *i,
+			   bool overwrite)
+{
+	struct bkey_s_c old = { &i->old_k, i->old_v };
+	int ret = 0;
+
+	if ((i->flags & BTREE_TRIGGER_NORUN) ||
+	    !(BTREE_NODE_TYPE_HAS_TRANS_TRIGGERS & (1U << i->bkey_type)))
+		return 0;
+
+	if (!overwrite) {
+		if (i->insert_trigger_run)
+			return 0;
+
+		BUG_ON(i->overwrite_trigger_run);
+		i->insert_trigger_run = true;
+	} else {
+		if (i->overwrite_trigger_run)
+			return 0;
+
+		BUG_ON(!i->insert_trigger_run);
+		i->overwrite_trigger_run = true;
 	}
 
-	/*
-	 * Check if the insert will fit in the leaf node with the write lock
-	 * held, otherwise another thread could write the node changing the
-	 * amount of space available:
-	 */
-	ret = btree_trans_check_can_insert(trans, stopped_at);
-	if (ret)
-		goto out;
+	if (overwrite) {
+		ret = bch2_trans_mark_old(trans, old, i->flags);
+	} else if (bch2_bkey_ops[old.k->type].trans_trigger ==
+		   bch2_bkey_ops[i->k->k.type].trans_trigger &&
+	    ((1U << old.k->type) & BTREE_TRIGGER_WANTS_OLD_AND_NEW)) {
+		i->overwrite_trigger_run = true;
+		ret = bch2_trans_mark_key(trans, old, i->k,
+				BTREE_TRIGGER_INSERT|BTREE_TRIGGER_OVERWRITE|i->flags);
+	} else {
+		ret = bch2_trans_mark_new(trans, i->k, i->flags);
+	}
 
-	trans_for_each_update(trans, i) {
-		if (!btree_node_type_needs_gc(i->iter->btree_id))
-			continue;
+	if (ret == -EINTR)
+		trace_trans_restart_mark(trans->fn, _RET_IP_,
+					 i->btree_id, &i->path->pos);
+	return ret ?: 1;
+}
 
-		if (!fs_usage) {
-			percpu_down_read(&c->mark_lock);
-			fs_usage = bch2_fs_usage_scratch_get(c);
-		}
+static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
+			      struct btree_insert_entry *btree_id_start)
+{
+	struct btree_insert_entry *i;
+	bool trans_trigger_run;
+	int ret, overwrite;
 
-		if (!bch2_bkey_replicas_marked_locked(c,
-			bkey_i_to_s_c(i->k), true)) {
-			ret = BTREE_INSERT_NEED_MARK_REPLICAS;
-			goto out;
-		}
+	for (overwrite = 0; overwrite < 2; overwrite++) {
+
+		/*
+		 * Running triggers will append more updates to the list of updates as
+		 * we're walking it:
+		 */
+		do {
+			trans_trigger_run = false;
+
+			for (i = btree_id_start;
+			     i < trans->updates + trans->nr_updates && i->btree_id <= btree_id;
+			     i++) {
+				if (i->btree_id != btree_id)
+					continue;
+
+				ret = run_one_trans_trigger(trans, i, overwrite);
+				if (ret < 0)
+					return ret;
+				if (ret)
+					trans_trigger_run = true;
+			}
+		} while (trans_trigger_run);
 	}
 
+	return 0;
+}
+
+static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
+{
+	struct btree_insert_entry *i = NULL, *btree_id_start = trans->updates;
+	unsigned btree_id = 0;
+	int ret = 0;
+
 	/*
-	 * Don't get journal reservation until after we know insert will
-	 * succeed:
+	 *
+	 * For a given btree, this algorithm runs insert triggers before
+	 * overwrite triggers: this is so that when extents are being moved
+	 * (e.g. by FALLOCATE_FL_INSERT_RANGE), we don't drop references before
+	 * they are re-added.
 	 */
-	if (likely(!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY))) {
-		trans->journal_u64s = 0;
+	for (btree_id = 0; btree_id < BTREE_ID_NR; btree_id++) {
+		if (btree_id == BTREE_ID_alloc)
+			continue;
 
-		trans_for_each_update(trans, i)
-			trans->journal_u64s += jset_u64s(i->k->k.u64s);
+		while (btree_id_start < trans->updates + trans->nr_updates &&
+		       btree_id_start->btree_id < btree_id)
+			btree_id_start++;
 
-		ret = bch2_trans_journal_res_get(trans, JOURNAL_RES_GET_NONBLOCK);
+		ret = run_btree_triggers(trans, btree_id, btree_id_start);
 		if (ret)
-			goto out;
+			return ret;
 	}
 
-	if (!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY)) {
-		if (journal_seq_verify(c))
-			trans_for_each_update(trans, i)
-				i->k->k.version.lo = trans->journal_res.seq;
-		else if (inject_invalid_keys(c))
-			trans_for_each_update(trans, i)
-				i->k->k.version = MAX_VERSION;
+	trans_for_each_update(trans, i) {
+		if (i->btree_id > BTREE_ID_alloc)
+			break;
+		if (i->btree_id == BTREE_ID_alloc) {
+			ret = run_btree_triggers(trans, BTREE_ID_alloc, i);
+			if (ret)
+				return ret;
+			break;
+		}
 	}
 
 	trans_for_each_update(trans, i)
-		if (update_has_triggers(trans, i) &&
-		    !update_triggers_transactional(trans, i))
-			bch2_mark_update(trans, i, fs_usage, mark_flags);
-
-	if (fs_usage && trans->fs_usage_deltas)
-		bch2_replicas_delta_list_apply(c, fs_usage,
-					       trans->fs_usage_deltas);
-
-	if (fs_usage)
-		bch2_trans_fs_usage_apply(trans, fs_usage);
-
-	if (likely(!(trans->flags & BTREE_INSERT_NOMARK)) &&
-	    unlikely(c->gc_pos.phase))
-		trans_for_each_update(trans, i)
-			if (gc_visited(c, gc_pos_btree_node(i->iter->l[0].b)))
-				bch2_mark_update(trans, i, NULL,
-						 mark_flags|
-						 BCH_BUCKET_MARK_GC);
+		BUG_ON(!(i->flags & BTREE_TRIGGER_NORUN) &&
+		       (BTREE_NODE_TYPE_HAS_TRANS_TRIGGERS & (1U << i->bkey_type)) &&
+		       (!i->insert_trigger_run || !i->overwrite_trigger_run));
 
-	trans_for_each_update(trans, i)
-		do_btree_insert_one(trans, i);
-out:
-	BUG_ON(ret &&
-	       (trans->flags & BTREE_INSERT_JOURNAL_RESERVED) &&
-	       trans->journal_res.ref);
+	return 0;
+}
 
-	btree_trans_lock_write(trans, false);
+static noinline int bch2_trans_commit_run_gc_triggers(struct btree_trans *trans)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_insert_entry *i;
+	int ret = 0;
 
-	if (fs_usage) {
-		bch2_fs_usage_scratch_put(c, fs_usage);
-		percpu_up_read(&c->mark_lock);
-	}
+	trans_for_each_update(trans, i) {
+		/*
+		 * XXX: synchronization of cached update triggers with gc
+		 * XXX: synchronization of interior node updates with gc
+		 */
+		BUG_ON(i->cached || i->level);
 
-	bch2_journal_res_put(&c->journal, &trans->journal_res);
-out_clear_replicas:
-	if (trans->fs_usage_deltas) {
-		memset(&trans->fs_usage_deltas->fs_usage, 0,
-		       sizeof(trans->fs_usage_deltas->fs_usage));
-		trans->fs_usage_deltas->used = 0;
+		if (gc_visited(c, gc_pos_btree_node(insert_l(i)->b))) {
+			ret = run_one_mem_trigger(trans, i, i->flags|BTREE_TRIGGER_GC);
+			if (ret)
+				break;
+		}
 	}
 
 	return ret;
 }
 
-static noinline
-int bch2_trans_commit_error(struct btree_trans *trans,
-			    struct btree_insert_entry *i,
-			    int ret)
+static inline int
+bch2_trans_commit_write_locked(struct btree_trans *trans,
+			       struct btree_insert_entry **stopped_at,
+			       unsigned long trace_ip)
 {
 	struct bch_fs *c = trans->c;
-	unsigned flags = trans->flags;
-
-	/*
-	 * BTREE_INSERT_NOUNLOCK means don't unlock _after_ successful btree
-	 * update; if we haven't done anything yet it doesn't apply
+	struct btree_insert_entry *i;
+	struct btree_trans_commit_hook *h;
+	unsigned u64s = 0;
+	bool marking = false;
+	int ret;
+
+	if (race_fault()) {
+		trace_trans_restart_fault_inject(trans->fn, trace_ip);
+		trans->restarted = true;
+		return -EINTR;
+	}
+
+	/*
+	 * Check if the insert will fit in the leaf node with the write lock
+	 * held, otherwise another thread could write the node changing the
+	 * amount of space available:
 	 */
-	flags &= ~BTREE_INSERT_NOUNLOCK;
 
-	switch (ret) {
-	case BTREE_INSERT_BTREE_NODE_FULL:
-		ret = bch2_btree_split_leaf(c, i->iter, flags);
+	prefetch(&trans->c->journal.flags);
+
+	h = trans->hooks;
+	while (h) {
+		ret = h->fn(trans, h);
+		if (ret)
+			return ret;
+		h = h->next;
+	}
+
+	trans_for_each_update(trans, i) {
+		/* Multiple inserts might go to same leaf: */
+		if (!same_leaf_as_prev(trans, i))
+			u64s = 0;
+
+		u64s += i->k->k.u64s;
+		ret = !i->cached
+			? btree_key_can_insert(trans, insert_l(i)->b, u64s)
+			: btree_key_can_insert_cached(trans, i->path, u64s);
+		if (ret) {
+			*stopped_at = i;
+			return ret;
+		}
+
+		if (btree_node_type_needs_gc(i->bkey_type))
+			marking = true;
 
 		/*
-		 * if the split succeeded without dropping locks the insert will
-		 * still be atomic (in the BTREE_INSERT_ATOMIC sense, what the
-		 * caller peeked() and is overwriting won't have changed)
-		 */
-#if 0
-		/*
-		 * XXX:
-		 * split -> btree node merging (of parent node) might still drop
-		 * locks when we're not passing it BTREE_INSERT_NOUNLOCK
+		 * Revalidate before calling mem triggers - XXX, ugly:
+		 *
+		 * - successful btree node splits don't cause transaction
+		 *   restarts and will have invalidated the pointer to the bkey
+		 *   value
+		 * - btree_node_lock_for_insert() -> btree_node_prep_for_write()
+		 *   when it has to resort
+		 * - btree_key_can_insert_cached() when it has to reallocate
 		 *
-		 * we don't want to pass BTREE_INSERT_NOUNLOCK to split as that
-		 * will inhibit merging - but we don't have a reliable way yet
-		 * (do we?) of checking if we dropped locks in this path
+		 *   Ugly because we currently have no way to tell if the
+		 *   pointer's been invalidated, which means it's debatabale
+		 *   whether we should be stashing the old key at all.
 		 */
-		if (!ret)
-			goto retry;
-#endif
+		i->old_v = bch2_btree_path_peek_slot(i->path, &i->old_k).v;
 
-		/*
-		 * don't care if we got ENOSPC because we told split it
-		 * couldn't block:
-		 */
-		if (!ret ||
-		    ret == -EINTR ||
-		    (flags & BTREE_INSERT_NOUNLOCK)) {
-			trace_trans_restart_btree_node_split(trans->ip);
-			ret = -EINTR;
+		if (unlikely(!test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags))) {
+			struct bkey_i *j_k =
+				bch2_journal_keys_peek(c, i->btree_id, i->level, i->k->k.p);
+
+			if (j_k && !bpos_cmp(j_k->k.p, i->k->k.p)) {
+				i->old_k = j_k->k;
+				i->old_v = &j_k->v;
+			}
 		}
-		break;
-	case BTREE_INSERT_ENOSPC:
-		ret = -ENOSPC;
-		break;
-	case BTREE_INSERT_NEED_MARK_REPLICAS:
-		bch2_trans_unlock(trans);
+	}
+
+	/*
+	 * Don't get journal reservation until after we know insert will
+	 * succeed:
+	 */
+	if (likely(!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY))) {
+		ret = bch2_trans_journal_res_get(trans,
+				JOURNAL_RES_GET_NONBLOCK);
+		if (ret)
+			return ret;
+
+		if (unlikely(trans->journal_transaction_names))
+			journal_transaction_name(trans);
+	} else {
+		trans->journal_res.seq = c->journal.replay_journal_seq;
+	}
+
+	if (unlikely(trans->extra_journal_entry_u64s)) {
+		memcpy_u64s_small(journal_res_entry(&c->journal, &trans->journal_res),
+				  trans->extra_journal_entries,
+				  trans->extra_journal_entry_u64s);
+
+		trans->journal_res.offset	+= trans->extra_journal_entry_u64s;
+		trans->journal_res.u64s		-= trans->extra_journal_entry_u64s;
+	}
+
+	/*
+	 * Not allowed to fail after we've gotten our journal reservation - we
+	 * have to use it:
+	 */
 
-		trans_for_each_update(trans, i) {
-			ret = bch2_mark_bkey_replicas(c, bkey_i_to_s_c(i->k));
+	if (!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY)) {
+		if (bch2_journal_seq_verify)
+			trans_for_each_update(trans, i)
+				i->k->k.version.lo = trans->journal_res.seq;
+		else if (bch2_inject_invalid_keys)
+			trans_for_each_update(trans, i)
+				i->k->k.version = MAX_VERSION;
+	}
+
+	if (trans->fs_usage_deltas &&
+	    bch2_trans_fs_usage_apply(trans, trans->fs_usage_deltas))
+		return BTREE_INSERT_NEED_MARK_REPLICAS;
+
+	trans_for_each_update(trans, i)
+		if (BTREE_NODE_TYPE_HAS_MEM_TRIGGERS & (1U << i->bkey_type)) {
+			ret = run_one_mem_trigger(trans, i, i->flags);
 			if (ret)
 				return ret;
 		}
 
+	if (unlikely(c->gc_pos.phase)) {
+		ret = bch2_trans_commit_run_gc_triggers(trans);
+		if  (ret)
+			return ret;
+	}
+
+	trans_for_each_update(trans, i)
+		do_btree_insert_one(trans, i);
+
+	return ret;
+}
+
+static inline void path_upgrade_readers(struct btree_trans *trans, struct btree_path *path)
+{
+	unsigned l;
+
+	for (l = 0; l < BTREE_MAX_DEPTH; l++)
+		if (btree_node_read_locked(path, l))
+			BUG_ON(!bch2_btree_node_upgrade(trans, path, l));
+}
+
+static inline void upgrade_readers(struct btree_trans *trans, struct btree_path *path)
+{
+	struct btree *b = path_l(path)->b;
+
+	do {
+		if (path->nodes_locked &&
+		    path->nodes_locked != path->nodes_intent_locked)
+			path_upgrade_readers(trans, path);
+	} while ((path = prev_btree_path(trans, path)) &&
+		 path_l(path)->b == b);
+}
+
+/*
+ * Check for nodes that we have both read and intent locks on, and upgrade the
+ * readers to intent:
+ */
+static inline void normalize_read_intent_locks(struct btree_trans *trans)
+{
+	struct btree_path *path;
+	unsigned i, nr_read = 0, nr_intent = 0;
+
+	trans_for_each_path_inorder(trans, path, i) {
+		struct btree_path *next = i + 1 < trans->nr_sorted
+			? trans->paths + trans->sorted[i + 1]
+			: NULL;
+
+		if (path->nodes_locked) {
+			if (path->nodes_intent_locked)
+				nr_intent++;
+			else
+				nr_read++;
+		}
+
+		if (!next || path_l(path)->b != path_l(next)->b) {
+			if (nr_read && nr_intent)
+				upgrade_readers(trans, path);
+
+			nr_read = nr_intent = 0;
+		}
+	}
+
+	bch2_trans_verify_locks(trans);
+}
+
+static inline bool have_conflicting_read_lock(struct btree_trans *trans, struct btree_path *pos)
+{
+	struct btree_path *path;
+	unsigned i;
+
+	trans_for_each_path_inorder(trans, path, i) {
+		//if (path == pos)
+		//	break;
+
+		if (path->nodes_locked != path->nodes_intent_locked &&
+		    !bch2_btree_path_upgrade(trans, path, path->level + 1))
+			return true;
+	}
+
+	return false;
+}
+
+static inline int trans_lock_write(struct btree_trans *trans)
+{
+	struct btree_insert_entry *i;
+
+	trans_for_each_update(trans, i) {
+		if (same_leaf_as_prev(trans, i))
+			continue;
+
+		if (!six_trylock_write(&insert_l(i)->b->c.lock)) {
+			if (have_conflicting_read_lock(trans, i->path))
+				goto fail;
+
+			btree_node_lock_type(trans, i->path,
+					     insert_l(i)->b,
+					     i->path->pos, i->level,
+					     SIX_LOCK_write, NULL, NULL);
+		}
+
+		bch2_btree_node_prep_for_write(trans, i->path, insert_l(i)->b);
+	}
+
+	return 0;
+fail:
+	while (--i >= trans->updates) {
+		if (same_leaf_as_prev(trans, i))
+			continue;
+
+		bch2_btree_node_unlock_write_inlined(trans, i->path, insert_l(i)->b);
+	}
+
+	trace_trans_restart_would_deadlock_write(trans->fn);
+	return btree_trans_restart(trans);
+}
+
+static noinline void bch2_drop_overwrites_from_journal(struct btree_trans *trans)
+{
+	struct btree_insert_entry *i;
+
+	trans_for_each_update(trans, i)
+		bch2_journal_key_overwritten(trans->c, i->btree_id, i->level, i->k->k.p);
+}
+
+/*
+ * Get journal reservation, take write locks, and attempt to do btree update(s):
+ */
+static inline int do_bch2_trans_commit(struct btree_trans *trans,
+				       struct btree_insert_entry **stopped_at,
+				       unsigned long trace_ip)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_insert_entry *i;
+	int ret, u64s_delta = 0;
+
+	trans_for_each_update(trans, i) {
+		const char *invalid = bch2_bkey_invalid(c,
+				bkey_i_to_s_c(i->k), i->bkey_type);
+		if (invalid) {
+			struct printbuf buf = PRINTBUF;
+
+			bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(i->k));
+			bch2_fs_fatal_error(c, "invalid bkey %s on insert from %s -> %ps: %s\n",
+					    buf.buf, trans->fn, (void *) i->ip_allocated, invalid);
+			printbuf_exit(&buf);
+			return -EINVAL;
+		}
+		btree_insert_entry_checks(trans, i);
+	}
+
+	trans_for_each_update(trans, i) {
+		if (i->cached)
+			continue;
+
+		u64s_delta += !bkey_deleted(&i->k->k) ? i->k->k.u64s : 0;
+		u64s_delta -= i->old_btree_u64s;
+
+		if (!same_leaf_as_next(trans, i)) {
+			if (u64s_delta <= 0) {
+				ret = bch2_foreground_maybe_merge(trans, i->path,
+							i->level, trans->flags);
+				if (unlikely(ret))
+					return ret;
+			}
+
+			u64s_delta = 0;
+		}
+	}
+
+	ret = bch2_journal_preres_get(&c->journal,
+			&trans->journal_preres, trans->journal_preres_u64s,
+			JOURNAL_RES_GET_NONBLOCK|
+			((trans->flags & BTREE_INSERT_JOURNAL_RESERVED)
+			 ? JOURNAL_RES_GET_RESERVED : 0));
+	if (unlikely(ret == -EAGAIN))
+		ret = bch2_trans_journal_preres_get_cold(trans,
+						trans->journal_preres_u64s, trace_ip);
+	if (unlikely(ret))
+		return ret;
+
+	normalize_read_intent_locks(trans);
+
+	ret = trans_lock_write(trans);
+	if (unlikely(ret))
+		return ret;
+
+	ret = bch2_trans_commit_write_locked(trans, stopped_at, trace_ip);
+
+	if (!ret && unlikely(!test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags)))
+		bch2_drop_overwrites_from_journal(trans);
+
+	trans_for_each_update(trans, i)
+		if (!same_leaf_as_prev(trans, i))
+			bch2_btree_node_unlock_write_inlined(trans, i->path,
+							insert_l(i)->b);
+
+	if (!ret && trans->journal_pin)
+		bch2_journal_pin_add(&c->journal, trans->journal_res.seq,
+				     trans->journal_pin, NULL);
+
+	/*
+	 * Drop journal reservation after dropping write locks, since dropping
+	 * the journal reservation may kick off a journal write:
+	 */
+	bch2_journal_res_put(&c->journal, &trans->journal_res);
+
+	if (unlikely(ret))
+		return ret;
+
+	bch2_trans_downgrade(trans);
+
+	return 0;
+}
+
+static int journal_reclaim_wait_done(struct bch_fs *c)
+{
+	int ret = bch2_journal_error(&c->journal) ?:
+		!bch2_btree_key_cache_must_wait(c);
+
+	if (!ret)
+		journal_reclaim_kick(&c->journal);
+	return ret;
+}
+
+static noinline
+int bch2_trans_commit_error(struct btree_trans *trans,
+			    struct btree_insert_entry *i,
+			    int ret, unsigned long trace_ip)
+{
+	struct bch_fs *c = trans->c;
+
+	switch (ret) {
+	case BTREE_INSERT_BTREE_NODE_FULL:
+		ret = bch2_btree_split_leaf(trans, i->path, trans->flags);
+		if (!ret)
+			return 0;
+
+		if (ret == -EINTR)
+			trace_trans_restart_btree_node_split(trans->fn, trace_ip,
+						i->btree_id, &i->path->pos);
+		break;
+	case BTREE_INSERT_NEED_MARK_REPLICAS:
+		bch2_trans_unlock(trans);
+
+		ret = bch2_replicas_delta_list_mark(c, trans->fs_usage_deltas);
+		if (ret)
+			break;
+
 		if (bch2_trans_relock(trans))
 			return 0;
 
-		trace_trans_restart_mark_replicas(trans->ip);
+		trace_trans_restart_mark_replicas(trans->fn, trace_ip);
 		ret = -EINTR;
 		break;
 	case BTREE_INSERT_NEED_JOURNAL_RES:
 		bch2_trans_unlock(trans);
 
+		if ((trans->flags & BTREE_INSERT_JOURNAL_RECLAIM) &&
+		    !(trans->flags & BTREE_INSERT_JOURNAL_RESERVED)) {
+			trans->restarted = true;
+			ret = -EAGAIN;
+			break;
+		}
+
 		ret = bch2_trans_journal_res_get(trans, JOURNAL_RES_GET_CHECK);
 		if (ret)
-			return ret;
+			break;
 
 		if (bch2_trans_relock(trans))
 			return 0;
 
-		trace_trans_restart_journal_res_get(trans->ip);
+		trace_trans_restart_journal_res_get(trans->fn, trace_ip);
 		ret = -EINTR;
 		break;
-	default:
-		BUG_ON(ret >= 0);
-		break;
-	}
+	case BTREE_INSERT_NEED_JOURNAL_RECLAIM:
+		bch2_trans_unlock(trans);
 
-	if (ret == -EINTR) {
-		int ret2 = bch2_btree_iter_traverse_all(trans);
+		trace_trans_blocked_journal_reclaim(trans->fn, trace_ip);
 
-		if (ret2) {
-			trace_trans_restart_traverse(trans->ip);
-			return ret2;
-		}
+		wait_event_freezable(c->journal.reclaim_wait,
+				     (ret = journal_reclaim_wait_done(c)));
+		if (ret < 0)
+			break;
 
-		/*
-		 * BTREE_ITER_ATOMIC means we have to return -EINTR if we
-		 * dropped locks:
-		 */
-		if (!(flags & BTREE_INSERT_ATOMIC))
+		if (bch2_trans_relock(trans))
 			return 0;
 
-		trace_trans_restart_atomic(trans->ip);
+		trace_trans_restart_journal_reclaim(trans->fn, trace_ip);
+		ret = -EINTR;
+		break;
+	default:
+		BUG_ON(ret >= 0);
+		break;
 	}
 
+	BUG_ON((ret == EINTR || ret == -EAGAIN) && !trans->restarted);
+	BUG_ON(ret == -ENOSPC &&
+	       !(trans->flags & BTREE_INSERT_NOWAIT) &&
+	       (trans->flags & BTREE_INSERT_NOFAIL));
+
 	return ret;
 }
 
-/**
- * __bch_btree_insert_at - insert keys at given iterator positions
- *
- * This is main entry point for btree updates.
- *
- * Return values:
- * -EINTR: locking changed, this function should be called again. Only returned
- *  if passed BTREE_INSERT_ATOMIC.
- * -EROFS: filesystem read only
- * -EIO: journal or btree node IO error
- */
-static int __bch2_trans_commit(struct btree_trans *trans,
-			       struct btree_insert_entry **stopped_at)
+static noinline int
+bch2_trans_commit_get_rw_cold(struct btree_trans *trans)
 {
 	struct bch_fs *c = trans->c;
-	struct btree_insert_entry *i;
-	unsigned iter;
 	int ret;
 
-	trans_for_each_update(trans, i) {
-		if (!bch2_btree_iter_upgrade(i->iter, 1)) {
-			trace_trans_restart_upgrade(trans->ip);
-			ret = -EINTR;
-			goto err;
-		}
+	if (likely(!(trans->flags & BTREE_INSERT_LAZY_RW)) ||
+	    test_bit(BCH_FS_STARTED, &c->flags))
+		return -EROFS;
 
-		ret = btree_iter_err(i->iter);
-		if (ret)
-			goto err;
-	}
+	bch2_trans_unlock(trans);
 
-	ret = do_btree_insert_at(trans, stopped_at);
-	if (unlikely(ret))
-		goto err;
+	ret = bch2_fs_read_write_early(c);
+	if (ret)
+		return ret;
 
-	if (trans->flags & BTREE_INSERT_NOUNLOCK)
-		trans->nounlock = true;
+	if (!bch2_trans_relock(trans))
+		return -EINTR;
 
-	trans_for_each_update_sorted(trans, i, iter)
-		if (!same_leaf_as_prev(trans, iter))
-			bch2_foreground_maybe_merge(c, i->iter,
-						    0, trans->flags);
+	percpu_ref_get(&c->writes);
+	return 0;
+}
 
-	trans->nounlock = false;
+/*
+ * This is for updates done in the early part of fsck - btree_gc - before we've
+ * gone RW. we only add the new key to the list of keys for journal replay to
+ * do.
+ */
+static noinline int
+do_bch2_trans_commit_to_journal_replay(struct btree_trans *trans)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_insert_entry *i;
+	int ret = 0;
 
-	trans_for_each_update(trans, i)
-		bch2_btree_iter_downgrade(i->iter);
-err:
-	/* make sure we didn't drop or screw up locks: */
-	bch2_btree_trans_verify_locks(trans);
+	trans_for_each_update(trans, i) {
+		ret = bch2_journal_key_insert(c, i->btree_id, i->level, i->k);
+		if (ret)
+			break;
+	}
 
 	return ret;
 }
 
-int bch2_trans_commit(struct btree_trans *trans,
-		      struct disk_reservation *disk_res,
-		      u64 *journal_seq,
-		      unsigned flags)
+int __bch2_trans_commit(struct btree_trans *trans)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_insert_entry *i = NULL;
-	struct btree_iter *iter;
-	unsigned orig_nr_updates	= trans->nr_updates;
-	unsigned orig_mem_top		= trans->mem_top;
+	unsigned u64s;
 	int ret = 0;
 
-	if (!trans->nr_updates)
-		goto out_noupdates;
-
-	/* for the sake of sanity: */
-	BUG_ON(trans->nr_updates > 1 && !(flags & BTREE_INSERT_ATOMIC));
+	if (!trans->nr_updates &&
+	    !trans->extra_journal_entry_u64s)
+		goto out_reset;
 
-	if (flags & BTREE_INSERT_GC_LOCK_HELD)
+	if (trans->flags & BTREE_INSERT_GC_LOCK_HELD)
 		lockdep_assert_held(&c->gc_lock);
 
-	if (!trans->commit_start)
-		trans->commit_start = local_clock();
+	ret = bch2_trans_commit_run_triggers(trans);
+	if (ret)
+		goto out_reset;
+
+	if (unlikely(!test_bit(BCH_FS_MAY_GO_RW, &c->flags))) {
+		ret = do_bch2_trans_commit_to_journal_replay(trans);
+		goto out_reset;
+	}
+
+	if (!(trans->flags & BTREE_INSERT_NOCHECK_RW) &&
+	    unlikely(!percpu_ref_tryget(&c->writes))) {
+		ret = bch2_trans_commit_get_rw_cold(trans);
+		if (ret)
+			goto out_reset;
+	}
 
-	memset(&trans->journal_res, 0, sizeof(trans->journal_res));
 	memset(&trans->journal_preres, 0, sizeof(trans->journal_preres));
-	trans->disk_res		= disk_res;
-	trans->journal_seq	= journal_seq;
-	trans->flags		= flags;
 
-	if (unlikely(!(trans->flags & BTREE_INSERT_NOCHECK_RW) &&
-		     !percpu_ref_tryget(&c->writes))) {
-		if (likely(!(trans->flags & BTREE_INSERT_LAZY_RW)))
-			return -EROFS;
+	trans->journal_u64s		= trans->extra_journal_entry_u64s;
+	trans->journal_preres_u64s	= 0;
 
-		bch2_trans_unlock(trans);
+	trans->journal_transaction_names = READ_ONCE(c->opts.journal_transaction_names);
 
-		ret = bch2_fs_read_write_early(c);
-		if (ret)
-			return ret;
+	if (trans->journal_transaction_names)
+		trans->journal_u64s += JSET_ENTRY_LOG_U64s;
 
-		percpu_ref_get(&c->writes);
+	trans_for_each_update(trans, i) {
+		BUG_ON(!i->path->should_be_locked);
 
-		if (!bch2_trans_relock(trans)) {
-			ret = -EINTR;
-			goto err;
+		if (unlikely(!bch2_btree_path_upgrade(trans, i->path, i->level + 1))) {
+			trace_trans_restart_upgrade(trans->fn, _RET_IP_,
+						    i->btree_id, &i->path->pos);
+			ret = btree_trans_restart(trans);
+			goto out;
 		}
+
+		BUG_ON(!btree_node_intent_locked(i->path, i->level));
+
+		u64s = jset_u64s(i->k->k.u64s);
+		if (i->cached &&
+		    likely(!(trans->flags & BTREE_INSERT_JOURNAL_REPLAY)))
+			trans->journal_preres_u64s += u64s;
+		trans->journal_u64s += u64s;
+	}
+
+	if (trans->extra_journal_res) {
+		ret = bch2_disk_reservation_add(c, trans->disk_res,
+				trans->extra_journal_res,
+				(trans->flags & BTREE_INSERT_NOFAIL)
+				? BCH_DISK_RESERVATION_NOFAIL : 0);
+		if (ret)
+			goto err;
 	}
 retry:
-	ret = bch2_trans_journal_preres_get(trans);
-	if (ret)
-		goto err;
+	BUG_ON(trans->restarted);
+	memset(&trans->journal_res, 0, sizeof(trans->journal_res));
+
+	ret = do_bch2_trans_commit(trans, &i, _RET_IP_);
+
+	/* make sure we didn't drop or screw up locks: */
+	bch2_trans_verify_locks(trans);
 
-	ret = __bch2_trans_commit(trans, &i);
 	if (ret)
 		goto err;
 out:
 	bch2_journal_preres_put(&c->journal, &trans->journal_preres);
 
-	if (unlikely(!(trans->flags & BTREE_INSERT_NOCHECK_RW)))
+	if (likely(!(trans->flags & BTREE_INSERT_NOCHECK_RW)))
 		percpu_ref_put(&c->writes);
-out_noupdates:
-	if (!ret && trans->commit_start) {
-		bch2_time_stats_update(&c->times[BCH_TIME_btree_update],
-				       trans->commit_start);
-		trans->commit_start = 0;
+out_reset:
+	trans_for_each_update(trans, i)
+		bch2_path_put(trans, i->path, true);
+
+	trans->extra_journal_res	= 0;
+	trans->nr_updates		= 0;
+	trans->hooks			= NULL;
+	trans->extra_journal_entries	= NULL;
+	trans->extra_journal_entry_u64s	= 0;
+
+	if (trans->fs_usage_deltas) {
+		trans->fs_usage_deltas->used = 0;
+		memset(&trans->fs_usage_deltas->memset_start, 0,
+		       (void *) &trans->fs_usage_deltas->memset_end -
+		       (void *) &trans->fs_usage_deltas->memset_start);
 	}
 
-	BUG_ON(!(trans->flags & BTREE_INSERT_ATOMIC) && ret == -EINTR);
+	return ret;
+err:
+	ret = bch2_trans_commit_error(trans, i, ret, _RET_IP_);
+	if (ret)
+		goto out;
+
+	goto retry;
+}
+
+static int check_pos_snapshot_overwritten(struct btree_trans *trans,
+					  enum btree_id id,
+					  struct bpos pos)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_iter iter;
+	struct bkey_s_c k;
+	int ret;
+
+	if (!btree_type_has_snapshots(id))
+		return 0;
+
+	if (!snapshot_t(c, pos.snapshot)->children[0])
+		return 0;
+
+	bch2_trans_iter_init(trans, &iter, id, pos,
+			     BTREE_ITER_NOT_EXTENTS|
+			     BTREE_ITER_ALL_SNAPSHOTS);
+	while (1) {
+		k = bch2_btree_iter_prev(&iter);
+		ret = bkey_err(k);
+		if (ret)
+			break;
 
-	trans_for_each_iter(trans, iter)
-		iter->flags &= ~BTREE_ITER_KEEP_UNTIL_COMMIT;
+		if (!k.k)
+			break;
+
+		if (bkey_cmp(pos, k.k->p))
+			break;
 
-	if (!ret) {
-		bch2_trans_unlink_iters(trans);
-		trans->iters_touched = 0;
+		if (bch2_snapshot_is_ancestor(c, k.k->p.snapshot, pos.snapshot)) {
+			ret = 1;
+			break;
+		}
 	}
-	trans->nr_updates	= 0;
-	trans->mem_top		= 0;
+	bch2_trans_iter_exit(trans, &iter);
 
 	return ret;
+}
+
+int bch2_trans_update_extent(struct btree_trans *trans,
+			     struct btree_iter *orig_iter,
+			     struct bkey_i *insert,
+			     enum btree_update_flags flags)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_iter iter, update_iter;
+	struct bpos start = bkey_start_pos(&insert->k);
+	struct bkey_i *update;
+	struct bkey_s_c k;
+	enum btree_id btree_id = orig_iter->btree_id;
+	int ret = 0, compressed_sectors;
+
+	bch2_trans_iter_init(trans, &iter, btree_id, start,
+			     BTREE_ITER_INTENT|
+			     BTREE_ITER_WITH_UPDATES|
+			     BTREE_ITER_NOT_EXTENTS);
+	k = bch2_btree_iter_peek_upto(&iter, POS(insert->k.p.inode, U64_MAX));
+	if ((ret = bkey_err(k)))
+		goto err;
+	if (!k.k)
+		goto out;
+
+	if (bch2_bkey_maybe_mergable(k.k, &insert->k)) {
+		/*
+		 * We can't merge extents if they belong to interior snapshot
+		 * tree nodes, and there's a snapshot in which one extent is
+		 * visible and the other is not - i.e. if visibility is
+		 * different.
+		 *
+		 * Instead of checking if visibilitiy of the two extents is
+		 * different, for now we just check if either has been
+		 * overwritten:
+		 */
+		ret = check_pos_snapshot_overwritten(trans, btree_id, insert->k.p);
+		if (ret < 0)
+			goto err;
+		if (ret)
+			goto nomerge1;
+
+		ret = check_pos_snapshot_overwritten(trans, btree_id, k.k->p);
+		if (ret < 0)
+			goto err;
+		if (ret)
+			goto nomerge1;
+
+		update = bch2_trans_kmalloc(trans, bkey_bytes(k.k));
+		if ((ret = PTR_ERR_OR_ZERO(update)))
+			goto err;
+
+		bkey_reassemble(update, k);
+
+		if (bch2_bkey_merge(c, bkey_i_to_s(update), bkey_i_to_s_c(insert))) {
+			ret = bch2_btree_delete_at(trans, &iter, flags);
+			if (ret)
+				goto err;
+
+			insert = update;
+			goto next;
+		}
+	}
+nomerge1:
+	ret = 0;
+	if (!bkey_cmp(k.k->p, start))
+		goto next;
+
+	while (bkey_cmp(insert->k.p, bkey_start_pos(k.k)) > 0) {
+		bool front_split = bkey_cmp(bkey_start_pos(k.k), start) < 0;
+		bool back_split  = bkey_cmp(k.k->p, insert->k.p) > 0;
+
+		/*
+		 * If we're going to be splitting a compressed extent, note it
+		 * so that __bch2_trans_commit() can increase our disk
+		 * reservation:
+		 */
+		if (((front_split && back_split) ||
+		     ((front_split || back_split) && k.k->p.snapshot != insert->k.p.snapshot)) &&
+		    (compressed_sectors = bch2_bkey_sectors_compressed(k)))
+			trans->extra_journal_res += compressed_sectors;
+
+		if (front_split) {
+			update = bch2_trans_kmalloc(trans, bkey_bytes(k.k));
+			if ((ret = PTR_ERR_OR_ZERO(update)))
+				goto err;
+
+			bkey_reassemble(update, k);
+
+			bch2_cut_back(start, update);
+
+			bch2_trans_iter_init(trans, &update_iter, btree_id, update->k.p,
+					     BTREE_ITER_NOT_EXTENTS|
+					     BTREE_ITER_ALL_SNAPSHOTS|
+					     BTREE_ITER_INTENT);
+			ret   = bch2_btree_iter_traverse(&update_iter) ?:
+				bch2_trans_update(trans, &update_iter, update,
+						  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
+						  flags);
+			bch2_trans_iter_exit(trans, &update_iter);
+
+			if (ret)
+				goto err;
+		}
+
+		if (k.k->p.snapshot != insert->k.p.snapshot &&
+		    (front_split || back_split)) {
+			update = bch2_trans_kmalloc(trans, bkey_bytes(k.k));
+			if ((ret = PTR_ERR_OR_ZERO(update)))
+				goto err;
+
+			bkey_reassemble(update, k);
+
+			bch2_cut_front(start, update);
+			bch2_cut_back(insert->k.p, update);
+
+			bch2_trans_iter_init(trans, &update_iter, btree_id, update->k.p,
+					     BTREE_ITER_NOT_EXTENTS|
+					     BTREE_ITER_ALL_SNAPSHOTS|
+					     BTREE_ITER_INTENT);
+			ret   = bch2_btree_iter_traverse(&update_iter) ?:
+				bch2_trans_update(trans, &update_iter, update,
+						  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
+						  flags);
+			bch2_trans_iter_exit(trans, &update_iter);
+			if (ret)
+				goto err;
+		}
+
+		if (bkey_cmp(k.k->p, insert->k.p) <= 0) {
+			update = bch2_trans_kmalloc(trans, sizeof(*update));
+			if ((ret = PTR_ERR_OR_ZERO(update)))
+				goto err;
+
+			bkey_init(&update->k);
+			update->k.p = k.k->p;
+
+			if (insert->k.p.snapshot != k.k->p.snapshot) {
+				update->k.p.snapshot = insert->k.p.snapshot;
+				update->k.type = KEY_TYPE_whiteout;
+			}
+
+			bch2_trans_iter_init(trans, &update_iter, btree_id, update->k.p,
+					     BTREE_ITER_NOT_EXTENTS|
+					     BTREE_ITER_INTENT);
+			ret   = bch2_btree_iter_traverse(&update_iter) ?:
+				bch2_trans_update(trans, &update_iter, update,
+						  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
+						  flags);
+			bch2_trans_iter_exit(trans, &update_iter);
+
+			if (ret)
+				goto err;
+		}
+
+		if (back_split) {
+			update = bch2_trans_kmalloc(trans, bkey_bytes(k.k));
+			if ((ret = PTR_ERR_OR_ZERO(update)))
+				goto err;
+
+			bkey_reassemble(update, k);
+			bch2_cut_front(insert->k.p, update);
+
+			ret = bch2_trans_update_by_path(trans, iter.path, update,
+						  BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE|
+						  flags);
+			if (ret)
+				goto err;
+			goto out;
+		}
+next:
+		bch2_btree_iter_advance(&iter);
+		k = bch2_btree_iter_peek_upto(&iter, POS(insert->k.p.inode, U64_MAX));
+		if ((ret = bkey_err(k)))
+			goto err;
+		if (!k.k)
+			goto out;
+	}
+
+	if (bch2_bkey_maybe_mergable(&insert->k, k.k)) {
+		ret = check_pos_snapshot_overwritten(trans, btree_id, insert->k.p);
+		if (ret < 0)
+			goto err;
+		if (ret)
+			goto nomerge2;
+
+		ret = check_pos_snapshot_overwritten(trans, btree_id, k.k->p);
+		if (ret < 0)
+			goto err;
+		if (ret)
+			goto nomerge2;
+
+		bch2_bkey_merge(c, bkey_i_to_s(insert), k);
+	}
+nomerge2:
+	ret = 0;
+out:
+	if (!bkey_deleted(&insert->k)) {
+		/*
+		 * Rewinding iterators is expensive: get a new one and the one
+		 * that points to the start of insert will be cloned from:
+		 */
+		bch2_trans_iter_exit(trans, &iter);
+		bch2_trans_iter_init(trans, &iter, btree_id, insert->k.p,
+				     BTREE_ITER_NOT_EXTENTS|
+				     BTREE_ITER_INTENT);
+		ret   = bch2_btree_iter_traverse(&iter) ?:
+			bch2_trans_update(trans, &iter, insert, flags);
+	}
 err:
-	ret = bch2_trans_commit_error(trans, i, ret);
+	bch2_trans_iter_exit(trans, &iter);
+
+	return ret;
+}
 
-	/* free updates and memory used by triggers, they'll be reexecuted: */
-	trans->nr_updates	= orig_nr_updates;
-	trans->mem_top		= orig_mem_top;
+/*
+ * When deleting, check if we need to emit a whiteout (because we're overwriting
+ * something in an ancestor snapshot)
+ */
+static int need_whiteout_for_snapshot(struct btree_trans *trans,
+				      enum btree_id btree_id, struct bpos pos)
+{
+	struct btree_iter iter;
+	struct bkey_s_c k;
+	u32 snapshot = pos.snapshot;
+	int ret;
 
-	/* can't loop if it was passed in and we changed it: */
-	if (unlikely(trans->flags & BTREE_INSERT_NO_CLEAR_REPLICAS) && !ret)
-		ret = -EINTR;
+	if (!bch2_snapshot_parent(trans->c, pos.snapshot))
+		return 0;
 
-	if (!ret)
-		goto retry;
+	pos.snapshot++;
+
+	for_each_btree_key_norestart(trans, iter, btree_id, pos,
+			   BTREE_ITER_ALL_SNAPSHOTS|
+			   BTREE_ITER_NOPRESERVE, k, ret) {
+		if (bkey_cmp(k.k->p, pos))
+			break;
+
+		if (bch2_snapshot_is_ancestor(trans->c, snapshot,
+					      k.k->p.snapshot)) {
+			ret = !bkey_whiteout(k.k);
+			break;
+		}
+	}
+	bch2_trans_iter_exit(trans, &iter);
+
+	return ret;
+}
+
+static int __must_check
+bch2_trans_update_by_path(struct btree_trans *trans, struct btree_path *path,
+			  struct bkey_i *k, enum btree_update_flags flags)
+{
+	struct bch_fs *c = trans->c;
+	struct btree_insert_entry *i, n;
+
+	BUG_ON(!path->should_be_locked);
+
+	BUG_ON(trans->nr_updates >= BTREE_ITER_MAX);
+	BUG_ON(bpos_cmp(k->k.p, path->pos));
+
+	n = (struct btree_insert_entry) {
+		.flags		= flags,
+		.bkey_type	= __btree_node_type(path->level, path->btree_id),
+		.btree_id	= path->btree_id,
+		.level		= path->level,
+		.cached		= path->cached,
+		.path		= path,
+		.k		= k,
+		.ip_allocated	= _RET_IP_,
+	};
+
+#ifdef CONFIG_BCACHEFS_DEBUG
+	trans_for_each_update(trans, i)
+		BUG_ON(i != trans->updates &&
+		       btree_insert_entry_cmp(i - 1, i) >= 0);
+#endif
+
+	/*
+	 * Pending updates are kept sorted: first, find position of new update,
+	 * then delete/trim any updates the new update overwrites:
+	 */
+	trans_for_each_update(trans, i)
+		if (btree_insert_entry_cmp(&n, i) <= 0)
+			break;
+
+	if (i < trans->updates + trans->nr_updates &&
+	    !btree_insert_entry_cmp(&n, i)) {
+		BUG_ON(i->insert_trigger_run || i->overwrite_trigger_run);
+
+		bch2_path_put(trans, i->path, true);
+		i->flags	= n.flags;
+		i->cached	= n.cached;
+		i->k		= n.k;
+		i->path		= n.path;
+		i->ip_allocated	= n.ip_allocated;
+	} else {
+		array_insert_item(trans->updates, trans->nr_updates,
+				  i - trans->updates, n);
+
+		i->old_v = bch2_btree_path_peek_slot(path, &i->old_k).v;
+		i->old_btree_u64s = !bkey_deleted(&i->old_k) ? i->old_k.u64s : 0;
+
+		if (unlikely(!test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags))) {
+			struct bkey_i *j_k =
+				bch2_journal_keys_peek(c, n.btree_id, n.level, k->k.p);
+
+			if (j_k && !bpos_cmp(j_k->k.p, i->k->k.p)) {
+				i->old_k = j_k->k;
+				i->old_v = &j_k->v;
+			}
+		}
+	}
+
+	__btree_path_get(n.path, true);
+	return 0;
+}
+
+int __must_check bch2_trans_update(struct btree_trans *trans, struct btree_iter *iter,
+				   struct bkey_i *k, enum btree_update_flags flags)
+{
+	struct btree_path *path = iter->update_path ?: iter->path;
+	struct bkey_cached *ck;
+	int ret;
+
+	if (iter->flags & BTREE_ITER_IS_EXTENTS)
+		return bch2_trans_update_extent(trans, iter, k, flags);
+
+	if (bkey_deleted(&k->k) &&
+	    !(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
+	    (iter->flags & BTREE_ITER_FILTER_SNAPSHOTS)) {
+		ret = need_whiteout_for_snapshot(trans, iter->btree_id, k->k.p);
+		if (unlikely(ret < 0))
+			return ret;
+
+		if (ret)
+			k->k.type = KEY_TYPE_whiteout;
+	}
+
+	if (!(flags & BTREE_UPDATE_KEY_CACHE_RECLAIM) &&
+	    !path->cached &&
+	    !path->level &&
+	    btree_id_cached(trans->c, path->btree_id)) {
+		if (!iter->key_cache_path ||
+		    !iter->key_cache_path->should_be_locked ||
+		    bpos_cmp(iter->key_cache_path->pos, k->k.p)) {
+			if (!iter->key_cache_path)
+				iter->key_cache_path =
+					bch2_path_get(trans, path->btree_id, path->pos, 1, 0,
+						      BTREE_ITER_INTENT|
+						      BTREE_ITER_CACHED, _THIS_IP_);
+
+			iter->key_cache_path =
+				bch2_btree_path_set_pos(trans, iter->key_cache_path, path->pos,
+							iter->flags & BTREE_ITER_INTENT,
+							_THIS_IP_);
+
+			ret = bch2_btree_path_traverse(trans, iter->key_cache_path,
+						       BTREE_ITER_CACHED);
+			if (unlikely(ret))
+				return ret;
+
+			ck = (void *) iter->key_cache_path->l[0].b;
+
+			if (test_bit(BKEY_CACHED_DIRTY, &ck->flags)) {
+				trace_trans_restart_key_cache_raced(trans->fn, _RET_IP_);
+				btree_trans_restart(trans);
+				return -EINTR;
+			}
 
-	goto out;
+			iter->key_cache_path->should_be_locked = true;
+		}
+
+		path = iter->key_cache_path;
+	}
+
+	return bch2_trans_update_by_path(trans, path, k, flags);
+}
+
+void bch2_trans_commit_hook(struct btree_trans *trans,
+			    struct btree_trans_commit_hook *h)
+{
+	h->next = trans->hooks;
+	trans->hooks = h;
+}
+
+int __bch2_btree_insert(struct btree_trans *trans,
+			enum btree_id id, struct bkey_i *k)
+{
+	struct btree_iter iter;
+	int ret;
+
+	bch2_trans_iter_init(trans, &iter, id, bkey_start_pos(&k->k),
+			     BTREE_ITER_INTENT);
+	ret   = bch2_btree_iter_traverse(&iter) ?:
+		bch2_trans_update(trans, &iter, k, 0);
+	bch2_trans_iter_exit(trans, &iter);
+	return ret;
 }
 
 /**
@@ -851,46 +1630,53 @@ err:
  * @hook:		insert callback
  */
 int bch2_btree_insert(struct bch_fs *c, enum btree_id id,
-		     struct bkey_i *k,
-		     struct disk_reservation *disk_res,
-		     u64 *journal_seq, int flags)
+		      struct bkey_i *k,
+		      struct disk_reservation *disk_res,
+		      u64 *journal_seq, int flags)
 {
-	struct btree_trans trans;
-	struct btree_iter *iter;
-	int ret;
-
-	bch2_trans_init(&trans, c, 0, 0);
-retry:
-	bch2_trans_begin(&trans);
-
-	iter = bch2_trans_get_iter(&trans, id, bkey_start_pos(&k->k),
-				   BTREE_ITER_INTENT);
+	return bch2_trans_do(c, disk_res, journal_seq, flags,
+			     __bch2_btree_insert(&trans, id, k));
+}
 
-	bch2_trans_update(&trans, iter, k);
+int bch2_btree_delete_at(struct btree_trans *trans,
+			 struct btree_iter *iter, unsigned update_flags)
+{
+	struct bkey_i *k;
 
-	ret = bch2_trans_commit(&trans, disk_res, journal_seq, flags);
-	if (ret == -EINTR)
-		goto retry;
-	bch2_trans_exit(&trans);
+	k = bch2_trans_kmalloc(trans, sizeof(*k));
+	if (IS_ERR(k))
+		return PTR_ERR(k);
 
-	return ret;
+	bkey_init(&k->k);
+	k->k.p = iter->pos;
+	return bch2_trans_update(trans, iter, k, update_flags);
 }
 
-int bch2_btree_delete_at_range(struct btree_trans *trans,
-			       struct btree_iter *iter,
-			       struct bpos end,
-			       u64 *journal_seq)
+int bch2_btree_delete_range_trans(struct btree_trans *trans, enum btree_id id,
+				  struct bpos start, struct bpos end,
+				  unsigned update_flags,
+				  u64 *journal_seq)
 {
+	struct btree_iter iter;
 	struct bkey_s_c k;
 	int ret = 0;
+
+	bch2_trans_iter_init(trans, &iter, id, start, BTREE_ITER_INTENT);
 retry:
-	while ((k = bch2_btree_iter_peek(iter)).k &&
+	while ((bch2_trans_begin(trans),
+	       (k = bch2_btree_iter_peek(&iter)).k) &&
 	       !(ret = bkey_err(k)) &&
-	       bkey_cmp(iter->pos, end) < 0) {
+	       bkey_cmp(iter.pos, end) < 0) {
+		struct disk_reservation disk_res =
+			bch2_disk_reservation_init(trans->c, 0);
 		struct bkey_i delete;
 
 		bkey_init(&delete.k);
 
+		/*
+		 * This could probably be more efficient for extents:
+		 */
+
 		/*
 		 * For extents, iter.pos won't necessarily be the same as
 		 * bkey_start_pos(k.k) (for non extents they always will be the
@@ -901,29 +1687,28 @@ retry:
 		 * (bch2_btree_iter_peek() does guarantee that iter.pos >=
 		 * bkey_start_pos(k.k)).
 		 */
-		delete.k.p = iter->pos;
+		delete.k.p = iter.pos;
 
-		if (iter->flags & BTREE_ITER_IS_EXTENTS) {
+		if (iter.flags & BTREE_ITER_IS_EXTENTS) {
 			unsigned max_sectors =
 				KEY_SIZE_MAX & (~0 << trans->c->block_bits);
 
 			/* create the biggest key we can */
 			bch2_key_resize(&delete.k, max_sectors);
-			bch2_cut_back(end, &delete.k);
+			bch2_cut_back(end, &delete);
 
-			ret = bch2_extent_trim_atomic(&delete, iter);
+			ret = bch2_extent_trim_atomic(trans, &iter, &delete);
 			if (ret)
 				break;
 		}
 
-		bch2_trans_update(trans, iter, &delete);
-		ret = bch2_trans_commit(trans, NULL, journal_seq,
-					BTREE_INSERT_ATOMIC|
-					BTREE_INSERT_NOFAIL);
+		ret   = bch2_trans_update(trans, &iter, &delete, 0) ?:
+			bch2_trans_commit(trans, &disk_res, journal_seq,
+					  BTREE_INSERT_NOFAIL|
+					  update_flags);
+		bch2_disk_reservation_put(trans->c, &disk_res);
 		if (ret)
 			break;
-
-		bch2_trans_cond_resched(trans);
 	}
 
 	if (ret == -EINTR) {
@@ -931,22 +1716,8 @@ retry:
 		goto retry;
 	}
 
+	bch2_trans_iter_exit(trans, &iter);
 	return ret;
-
-}
-
-int bch2_btree_delete_at(struct btree_trans *trans,
-			 struct btree_iter *iter, unsigned flags)
-{
-	struct bkey_i k;
-
-	bkey_init(&k.k);
-	k.k.p = iter->pos;
-
-	bch2_trans_update(trans, iter, &k);
-	return bch2_trans_commit(trans, NULL, NULL,
-				 BTREE_INSERT_NOFAIL|
-				 BTREE_INSERT_USE_RESERVE|flags);
 }
 
 /*
@@ -956,23 +1727,10 @@ int bch2_btree_delete_at(struct btree_trans *trans,
  */
 int bch2_btree_delete_range(struct bch_fs *c, enum btree_id id,
 			    struct bpos start, struct bpos end,
+			    unsigned update_flags,
 			    u64 *journal_seq)
 {
-	struct btree_trans trans;
-	struct btree_iter *iter;
-	int ret = 0;
-
-	/*
-	 * XXX: whether we need mem/more iters depends on whether this btree id
-	 * has triggers
-	 */
-	bch2_trans_init(&trans, c, BTREE_ITER_MAX, 512);
-
-	iter = bch2_trans_get_iter(&trans, id, start, BTREE_ITER_INTENT);
-
-	ret = bch2_btree_delete_at_range(&trans, iter, end, journal_seq);
-	ret = bch2_trans_exit(&trans) ?: ret;
-
-	BUG_ON(ret == -EINTR);
-	return ret;
+	return bch2_trans_do(c, NULL, journal_seq, 0,
+			     bch2_btree_delete_range_trans(&trans, id, start, end,
+							   update_flags, journal_seq));
 }