Merge with d7dbd2ce12 bcachefs: Add some printks for error paths

16 files changed, 172 insertions, 386 deletions

diff --git a/fs/bcachefs/bcachefs.h b/fs/bcachefs/bcachefs.h
index 289d7ae4e98c..fa9593764f0c 100644
--- a/fs/bcachefs/bcachefs.h
+++ b/fs/bcachefs/bcachefs.h
@@ -732,7 +732,7 @@ struct bch_fs {
 	ZSTD_parameters		zstd_params;
 
 	struct crypto_shash	*sha256;
-	struct crypto_skcipher	*chacha20;
+	struct crypto_sync_skcipher *chacha20;
 	struct crypto_shash	*poly1305;
 
 	atomic64_t		key_version;
diff --git a/fs/bcachefs/bkey_methods.c b/fs/bcachefs/bkey_methods.c
index c97e1e9002cb..55ef4032b37c 100644
--- a/fs/bcachefs/bkey_methods.c
+++ b/fs/bcachefs/bkey_methods.c
@@ -283,49 +283,64 @@ void __bch2_bkey_compat(unsigned level, enum btree_id btree_id,
 	const struct bkey_ops *ops;
 	struct bkey uk;
 	struct bkey_s u;
-
-	if (big_endian != CPU_BIG_ENDIAN)
-		bch2_bkey_swab_key(f, k);
-
-	if (version < bcachefs_metadata_version_bkey_renumber)
-		bch2_bkey_renumber(__btree_node_type(level, btree_id), k, write);
-
-	if (version < bcachefs_metadata_version_inode_btree_change &&
-	    btree_id == BTREE_ID_INODES) {
+	int i;
+
+	/*
+	 * Do these operations in reverse order in the write path:
+	 */
+
+	for (i = 0; i < 4; i++)
+	switch (!write ? i : 3 - i) {
+	case 0:
+		if (big_endian != CPU_BIG_ENDIAN)
+			bch2_bkey_swab_key(f, k);
+		break;
+	case 1:
+		if (version < bcachefs_metadata_version_bkey_renumber)
+			bch2_bkey_renumber(__btree_node_type(level, btree_id), k, write);
+		break;
+	case 2:
+		if (version < bcachefs_metadata_version_inode_btree_change &&
+		    btree_id == BTREE_ID_INODES) {
+			if (!bkey_packed(k)) {
+				struct bkey_i *u = packed_to_bkey(k);
+				swap(u->k.p.inode, u->k.p.offset);
+			} else if (f->bits_per_field[BKEY_FIELD_INODE] &&
+				   f->bits_per_field[BKEY_FIELD_OFFSET]) {
+				struct bkey_format tmp = *f, *in = f, *out = &tmp;
+
+				swap(tmp.bits_per_field[BKEY_FIELD_INODE],
+				     tmp.bits_per_field[BKEY_FIELD_OFFSET]);
+				swap(tmp.field_offset[BKEY_FIELD_INODE],
+				     tmp.field_offset[BKEY_FIELD_OFFSET]);
+
+				if (!write)
+					swap(in, out);
+
+				uk = __bch2_bkey_unpack_key(in, k);
+				swap(uk.p.inode, uk.p.offset);
+				BUG_ON(!bch2_bkey_pack_key(k, &uk, out));
+			}
+		}
+		break;
+	case 3:
 		if (!bkey_packed(k)) {
-			struct bkey_i *u = packed_to_bkey(k);
-			swap(u->k.p.inode, u->k.p.offset);
-		} else if (f->bits_per_field[BKEY_FIELD_INODE] &&
-			   f->bits_per_field[BKEY_FIELD_OFFSET]) {
-			struct bkey_format tmp = *f, *in = f, *out = &tmp;
-
-			swap(tmp.bits_per_field[BKEY_FIELD_INODE],
-			     tmp.bits_per_field[BKEY_FIELD_OFFSET]);
-			swap(tmp.field_offset[BKEY_FIELD_INODE],
-			     tmp.field_offset[BKEY_FIELD_OFFSET]);
-
-			if (!write)
-				swap(in, out);
-
-			uk = __bch2_bkey_unpack_key(in, k);
-			swap(uk.p.inode, uk.p.offset);
-			BUG_ON(!bch2_bkey_pack_key(k, &uk, out));
+			u = bkey_i_to_s(packed_to_bkey(k));
+		} else {
+			uk = __bch2_bkey_unpack_key(f, k);
+			u.k = &uk;
+			u.v = bkeyp_val(f, k);
 		}
-	}
 
-	if (!bkey_packed(k)) {
-		u = bkey_i_to_s(packed_to_bkey(k));
-	} else {
-		uk = __bch2_bkey_unpack_key(f, k);
-		u.k = &uk;
-		u.v = bkeyp_val(f, k);
-	}
+		if (big_endian != CPU_BIG_ENDIAN)
+			bch2_bkey_swab_val(u);
 
-	if (big_endian != CPU_BIG_ENDIAN)
-		bch2_bkey_swab_val(u);
+		ops = &bch2_bkey_ops[k->type];
 
-	ops = &bch2_bkey_ops[k->type];
-
-	if (ops->compat)
-		ops->compat(btree_id, version, big_endian, write, u);
+		if (ops->compat)
+			ops->compat(btree_id, version, big_endian, write, u);
+		break;
+	default:
+		BUG();
+	}
 }
diff --git a/fs/bcachefs/btree_gc.c b/fs/bcachefs/btree_gc.c
index cef8e148f784..146f2428fe04 100644
--- a/fs/bcachefs/btree_gc.c
+++ b/fs/bcachefs/btree_gc.c
@@ -699,8 +699,10 @@ static int bch2_gc_start(struct bch_fs *c,
 
 	c->usage_gc = __alloc_percpu_gfp(fs_usage_u64s(c) * sizeof(u64),
 					 sizeof(u64), GFP_KERNEL);
-	if (!c->usage_gc)
+	if (!c->usage_gc) {
+		bch_err(c, "error allocating c->usage_gc");
 		return -ENOMEM;
+	}
 
 	for_each_member_device(ca, c, i) {
 		BUG_ON(ca->buckets[1]);
@@ -711,19 +713,23 @@ static int bch2_gc_start(struct bch_fs *c,
 				GFP_KERNEL|__GFP_ZERO);
 		if (!ca->buckets[1]) {
 			percpu_ref_put(&ca->ref);
+			bch_err(c, "error allocating ca->buckets[gc]");
 			return -ENOMEM;
 		}
 
 		ca->usage[1] = alloc_percpu(struct bch_dev_usage);
 		if (!ca->usage[1]) {
+			bch_err(c, "error allocating ca->usage[gc]");
 			percpu_ref_put(&ca->ref);
 			return -ENOMEM;
 		}
 	}
 
 	ret = bch2_ec_mem_alloc(c, true);
-	if (ret)
+	if (ret) {
+		bch_err(c, "error allocating ec gc mem");
 		return ret;
+	}
 
 	percpu_down_write(&c->mark_lock);
 
diff --git a/fs/bcachefs/btree_io.c b/fs/bcachefs/btree_io.c
index 61d4c98d3476..ac8b98861aae 100644
--- a/fs/bcachefs/btree_io.c
+++ b/fs/bcachefs/btree_io.c
@@ -620,7 +620,7 @@ static void bset_encrypt(struct bch_fs *c, struct bset *i, unsigned offset)
 		bch2_encrypt(c, BSET_CSUM_TYPE(i), nonce, &bn->flags,
 			     bytes);
 
-		nonce = nonce_add(nonce, round_up(bytes, CHACHA20_BLOCK_SIZE));
+		nonce = nonce_add(nonce, round_up(bytes, CHACHA_BLOCK_SIZE));
 	}
 
 	bch2_encrypt(c, BSET_CSUM_TYPE(i), nonce, i->_data,
diff --git a/fs/bcachefs/btree_locking.h b/fs/bcachefs/btree_locking.h
index aaad2d289e79..9081d3fc238a 100644
--- a/fs/bcachefs/btree_locking.h
+++ b/fs/bcachefs/btree_locking.h
@@ -165,8 +165,7 @@ static inline bool btree_node_lock_increment(struct btree_iter *iter,
 	struct btree_iter *linked;
 
 	trans_for_each_iter(iter->trans, linked)
-		if (linked != iter &&
-		    linked->l[level].b == b &&
+		if (linked->l[level].b == b &&
 		    btree_node_locked_type(linked, level) >= want) {
 			six_lock_increment(&b->lock, want);
 			return true;
diff --git a/fs/bcachefs/checksum.c b/fs/bcachefs/checksum.c
index 9ef2cb9b9904..6f1afa4a3119 100644
--- a/fs/bcachefs/checksum.c
+++ b/fs/bcachefs/checksum.c
@@ -10,7 +10,7 @@
 #include <linux/random.h>
 #include <linux/scatterlist.h>
 #include <crypto/algapi.h>
-#include <crypto/chacha20.h>
+#include <crypto/chacha.h>
 #include <crypto/hash.h>
 #include <crypto/poly1305.h>
 #include <keys/user-type.h>
@@ -67,21 +67,21 @@ static u64 bch2_checksum_update(unsigned type, u64 crc, const void *data, size_t
 	}
 }
 
-static inline void do_encrypt_sg(struct crypto_skcipher *tfm,
+static inline void do_encrypt_sg(struct crypto_sync_skcipher *tfm,
 				 struct nonce nonce,
 				 struct scatterlist *sg, size_t len)
 {
-	SKCIPHER_REQUEST_ON_STACK(req, tfm);
+	SYNC_SKCIPHER_REQUEST_ON_STACK(req, tfm);
 	int ret;
 
-	skcipher_request_set_tfm(req, tfm);
+	skcipher_request_set_sync_tfm(req, tfm);
 	skcipher_request_set_crypt(req, sg, sg, len, nonce.d);
 
 	ret = crypto_skcipher_encrypt(req);
 	BUG_ON(ret);
 }
 
-static inline void do_encrypt(struct crypto_skcipher *tfm,
+static inline void do_encrypt(struct crypto_sync_skcipher *tfm,
 			      struct nonce nonce,
 			      void *buf, size_t len)
 {
@@ -94,8 +94,8 @@ static inline void do_encrypt(struct crypto_skcipher *tfm,
 int bch2_chacha_encrypt_key(struct bch_key *key, struct nonce nonce,
 			    void *buf, size_t len)
 {
-	struct crypto_skcipher *chacha20 =
-		crypto_alloc_skcipher("chacha20", 0, 0);
+	struct crypto_sync_skcipher *chacha20 =
+		crypto_alloc_sync_skcipher("chacha20", 0, 0);
 	int ret;
 
 	if (!chacha20) {
@@ -103,7 +103,8 @@ int bch2_chacha_encrypt_key(struct bch_key *key, struct nonce nonce,
 		return PTR_ERR(chacha20);
 	}
 
-	ret = crypto_skcipher_setkey(chacha20, (void *) key, sizeof(*key));
+	ret = crypto_skcipher_setkey(&chacha20->base,
+				     (void *) key, sizeof(*key));
 	if (ret) {
 		pr_err("crypto_skcipher_setkey() error: %i", ret);
 		goto err;
@@ -111,7 +112,7 @@ int bch2_chacha_encrypt_key(struct bch_key *key, struct nonce nonce,
 
 	do_encrypt(chacha20, nonce, buf, len);
 err:
-	crypto_free_skcipher(chacha20);
+	crypto_free_sync_skcipher(chacha20);
 	return ret;
 }
 
@@ -198,7 +199,7 @@ static struct bch_csum __bch2_checksum_bio(struct bch_fs *c, unsigned type,
 			kunmap_atomic(p);
 		}
 #else
-		__bio_for_each_contig_segment(bv, bio, *iter, *iter)
+		__bio_for_each_bvec(bv, bio, *iter, *iter)
 			crc = bch2_checksum_update(type, crc,
 				page_address(bv.bv_page) + bv.bv_offset,
 				bv.bv_len);
@@ -223,7 +224,7 @@ static struct bch_csum __bch2_checksum_bio(struct bch_fs *c, unsigned type,
 			kunmap_atomic(p);
 		}
 #else
-		__bio_for_each_contig_segment(bv, bio, *iter, *iter)
+		__bio_for_each_bvec(bv, bio, *iter, *iter)
 			crypto_shash_update(desc,
 				page_address(bv.bv_page) + bv.bv_offset,
 				bv.bv_len);
@@ -462,7 +463,7 @@ err:
 static int bch2_alloc_ciphers(struct bch_fs *c)
 {
 	if (!c->chacha20)
-		c->chacha20 = crypto_alloc_skcipher("chacha20", 0, 0);
+		c->chacha20 = crypto_alloc_sync_skcipher("chacha20", 0, 0);
 	if (IS_ERR(c->chacha20)) {
 		bch_err(c, "error requesting chacha20 module: %li",
 			PTR_ERR(c->chacha20));
@@ -545,7 +546,7 @@ int bch2_enable_encryption(struct bch_fs *c, bool keyed)
 			goto err;
 	}
 
-	ret = crypto_skcipher_setkey(c->chacha20,
+	ret = crypto_skcipher_setkey(&c->chacha20->base,
 			(void *) &key.key, sizeof(key.key));
 	if (ret)
 		goto err;
@@ -573,7 +574,7 @@ void bch2_fs_encryption_exit(struct bch_fs *c)
 	if (!IS_ERR_OR_NULL(c->poly1305))
 		crypto_free_shash(c->poly1305);
 	if (!IS_ERR_OR_NULL(c->chacha20))
-		crypto_free_skcipher(c->chacha20);
+		crypto_free_sync_skcipher(c->chacha20);
 	if (!IS_ERR_OR_NULL(c->sha256))
 		crypto_free_shash(c->sha256);
 }
@@ -605,7 +606,7 @@ int bch2_fs_encryption_init(struct bch_fs *c)
 	if (ret)
 		goto out;
 
-	ret = crypto_skcipher_setkey(c->chacha20,
+	ret = crypto_skcipher_setkey(&c->chacha20->base,
 			(void *) &key.key, sizeof(key.key));
 	if (ret)
 		goto out;
diff --git a/fs/bcachefs/checksum.h b/fs/bcachefs/checksum.h
index 833537cc8fd0..24dee8039d57 100644
--- a/fs/bcachefs/checksum.h
+++ b/fs/bcachefs/checksum.h
@@ -7,7 +7,7 @@
 #include "super-io.h"
 
 #include <linux/crc64.h>
-#include <crypto/chacha20.h>
+#include <crypto/chacha.h>
 
 static inline bool bch2_checksum_mergeable(unsigned type)
 {
@@ -138,9 +138,9 @@ static inline bool bch2_crc_cmp(struct bch_csum l, struct bch_csum r)
 /* for skipping ahead and encrypting/decrypting at an offset: */
 static inline struct nonce nonce_add(struct nonce nonce, unsigned offset)
 {
-	EBUG_ON(offset & (CHACHA20_BLOCK_SIZE - 1));
+	EBUG_ON(offset & (CHACHA_BLOCK_SIZE - 1));
 
-	le32_add_cpu(&nonce.d[0], offset / CHACHA20_BLOCK_SIZE);
+	le32_add_cpu(&nonce.d[0], offset / CHACHA_BLOCK_SIZE);
 	return nonce;
 }
 
diff --git a/fs/bcachefs/compress.c b/fs/bcachefs/compress.c
index 4a163b6d59da..0713286d7999 100644
--- a/fs/bcachefs/compress.c
+++ b/fs/bcachefs/compress.c
@@ -55,7 +55,7 @@ static struct bbuf __bio_map_or_bounce(struct bch_fs *c, struct bio *bio,
 	BUG_ON(bvec_iter_sectors(start) > c->sb.encoded_extent_max);
 
 #ifndef CONFIG_HIGHMEM
-	__bio_for_each_contig_segment(bv, bio, iter, start) {
+	__bio_for_each_bvec(bv, bio, iter, start) {
 		if (bv.bv_len == start.bi_size)
 			return (struct bbuf) {
 				.b = page_address(bv.bv_page) + bv.bv_offset,
diff --git a/fs/bcachefs/extent_update.c b/fs/bcachefs/extent_update.c
index 2a7d913bdda3..d0af1bc17018 100644
--- a/fs/bcachefs/extent_update.c
+++ b/fs/bcachefs/extent_update.c
@@ -34,16 +34,10 @@ static int count_iters_for_insert(struct btree_trans *trans,
 				  unsigned offset,
 				  struct bpos *end,
 				  unsigned *nr_iters,
-				  unsigned max_iters,
-				  bool overwrite)
+				  unsigned max_iters)
 {
-	int ret = 0;
+	int ret = 0, ret2 = 0;
 
-	/*
-	 * The extent update path requires an _additional_ iterator for each
-	 * extent we're inserting and overwriting:
-	 */
-	*nr_iters += 1;
 	if (*nr_iters >= max_iters) {
 		*end = bpos_min(*end, k.k->p);
 		ret = 1;
@@ -70,11 +64,14 @@ static int count_iters_for_insert(struct btree_trans *trans,
 
 		for_each_btree_key(trans, iter,
 				   BTREE_ID_REFLINK, POS(0, idx + offset),
-				   BTREE_ITER_SLOTS, r_k, ret) {
+				   BTREE_ITER_SLOTS, r_k, ret2) {
 			if (bkey_cmp(bkey_start_pos(r_k.k),
 				     POS(0, idx + sectors)) >= 0)
 				break;
 
+			/* extent_update_to_keys(), for the reflink_v update */
+			*nr_iters += 1;
+
 			*nr_iters += 1 + bch2_bkey_nr_alloc_ptrs(r_k);
 
 			if (*nr_iters >= max_iters) {
@@ -92,7 +89,7 @@ static int count_iters_for_insert(struct btree_trans *trans,
 	}
 	}
 
-	return ret;
+	return ret2 ?: ret;
 }
 
 #define EXTENT_ITERS_MAX	(BTREE_ITER_MAX / 3)
@@ -121,8 +118,11 @@ int bch2_extent_atomic_end(struct btree_iter *iter,
 
 	*end = bpos_min(insert->k.p, b->key.k.p);
 
+	/* extent_update_to_keys(): */
+	nr_iters += 1;
+
 	ret = count_iters_for_insert(trans, bkey_i_to_s_c(insert), 0, end,
-				     &nr_iters, EXTENT_ITERS_MAX / 2, false);
+				     &nr_iters, EXTENT_ITERS_MAX / 2);
 	if (ret < 0)
 		return ret;
 
@@ -139,8 +139,20 @@ int bch2_extent_atomic_end(struct btree_iter *iter,
 			offset = bkey_start_offset(&insert->k) -
 				bkey_start_offset(k.k);
 
+		/* extent_handle_overwrites(): */
+		switch (bch2_extent_overlap(&insert->k, k.k)) {
+		case BCH_EXTENT_OVERLAP_ALL:
+		case BCH_EXTENT_OVERLAP_FRONT:
+			nr_iters += 1;
+			break;
+		case BCH_EXTENT_OVERLAP_BACK:
+		case BCH_EXTENT_OVERLAP_MIDDLE:
+			nr_iters += 2;
+			break;
+		}
+
 		ret = count_iters_for_insert(trans, k, offset, end,
-					&nr_iters, EXTENT_ITERS_MAX, true);
+					&nr_iters, EXTENT_ITERS_MAX);
 		if (ret)
 			break;
 
diff --git a/fs/bcachefs/fs-io.c b/fs/bcachefs/fs-io.c
index 707ce27eace6..7de61f7f4314 100644
--- a/fs/bcachefs/fs-io.c
+++ b/fs/bcachefs/fs-io.c
@@ -602,7 +602,7 @@ int bch2_migrate_page(struct address_space *mapping, struct page *newpage,
 	EBUG_ON(!PageLocked(page));
 	EBUG_ON(!PageLocked(newpage));
 
-	ret = migrate_page_move_mapping(mapping, newpage, page, NULL, mode, 0);
+	ret = migrate_page_move_mapping(mapping, newpage, page, 0);
 	if (ret != MIGRATEPAGE_SUCCESS)
 		return ret;
 
@@ -627,10 +627,10 @@ int bch2_migrate_page(struct address_space *mapping, struct page *newpage,
 
 static void bch2_readpages_end_io(struct bio *bio)
 {
+	struct bvec_iter_all iter;
 	struct bio_vec *bv;
-	unsigned i;
 
-	bio_for_each_segment_all(bv, bio, i) {
+	bio_for_each_segment_all(bv, bio, iter) {
 		struct page *page = bv->bv_page;
 
 		if (!bio->bi_status) {
@@ -782,11 +782,8 @@ static void readpage_bio_extend(struct readpages_iter *iter,
 			if (!get_more)
 				break;
 
-			rcu_read_lock();
-			page = radix_tree_lookup(&iter->mapping->i_pages, page_offset);
-			rcu_read_unlock();
-
-			if (page && !radix_tree_exceptional_entry(page))
+			page = xa_load(&iter->mapping->i_pages, page_offset);
+			if (page && !xa_is_value(page))
 				break;
 
 			page = __page_cache_alloc(readahead_gfp_mask(iter->mapping));
@@ -1037,32 +1034,33 @@ static void bch2_writepage_io_done(struct closure *cl)
 					struct bch_writepage_io, cl);
 	struct bch_fs *c = io->op.c;
 	struct bio *bio = &io->op.wbio.bio;
+	struct bvec_iter_all iter;
 	struct bio_vec *bvec;
-	unsigned i, j;
+	unsigned i;
 
 	if (io->op.error) {
-		bio_for_each_segment_all(bvec, bio, i) {
+		bio_for_each_segment_all(bvec, bio, iter) {
 			struct bch_page_state *s;
 
 			SetPageError(bvec->bv_page);
-			mapping_set_error(io->inode->v.i_mapping, -EIO);
+			mapping_set_error(bvec->bv_page->mapping, -EIO);
 
 			s = __bch2_page_state(bvec->bv_page);
 			spin_lock(&s->lock);
-			for (j = 0; j < PAGE_SECTORS; j++)
-				s->s[j].nr_replicas = 0;
+			for (i = 0; i < PAGE_SECTORS; i++)
+				s->s[i].nr_replicas = 0;
 			spin_unlock(&s->lock);
 		}
 	}
 
 	if (io->op.flags & BCH_WRITE_WROTE_DATA_INLINE) {
-		bio_for_each_segment_all(bvec, bio, i) {
+		bio_for_each_segment_all(bvec, bio, iter) {
 			struct bch_page_state *s;
 
 			s = __bch2_page_state(bvec->bv_page);
 			spin_lock(&s->lock);
-			for (j = 0; j < PAGE_SECTORS; j++)
-				s->s[j].nr_replicas = 0;
+			for (i = 0; i < PAGE_SECTORS; i++)
+				s->s[i].nr_replicas = 0;
 			spin_unlock(&s->lock);
 		}
 	}
@@ -1086,7 +1084,7 @@ static void bch2_writepage_io_done(struct closure *cl)
 	 */
 	i_sectors_acct(c, io->inode, NULL, io->op.i_sectors_delta);
 
-	bio_for_each_segment_all(bvec, bio, i) {
+	bio_for_each_segment_all(bvec, bio, iter) {
 		struct bch_page_state *s = __bch2_page_state(bvec->bv_page);
 
 		if (atomic_dec_and_test(&s->write_count))
@@ -1240,8 +1238,9 @@ do_io:
 
 		if (w->io &&
 		    (w->io->op.res.nr_replicas != nr_replicas_this_write ||
-		     bio_full(&w->io->op.wbio.bio) ||
-		     w->io->op.wbio.bio.bi_iter.bi_size >= (256U << 20) ||
+		     bio_full(&w->io->op.wbio.bio, PAGE_SIZE) ||
+		     w->io->op.wbio.bio.bi_iter.bi_size + (sectors << 9) >=
+		     (BIO_MAX_PAGES * PAGE_SIZE) ||
 		     bio_end_sector(&w->io->op.wbio.bio) != sector))
 			bch2_writepage_do_io(w);
 
@@ -1805,8 +1804,9 @@ static long bch2_dio_write_loop(struct dio_write *dio)
 	struct address_space *mapping = req->ki_filp->f_mapping;
 	struct bch_inode_info *inode = file_bch_inode(req->ki_filp);
 	struct bio *bio = &dio->op.wbio.bio;
+	struct bvec_iter_all iter;
 	struct bio_vec *bv;
-	unsigned i, unaligned;
+	unsigned unaligned;
 	u64 new_i_size;
 	bool sync = dio->sync;
 	long ret;
@@ -1815,12 +1815,22 @@ static long bch2_dio_write_loop(struct dio_write *dio)
 		goto loop;
 
 	while (1) {
+		size_t extra = dio->iter.count -
+			min(BIO_MAX_PAGES * PAGE_SIZE, dio->iter.count);
+
 		if (kthread)
 			use_mm(dio->mm);
 		BUG_ON(current->faults_disabled_mapping);
 		current->faults_disabled_mapping = mapping;
 
+		/*
+		 * Don't issue more than 2MB at once, the bcachefs io path in
+		 * io.c can't bounce more than that:
+		 */
+
+		dio->iter.count -= extra;
 		ret = bio_iov_iter_get_pages(bio, &dio->iter);
+		dio->iter.count += extra;
 
 		current->faults_disabled_mapping = NULL;
 		if (kthread)
@@ -1838,7 +1848,7 @@ static long bch2_dio_write_loop(struct dio_write *dio)
 			 * bio_iov_iter_get_pages was only able to get <
 			 * blocksize worth of pages:
 			 */
-			bio_for_each_segment_all(bv, bio, i)
+			bio_for_each_segment_all(bv, bio, iter)
 				put_page(bv->bv_page);
 			ret = -EFAULT;
 			goto err;
@@ -1886,7 +1896,7 @@ loop:
 			i_size_write(&inode->v, new_i_size);
 		spin_unlock(&inode->v.i_lock);
 
-		bio_for_each_segment_all(bv, bio, i)
+		bio_for_each_segment_all(bv, bio, iter)
 			put_page(bv->bv_page);
 		if (!dio->iter.count || dio->op.error)
 			break;
@@ -2823,235 +2833,6 @@ static void mark_range_unallocated(struct bch_inode_info *inode,
 	} while (index <= end_index);
 }
 
-static int generic_access_check_limits(struct file *file, loff_t pos,
-				       loff_t *count)
-{
-	struct inode *inode = file->f_mapping->host;
-	loff_t max_size = inode->i_sb->s_maxbytes;
-
-	if (!(file->f_flags & O_LARGEFILE))
-		max_size = MAX_NON_LFS;
-
-	if (unlikely(pos >= max_size))
-		return -EFBIG;
-	*count = min(*count, max_size - pos);
-	return 0;
-}
-
-static int generic_write_check_limits(struct file *file, loff_t pos,
-				      loff_t *count)
-{
-	loff_t limit = rlimit(RLIMIT_FSIZE);
-
-	if (limit != RLIM_INFINITY) {
-		if (pos >= limit) {
-			send_sig(SIGXFSZ, current, 0);
-			return -EFBIG;
-		}
-		*count = min(*count, limit - pos);
-	}
-
-	return generic_access_check_limits(file, pos, count);
-}
-
-static int generic_remap_checks(struct file *file_in, loff_t pos_in,
-			 struct file *file_out, loff_t pos_out,
-			 loff_t *req_count, unsigned int remap_flags)
-{
-	struct inode *inode_in = file_in->f_mapping->host;
-	struct inode *inode_out = file_out->f_mapping->host;
-	uint64_t count = *req_count;
-	uint64_t bcount;
-	loff_t size_in, size_out;
-	loff_t bs = inode_out->i_sb->s_blocksize;
-	int ret;
-
-	/* The start of both ranges must be aligned to an fs block. */
-	if (!IS_ALIGNED(pos_in, bs) || !IS_ALIGNED(pos_out, bs))
-		return -EINVAL;
-
-	/* Ensure offsets don't wrap. */
-	if (pos_in + count < pos_in || pos_out + count < pos_out)
-		return -EINVAL;
-
-	size_in = i_size_read(inode_in);
-	size_out = i_size_read(inode_out);
-
-	/* Dedupe requires both ranges to be within EOF. */
-	if ((remap_flags & REMAP_FILE_DEDUP) &&
-	    (pos_in >= size_in || pos_in + count > size_in ||
-	     pos_out >= size_out || pos_out + count > size_out))
-		return -EINVAL;
-
-	/* Ensure the infile range is within the infile. */
-	if (pos_in >= size_in)
-		return -EINVAL;
-	count = min(count, size_in - (uint64_t)pos_in);
-
-	ret = generic_access_check_limits(file_in, pos_in, &count);
-	if (ret)
-		return ret;
-
-	ret = generic_write_check_limits(file_out, pos_out, &count);
-	if (ret)
-		return ret;
-
-	/*
-	 * If the user wanted us to link to the infile's EOF, round up to the
-	 * next block boundary for this check.
-	 *
-	 * Otherwise, make sure the count is also block-aligned, having
-	 * already confirmed the starting offsets' block alignment.
-	 */
-	if (pos_in + count == size_in) {
-		bcount = ALIGN(size_in, bs) - pos_in;
-	} else {
-		if (!IS_ALIGNED(count, bs))
-			count = ALIGN_DOWN(count, bs);
-		bcount = count;
-	}
-
-	/* Don't allow overlapped cloning within the same file. */
-	if (inode_in == inode_out &&
-	    pos_out + bcount > pos_in &&
-	    pos_out < pos_in + bcount)
-		return -EINVAL;
-
-	/*
-	 * We shortened the request but the caller can't deal with that, so
-	 * bounce the request back to userspace.
-	 */
-	if (*req_count != count && !(remap_flags & REMAP_FILE_CAN_SHORTEN))
-		return -EINVAL;
-
-	*req_count = count;
-	return 0;
-}
-
-static int generic_remap_check_len(struct inode *inode_in,
-				   struct inode *inode_out,
-				   loff_t pos_out,
-				   loff_t *len,
-				   unsigned int remap_flags)
-{
-	u64 blkmask = i_blocksize(inode_in) - 1;
-	loff_t new_len = *len;
-
-	if ((*len & blkmask) == 0)
-		return 0;
-
-	if ((remap_flags & REMAP_FILE_DEDUP) ||
-	    pos_out + *len < i_size_read(inode_out))
-		new_len &= ~blkmask;
-
-	if (new_len == *len)
-		return 0;
-
-	if (remap_flags & REMAP_FILE_CAN_SHORTEN) {
-		*len = new_len;
-		return 0;
-	}
-
-	return (remap_flags & REMAP_FILE_DEDUP) ? -EBADE : -EINVAL;
-}
-
-static int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in,
-				  struct file *file_out, loff_t pos_out,
-				  loff_t *len, unsigned int remap_flags)
-{
-	struct inode *inode_in = file_inode(file_in);
-	struct inode *inode_out = file_inode(file_out);
-	bool same_inode = (inode_in == inode_out);
-	int ret;
-
-	/* Don't touch certain kinds of inodes */
-	if (IS_IMMUTABLE(inode_out))
-		return -EPERM;
-
-	if (IS_SWAPFILE(inode_in) || IS_SWAPFILE(inode_out))
-		return -ETXTBSY;
-
-	/* Don't reflink dirs, pipes, sockets... */
-	if (S_ISDIR(inode_in->i_mode) || S_ISDIR(inode_out->i_mode))
-		return -EISDIR;
-	if (!S_ISREG(inode_in->i_mode) || !S_ISREG(inode_out->i_mode))
-		return -EINVAL;
-
-	/* Zero length dedupe exits immediately; reflink goes to EOF. */
-	if (*len == 0) {
-		loff_t isize = i_size_read(inode_in);
-
-		if ((remap_flags & REMAP_FILE_DEDUP) || pos_in == isize)
-			return 0;
-		if (pos_in > isize)
-			return -EINVAL;
-		*len = isize - pos_in;
-		if (*len == 0)
-			return 0;
-	}
-
-	/* Check that we don't violate system file offset limits. */
-	ret = generic_remap_checks(file_in, pos_in, file_out, pos_out, len,
-			remap_flags);
-	if (ret)
-		return ret;
-
-	/* Wait for the completion of any pending IOs on both files */
-	inode_dio_wait(inode_in);
-	if (!same_inode)
-		inode_dio_wait(inode_out);
-
-	ret = filemap_write_and_wait_range(inode_in->i_mapping,
-			pos_in, pos_in + *len - 1);
-	if (ret)
-		return ret;
-
-	ret = filemap_write_and_wait_range(inode_out->i_mapping,
-			pos_out, pos_out + *len - 1);
-	if (ret)
-		return ret;
-
-	/*
-	 * Check that the extents are the same.
-	 */
-	if (remap_flags & REMAP_FILE_DEDUP) {
-		bool		is_same = false;
-
-		ret = vfs_dedupe_file_range_compare(inode_in, pos_in,
-				inode_out, pos_out, *len, &is_same);
-		if (ret)
-			return ret;
-		if (!is_same)
-			return -EBADE;
-	}
-
-	ret = generic_remap_check_len(inode_in, inode_out, pos_out, len,
-			remap_flags);
-	if (ret)
-		return ret;
-
-	/* If can't alter the file contents, we're done. */
-	if (!(remap_flags & REMAP_FILE_DEDUP)) {
-		/* Update the timestamps, since we can alter file contents. */
-		if (!(file_out->f_mode & FMODE_NOCMTIME)) {
-			ret = file_update_time(file_out);
-			if (ret)
-				return ret;
-		}
-
-		/*
-		 * Clear the security bits if the process is not being run by
-		 * root.  This keeps people from modifying setuid and setgid
-		 * binaries.
-		 */
-		ret = file_remove_privs(file_out);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
-}
-
 loff_t bch2_remap_file_range(struct file *file_src, loff_t pos_src,
 			     struct file *file_dst, loff_t pos_dst,
 			     loff_t len, unsigned remap_flags)
@@ -3244,7 +3025,7 @@ static loff_t page_hole_offset(struct address_space *mapping, loff_t offset)
 	loff_t ret = -1;
 
 	page = find_lock_entry(mapping, index);
-	if (!page || radix_tree_exception(page))
+	if (!page || xa_is_value(page))
 		return offset;
 
 	pg_offset = __page_hole_offset(page, offset & (PAGE_SIZE - 1));
diff --git a/fs/bcachefs/fs-io.h b/fs/bcachefs/fs-io.h
index 1b593ea707d5..7063556d289b 100644
--- a/fs/bcachefs/fs-io.h
+++ b/fs/bcachefs/fs-io.h
@@ -35,10 +35,6 @@ int bch2_fsync(struct file *, loff_t, loff_t, int);
 int bch2_truncate(struct bch_inode_info *, struct iattr *);
 long bch2_fallocate_dispatch(struct file *, int, loff_t, loff_t);
 
-#define REMAP_FILE_ADVISORY		(0)
-#define REMAP_FILE_DEDUP		(1 << 0)
-#define REMAP_FILE_CAN_SHORTEN		(1 << 1)
-
 loff_t bch2_remap_file_range(struct file *, loff_t, struct file *,
 			     loff_t, loff_t, unsigned);
 
diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c
index b1445bdc3e9d..1c89a1b2c2d0 100644
--- a/fs/bcachefs/fs.c
+++ b/fs/bcachefs/fs.c
@@ -966,15 +966,6 @@ static int bch2_vfs_readdir(struct file *file, struct dir_context *ctx)
 	return bch2_readdir(c, inode->v.i_ino, ctx);
 }
 
-static int bch2_clone_file_range(struct file *file_src, loff_t pos_src,
-				 struct file *file_dst, loff_t pos_dst,
-				 u64 len)
-{
-	return bch2_remap_file_range(file_src, pos_src,
-				     file_dst, pos_dst,
-				     len, 0);
-}
-
 static const struct file_operations bch_file_operations = {
 	.llseek		= bch2_llseek,
 	.read_iter	= bch2_read_iter,
@@ -992,7 +983,7 @@ static const struct file_operations bch_file_operations = {
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= bch2_compat_fs_ioctl,
 #endif
-	.clone_file_range = bch2_clone_file_range,
+	.remap_file_range = bch2_remap_file_range,
 };
 
 static const struct inode_operations bch_file_inode_operations = {
@@ -1523,7 +1514,7 @@ static struct dentry *bch2_mount(struct file_system_type *fs_type,
 
 	sb->s_bdi->congested_fn		= bch2_congested;
 	sb->s_bdi->congested_data	= c;
-	sb->s_bdi->ra_pages		= VM_MAX_READAHEAD * 1024 / PAGE_SIZE;
+	sb->s_bdi->ra_pages		= VM_READAHEAD_PAGES;
 
 	for_each_online_member(ca, c, i) {
 		struct block_device *bdev = ca->disk_sb.bdev;
diff --git a/fs/bcachefs/io.c b/fs/bcachefs/io.c
index ff3dd31c134c..19059702428a 100644
--- a/fs/bcachefs/io.c
+++ b/fs/bcachefs/io.c
@@ -124,10 +124,10 @@ void bch2_latency_acct(struct bch_dev *ca, u64 submit_time, int rw)
 
 void bch2_bio_free_pages_pool(struct bch_fs *c, struct bio *bio)
 {
+	struct bvec_iter_all iter;
 	struct bio_vec *bv;
-	unsigned i;
 
-	bio_for_each_segment_all(bv, bio, i)
+	bio_for_each_segment_all(bv, bio, iter)
 		if (bv->bv_page != ZERO_PAGE(0))
 			mempool_free(bv->bv_page, &c->bio_bounce_pages);
 	bio->bi_vcnt = 0;
diff --git a/fs/bcachefs/move.c b/fs/bcachefs/move.c
index 728c7554779f..67e495bc8aba 100644
--- a/fs/bcachefs/move.c
+++ b/fs/bcachefs/move.c
@@ -70,19 +70,26 @@ static int bch2_migrate_index_update(struct bch_write_op *op)
 				   BTREE_ITER_SLOTS|BTREE_ITER_INTENT);
 
 	while (1) {
-		struct bkey_s_c k = bch2_btree_iter_peek_slot(iter);
+		struct bkey_s_c k;
 		struct bkey_i *insert;
-		struct bkey_i_extent *new =
-			bkey_i_to_extent(bch2_keylist_front(keys));
+		struct bkey_i_extent *new;
 		BKEY_PADDED(k) _new, _insert;
 		const union bch_extent_entry *entry;
 		struct extent_ptr_decoded p;
 		bool did_work = false;
 		int nr;
 
+		bch2_trans_reset(&trans, 0);
+
+		k = bch2_btree_iter_peek_slot(iter);
 		ret = bkey_err(k);
-		if (ret)
+		if (ret) {
+			if (ret == -EINTR)
+				continue;
 			break;
+		}
+
+		new = bkey_i_to_extent(bch2_keylist_front(keys));
 
 		if (bversion_cmp(k.k->version, new->k.version) ||
 		    !bch2_bkey_matches_ptr(c, k, m->ptr, m->offset))
@@ -306,12 +313,12 @@ static void move_free(struct closure *cl)
 {
 	struct moving_io *io = container_of(cl, struct moving_io, cl);
 	struct moving_context *ctxt = io->write.ctxt;
+	struct bvec_iter_all iter;
 	struct bio_vec *bv;
-	unsigned i;
 
 	bch2_disk_reservation_put(io->write.op.c, &io->write.op.res);
 
-	bio_for_each_segment_all(bv, &io->write.op.wbio.bio, i)
+	bio_for_each_segment_all(bv, &io->write.op.wbio.bio, iter)
 		if (bv->bv_page)
 			__free_page(bv->bv_page);
 
diff --git a/fs/bcachefs/replicas.c b/fs/bcachefs/replicas.c
index be4908575f72..67a7128fd9af 100644
--- a/fs/bcachefs/replicas.c
+++ b/fs/bcachefs/replicas.c
@@ -299,8 +299,10 @@ static int replicas_table_update(struct bch_fs *c,
 						GFP_NOIO)) ||
 	    !(new_scratch  = kmalloc(bytes, GFP_NOIO)) ||
 	    (c->usage_gc &&
-	     !(new_gc = __alloc_percpu_gfp(bytes, sizeof(u64), GFP_NOIO))))
+	     !(new_gc = __alloc_percpu_gfp(bytes, sizeof(u64), GFP_NOIO)))) {
+		bch_err(c, "error updating replicas table: memory allocation failure");
 		goto err;
+	}
 
 	if (c->usage_base)
 		__replicas_table_update(new_base,		new_r,
@@ -362,7 +364,7 @@ static int bch2_mark_replicas_slowpath(struct bch_fs *c,
 				struct bch_replicas_entry *new_entry)
 {
 	struct bch_replicas_cpu new_r, new_gc;
-	int ret = -ENOMEM;
+	int ret = 0;
 
 	verify_replicas_entry(new_entry);
 
@@ -409,14 +411,16 @@ static int bch2_mark_replicas_slowpath(struct bch_fs *c,
 		swap(new_gc, c->replicas_gc);
 	percpu_up_write(&c->mark_lock);
 out:
-	ret = 0;
-err:
 	mutex_unlock(&c->sb_lock);
 
 	kfree(new_r.entries);
 	kfree(new_gc.entries);
 
 	return ret;
+err:
+	bch_err(c, "error adding replicas entry: memory allocation failure");
+	ret = -ENOMEM;
+	goto out;
 }
 
 int bch2_mark_replicas(struct bch_fs *c,
@@ -561,6 +565,7 @@ int bch2_replicas_gc_start(struct bch_fs *c, unsigned typemask)
 					 GFP_NOIO);
 	if (!c->replicas_gc.entries) {
 		mutex_unlock(&c->sb_lock);
+		bch_err(c, "error allocating c->replicas_gc");
 		return -ENOMEM;
 	}
 
@@ -586,8 +591,10 @@ retry:
 	nr		= READ_ONCE(c->replicas.nr);
 	new.entry_size	= READ_ONCE(c->replicas.entry_size);
 	new.entries	= kcalloc(nr, new.entry_size, GFP_KERNEL);
-	if (!new.entries)
+	if (!new.entries) {
+		bch_err(c, "error allocating c->replicas_gc");
 		return -ENOMEM;
+	}
 
 	mutex_lock(&c->sb_lock);
 	percpu_down_write(&c->mark_lock);
diff --git a/fs/bcachefs/util.h b/fs/bcachefs/util.h
index 2b19a0038045..0128daba5970 100644
--- a/fs/bcachefs/util.h
+++ b/fs/bcachefs/util.h
@@ -664,35 +664,6 @@ static inline void memset_u64s_tail(void *s, int c, unsigned bytes)
 	memset(s + bytes, c, rem);
 }
 
-static inline struct bio_vec next_contig_bvec(struct bio *bio,
-					      struct bvec_iter *iter)
-{
-	struct bio_vec bv = bio_iter_iovec(bio, *iter);
-
-	bio_advance_iter(bio, iter, bv.bv_len);
-#ifndef CONFIG_HIGHMEM
-	while (iter->bi_size) {
-		struct bio_vec next = bio_iter_iovec(bio, *iter);
-
-		if (page_address(bv.bv_page) + bv.bv_offset + bv.bv_len !=
-		    page_address(next.bv_page) + next.bv_offset)
-			break;
-
-		bv.bv_len += next.bv_len;
-		bio_advance_iter(bio, iter, next.bv_len);
-	}
-#endif
-	return bv;
-}
-
-#define __bio_for_each_contig_segment(bv, bio, iter, start)		\
-	for (iter = (start);						\
-	     (iter).bi_size &&						\
-		((bv = next_contig_bvec((bio), &(iter))), 1);)
-
-#define bio_for_each_contig_segment(bv, bio, iter)			\
-	__bio_for_each_contig_segment(bv, bio, iter, (bio)->bi_iter)
-
 void sort_cmp_size(void *base, size_t num, size_t size,
 	  int (*cmp_func)(const void *, const void *, size_t),
 	  void (*swap_func)(void *, void *, size_t));