bcachefs: Fixes for 4.19bcachefs-v4.19-backport-2.0

44 files changed, 977 insertions, 426 deletions

diff --git a/fs/bcachefs/acl.c b/fs/bcachefs/acl.c
index 5c6ccf685094..0ed933f82189 100644
--- a/fs/bcachefs/acl.c
+++ b/fs/bcachefs/acl.c
@@ -212,7 +212,7 @@ bch2_acl_to_xattr(struct btree_trans *trans,
 	return xattr;
 }
 
-struct posix_acl *bch2_get_acl(struct inode *vinode, int type, bool rcu)
+struct posix_acl *bch2_get_acl(struct inode *vinode, int type)
 {
 	struct bch_inode_info *inode = to_bch_ei(vinode);
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
@@ -224,9 +224,6 @@ struct posix_acl *bch2_get_acl(struct inode *vinode, int type, bool rcu)
 	struct bkey_s_c k;
 	int ret;
 
-	if (rcu)
-		return ERR_PTR(-ECHILD);
-
 	bch2_trans_init(&trans, c, 0, 0);
 retry:
 	bch2_trans_begin(&trans);
@@ -292,8 +289,7 @@ int bch2_set_acl_trans(struct btree_trans *trans, subvol_inum inum,
 	return ret == -ENOENT ? 0 : ret;
 }
 
-int bch2_set_acl(struct user_namespace *mnt_userns,
-		 struct inode *vinode, struct posix_acl *_acl, int type)
+int bch2_set_acl(struct inode *vinode, struct posix_acl *_acl, int type)
 {
 	struct bch_inode_info *inode = to_bch_ei(vinode);
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
@@ -318,7 +314,7 @@ retry:
 	mode = inode_u.bi_mode;
 
 	if (type == ACL_TYPE_ACCESS) {
-		ret = posix_acl_update_mode(mnt_userns, &inode->v, &mode, &acl);
+		ret = posix_acl_update_mode(&inode->v, &mode, &acl);
 		if (ret)
 			goto btree_err;
 	}
diff --git a/fs/bcachefs/acl.h b/fs/bcachefs/acl.h
index 2d76a4897ba8..14cabbc91808 100644
--- a/fs/bcachefs/acl.h
+++ b/fs/bcachefs/acl.h
@@ -26,12 +26,12 @@ typedef struct {
 	__le32		a_version;
 } bch_acl_header;
 
-struct posix_acl *bch2_get_acl(struct inode *, int, bool);
+struct posix_acl *bch2_get_acl(struct inode *, int);
 
 int bch2_set_acl_trans(struct btree_trans *, subvol_inum,
 		       struct bch_inode_unpacked *,
 		       struct posix_acl *, int);
-int bch2_set_acl(struct user_namespace *, struct inode *, struct posix_acl *, int);
+int bch2_set_acl(struct inode *, struct posix_acl *, int);
 int bch2_acl_chmod(struct btree_trans *, subvol_inum,
 		   struct bch_inode_unpacked *,
 		   umode_t, struct posix_acl **);
diff --git a/fs/bcachefs/alloc_background.c b/fs/bcachefs/alloc_background.c
index d0d7690a4940..c2f97d03dad6 100644
--- a/fs/bcachefs/alloc_background.c
+++ b/fs/bcachefs/alloc_background.c
@@ -1072,7 +1072,7 @@ static int bch2_discard_one_bucket(struct btree_trans *trans,
 		blkdev_issue_discard(ca->disk_sb.bdev,
 				     k.k->p.offset * ca->mi.bucket_size,
 				     ca->mi.bucket_size,
-				     GFP_KERNEL);
+				     GFP_KERNEL, 0);
 
 		ret = bch2_trans_relock(trans);
 		if (ret)
@@ -1386,7 +1386,7 @@ void bch2_recalc_capacity(struct bch_fs *c)
 	lockdep_assert_held(&c->state_lock);
 
 	for_each_online_member(ca, c, i) {
-		struct backing_dev_info *bdi = ca->disk_sb.bdev->bd_disk->bdi;
+		struct backing_dev_info *bdi = ca->disk_sb.bdev->bd_bdi;
 
 		ra_pages += bdi->ra_pages;
 	}
diff --git a/fs/bcachefs/bcachefs.h b/fs/bcachefs/bcachefs.h
index ccac2a3fcdf7..38f5753107c6 100644
--- a/fs/bcachefs/bcachefs.h
+++ b/fs/bcachefs/bcachefs.h
@@ -204,6 +204,7 @@
 #include <linux/zstd.h>
 
 #include "bcachefs_format.h"
+#include "darray.h"
 #include "errcode.h"
 #include "fifo.h"
 #include "opts.h"
@@ -211,6 +212,7 @@
 
 #define dynamic_fault(...)		0
 #define race_fault(...)			0
+#define fallthrough			do {} while (0)  /* fallthrough */
 
 #define trace_and_count(_c, _name, ...)					\
 do {									\
@@ -827,7 +829,7 @@ struct bch_fs {
 	ZSTD_parameters		zstd_params;
 
 	struct crypto_shash	*sha256;
-	struct crypto_sync_skcipher *chacha20;
+	struct crypto_skcipher	*chacha20;
 	struct crypto_shash	*poly1305;
 
 	atomic64_t		key_version;
diff --git a/fs/bcachefs/btree_cache.c b/fs/bcachefs/btree_cache.c
index f84b50869de2..67661cbf0ae8 100644
--- a/fs/bcachefs/btree_cache.c
+++ b/fs/bcachefs/btree_cache.c
@@ -93,7 +93,8 @@ static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp)
 	if (!b->data)
 		return -ENOMEM;
 #ifdef __KERNEL__
-	b->aux_data = vmalloc_exec(btree_aux_data_bytes(b), gfp);
+	b->aux_data = __vmalloc(btree_aux_data_bytes(b), gfp,
+				PAGE_KERNEL_EXEC);
 #else
 	b->aux_data = mmap(NULL, btree_aux_data_bytes(b),
 			   PROT_READ|PROT_WRITE|PROT_EXEC,
@@ -118,9 +119,6 @@ static struct btree *__btree_node_mem_alloc(struct bch_fs *c, gfp_t gfp)
 
 	bkey_btree_ptr_init(&b->key);
 	__six_lock_init(&b->c.lock, "b->c.lock", &bch2_btree_node_lock_key);
-#ifdef CONFIG_DEBUG_LOCK_ALLOC
-	lockdep_set_no_check_recursion(&b->c.lock.dep_map);
-#endif
 	INIT_LIST_HEAD(&b->list);
 	INIT_LIST_HEAD(&b->write_blocked);
 	b->byte_order = ilog2(btree_bytes(c));
@@ -385,14 +383,6 @@ static unsigned long bch2_btree_cache_count(struct shrinker *shrink,
 	return btree_cache_can_free(bc);
 }
 
-static void bch2_btree_cache_shrinker_to_text(struct printbuf *out, struct shrinker *shrink)
-{
-	struct bch_fs *c = container_of(shrink, struct bch_fs,
-					btree_cache.shrink);
-
-	bch2_btree_cache_to_text(out, c);
-}
-
 void bch2_fs_btree_cache_exit(struct bch_fs *c)
 {
 	struct btree_cache *bc = &c->btree_cache;
@@ -476,9 +466,8 @@ int bch2_fs_btree_cache_init(struct bch_fs *c)
 
 	bc->shrink.count_objects	= bch2_btree_cache_count;
 	bc->shrink.scan_objects		= bch2_btree_cache_scan;
-	bc->shrink.to_text		= bch2_btree_cache_shrinker_to_text;
 	bc->shrink.seeks		= 4;
-	ret = register_shrinker(&bc->shrink, "%s/btree_cache", c->name);
+	ret = register_shrinker(&bc->shrink);
 out:
 	pr_verbose_init(c->opts, "ret %i", ret);
 	return ret;
diff --git a/fs/bcachefs/btree_io.c b/fs/bcachefs/btree_io.c
index dd6b536ced6a..da55b8d9a173 100644
--- a/fs/bcachefs/btree_io.c
+++ b/fs/bcachefs/btree_io.c
@@ -33,8 +33,6 @@ void bch2_btree_node_io_unlock(struct btree *b)
 
 void bch2_btree_node_io_lock(struct btree *b)
 {
-	BUG_ON(lock_class_is_held(&bch2_btree_node_lock_key));
-
 	wait_on_bit_lock_io(&b->flags, BTREE_NODE_write_in_flight,
 			    TASK_UNINTERRUPTIBLE);
 }
@@ -53,16 +51,12 @@ void __bch2_btree_node_wait_on_write(struct btree *b)
 
 void bch2_btree_node_wait_on_read(struct btree *b)
 {
-	BUG_ON(lock_class_is_held(&bch2_btree_node_lock_key));
-
 	wait_on_bit_io(&b->flags, BTREE_NODE_read_in_flight,
 		       TASK_UNINTERRUPTIBLE);
 }
 
 void bch2_btree_node_wait_on_write(struct btree *b)
 {
-	BUG_ON(lock_class_is_held(&bch2_btree_node_lock_key));
-
 	wait_on_bit_io(&b->flags, BTREE_NODE_write_in_flight,
 		       TASK_UNINTERRUPTIBLE);
 }
@@ -1156,7 +1150,8 @@ static void btree_node_read_work(struct work_struct *work)
 		bch_info(c, "retrying read");
 		ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev);
 		rb->have_ioref		= bch2_dev_get_ioref(ca, READ);
-		bio_reset(bio, NULL, REQ_OP_READ|REQ_SYNC|REQ_META);
+		bio_reset(bio);
+		bio->bi_opf		= REQ_OP_READ|REQ_SYNC|REQ_META;
 		bio->bi_iter.bi_sector	= rb->pick.ptr.offset;
 		bio->bi_iter.bi_size	= btree_bytes(c);
 
@@ -1434,10 +1429,8 @@ static int btree_node_read_all_replicas(struct bch_fs *c, struct btree *b, bool
 
 	for (i = 0; i < ra->nr; i++) {
 		ra->buf[i] = mempool_alloc(&c->btree_bounce_pool, GFP_NOFS);
-		ra->bio[i] = bio_alloc_bioset(NULL,
-					      buf_pages(ra->buf[i], btree_bytes(c)),
-					      REQ_OP_READ|REQ_SYNC|REQ_META,
-					      GFP_NOFS,
+		ra->bio[i] = bio_alloc_bioset(GFP_NOFS, buf_pages(ra->buf[i],
+								  btree_bytes(c)),
 					      &c->btree_bio);
 	}
 
@@ -1453,6 +1446,7 @@ static int btree_node_read_all_replicas(struct bch_fs *c, struct btree *b, bool
 		rb->have_ioref		= bch2_dev_get_ioref(ca, READ);
 		rb->idx			= i;
 		rb->pick		= pick;
+		rb->bio.bi_opf		= REQ_OP_READ|REQ_SYNC|REQ_META;
 		rb->bio.bi_iter.bi_sector = pick.ptr.offset;
 		rb->bio.bi_end_io	= btree_node_read_all_replicas_endio;
 		bch2_bio_map(&rb->bio, ra->buf[i], btree_bytes(c));
@@ -1519,10 +1513,8 @@ void bch2_btree_node_read(struct bch_fs *c, struct btree *b,
 
 	ca = bch_dev_bkey_exists(c, pick.ptr.dev);
 
-	bio = bio_alloc_bioset(NULL,
-			       buf_pages(b->data, btree_bytes(c)),
-			       REQ_OP_READ|REQ_SYNC|REQ_META,
-			       GFP_NOIO,
+	bio = bio_alloc_bioset(GFP_NOIO, buf_pages(b->data,
+						   btree_bytes(c)),
 			       &c->btree_bio);
 	rb = container_of(bio, struct btree_read_bio, bio);
 	rb->c			= c;
@@ -1532,6 +1524,7 @@ void bch2_btree_node_read(struct bch_fs *c, struct btree *b,
 	rb->have_ioref		= bch2_dev_get_ioref(ca, READ);
 	rb->pick		= pick;
 	INIT_WORK(&rb->work, btree_node_read_work);
+	bio->bi_opf		= REQ_OP_READ|REQ_SYNC|REQ_META;
 	bio->bi_iter.bi_sector	= pick.ptr.offset;
 	bio->bi_end_io		= btree_node_read_endio;
 	bch2_bio_map(bio, b->data, btree_bytes(c));
@@ -1994,10 +1987,8 @@ do_write:
 
 	trace_and_count(c, btree_node_write, b, bytes_to_write, sectors_to_write);
 
-	wbio = container_of(bio_alloc_bioset(NULL,
+	wbio = container_of(bio_alloc_bioset(GFP_NOIO,
 				buf_pages(data, sectors_to_write << 9),
-				REQ_OP_WRITE|REQ_META,
-				GFP_NOIO,
 				&c->btree_bio),
 			    struct btree_write_bio, wbio.bio);
 	wbio_init(&wbio->wbio.bio);
@@ -2007,6 +1998,7 @@ do_write:
 	wbio->wbio.c			= c;
 	wbio->wbio.used_mempool		= used_mempool;
 	wbio->wbio.first_btree_write	= !b->written;
+	wbio->wbio.bio.bi_opf		= REQ_OP_WRITE|REQ_META;
 	wbio->wbio.bio.bi_end_io	= btree_node_write_endio;
 	wbio->wbio.bio.bi_private	= b;
 
diff --git a/fs/bcachefs/btree_io.h b/fs/bcachefs/btree_io.h
index 8af853642123..a24e19039ac0 100644
--- a/fs/bcachefs/btree_io.h
+++ b/fs/bcachefs/btree_io.h
@@ -114,7 +114,7 @@ static inline int bset_encrypt(struct bch_fs *c, struct bset *i, unsigned offset
 		if (ret)
 			return ret;
 
-		nonce = nonce_add(nonce, round_up(bytes, CHACHA_BLOCK_SIZE));
+		nonce = nonce_add(nonce, round_up(bytes, CHACHA20_BLOCK_SIZE));
 	}
 
 	return bch2_encrypt(c, BSET_CSUM_TYPE(i), nonce, i->_data,
diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c
index af658390bd67..176c488a25ea 100644
--- a/fs/bcachefs/btree_iter.c
+++ b/fs/bcachefs/btree_iter.c
@@ -16,7 +16,7 @@
 #include "replicas.h"
 #include "subvolume.h"
 
-#include <linux/prandom.h>
+#include <linux/random.h>
 #include <linux/prefetch.h>
 #include <trace/events/bcachefs.h>
 
@@ -2845,8 +2845,6 @@ void __bch2_trans_init(struct btree_trans *trans, struct bch_fs *c, const char *
 	struct btree_transaction_stats *s;
 	struct btree_trans *pos;
 
-	BUG_ON(lock_class_is_held(&bch2_btree_node_lock_key));
-
 	memset(trans, 0, sizeof(*trans));
 	trans->c		= c;
 	trans->fn		= fn;
diff --git a/fs/bcachefs/btree_iter.h b/fs/bcachefs/btree_iter.h
index 910f6d7bc961..7bdedf786345 100644
--- a/fs/bcachefs/btree_iter.h
+++ b/fs/bcachefs/btree_iter.h
@@ -141,13 +141,13 @@ int __must_check bch2_btree_path_traverse(struct btree_trans *,
 					  struct btree_path *, unsigned);
 struct btree_path *bch2_path_get(struct btree_trans *, enum btree_id, struct bpos,
 				 unsigned, unsigned, unsigned, unsigned long);
-inline struct bkey_s_c bch2_btree_path_peek_slot(struct btree_path *, struct bkey *);
+struct bkey_s_c bch2_btree_path_peek_slot(struct btree_path *, struct bkey *);
 
 struct bkey_i *bch2_btree_journal_peek_slot(struct btree_trans *,
 					struct btree_iter *, struct bpos);
 
-inline void bch2_btree_path_level_init(struct btree_trans *,
-				       struct btree_path *, struct btree *);
+void bch2_btree_path_level_init(struct btree_trans *,
+				struct btree_path *, struct btree *);
 
 #ifdef CONFIG_BCACHEFS_DEBUG
 void bch2_trans_verify_paths(struct btree_trans *);
diff --git a/fs/bcachefs/btree_key_cache.c b/fs/bcachefs/btree_key_cache.c
index 35e941949f49..7dc68afdb96a 100644
--- a/fs/bcachefs/btree_key_cache.c
+++ b/fs/bcachefs/btree_key_cache.c
@@ -813,7 +813,7 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 	do {
 		struct rhash_head *pos, *next;
 
-		pos = rht_ptr_rcu(rht_bucket(tbl, bc->shrink_iter));
+		pos = *rht_bucket(tbl, bc->shrink_iter);
 
 		while (!rht_is_a_nulls(pos)) {
 			next = rht_dereference_bucket_rcu(pos->next, tbl, bc->shrink_iter);
@@ -930,17 +930,8 @@ void bch2_fs_btree_key_cache_init_early(struct btree_key_cache *c)
 	INIT_LIST_HEAD(&c->freed_nonpcpu);
 }
 
-static void bch2_btree_key_cache_shrinker_to_text(struct printbuf *out, struct shrinker *shrink)
-{
-	struct btree_key_cache *bc =
-		container_of(shrink, struct btree_key_cache, shrink);
-
-	bch2_btree_key_cache_to_text(out, bc);
-}
-
 int bch2_fs_btree_key_cache_init(struct btree_key_cache *bc)
 {
-	struct bch_fs *c = container_of(bc, struct bch_fs, btree_key_cache);
 	int ret;
 
 	bc->pcpu_freed = alloc_percpu(struct btree_key_cache_freelist);
@@ -956,8 +947,7 @@ int bch2_fs_btree_key_cache_init(struct btree_key_cache *bc)
 	bc->shrink.seeks		= 1;
 	bc->shrink.count_objects	= bch2_btree_key_cache_count;
 	bc->shrink.scan_objects		= bch2_btree_key_cache_scan;
-	bc->shrink.to_text		= bch2_btree_key_cache_shrinker_to_text;
-	return register_shrinker(&bc->shrink, "%s/btree_key_cache", c->name);
+	return register_shrinker(&bc->shrink);
 }
 
 void bch2_btree_key_cache_to_text(struct printbuf *out, struct btree_key_cache *c)
diff --git a/fs/bcachefs/btree_locking.c b/fs/bcachefs/btree_locking.c
index 93a6ebed3aba..93f4eb782e7a 100644
--- a/fs/bcachefs/btree_locking.c
+++ b/fs/bcachefs/btree_locking.c
@@ -622,13 +622,6 @@ void bch2_trans_unlock(struct btree_trans *trans)
 
 	trans_for_each_path(trans, path)
 		__bch2_btree_path_unlock(trans, path);
-
-	/*
-	 * bch2_gc_btree_init_recurse() doesn't use btree iterators for walking
-	 * btree nodes, it implements its own walking:
-	 */
-	EBUG_ON(!trans->is_initial_gc &&
-		lock_class_is_held(&bch2_btree_node_lock_key));
 }
 
 bool bch2_trans_locked(struct btree_trans *trans)
diff --git a/fs/bcachefs/checksum.c b/fs/bcachefs/checksum.c
index b5850a761b91..eab1e7b51ec8 100644
--- a/fs/bcachefs/checksum.c
+++ b/fs/bcachefs/checksum.c
@@ -12,7 +12,7 @@
 #include <linux/random.h>
 #include <linux/scatterlist.h>
 #include <crypto/algapi.h>
-#include <crypto/chacha.h>
+#include <crypto/chacha20.h>
 #include <crypto/hash.h>
 #include <crypto/poly1305.h>
 #include <crypto/skcipher.h>
@@ -94,14 +94,14 @@ static void bch2_checksum_update(struct bch2_checksum_state *state, const void *
 	}
 }
 
-static inline int do_encrypt_sg(struct crypto_sync_skcipher *tfm,
+static inline int do_encrypt_sg(struct crypto_skcipher *tfm,
 				struct nonce nonce,
 				struct scatterlist *sg, size_t len)
 {
-	SYNC_SKCIPHER_REQUEST_ON_STACK(req, tfm);
+	SKCIPHER_REQUEST_ON_STACK(req, tfm);
 	int ret;
 
-	skcipher_request_set_sync_tfm(req, tfm);
+	skcipher_request_set_tfm(req, tfm);
 	skcipher_request_set_crypt(req, sg, sg, len, nonce.d);
 
 	ret = crypto_skcipher_encrypt(req);
@@ -111,7 +111,7 @@ static inline int do_encrypt_sg(struct crypto_sync_skcipher *tfm,
 	return ret;
 }
 
-static inline int do_encrypt(struct crypto_sync_skcipher *tfm,
+static inline int do_encrypt(struct crypto_skcipher *tfm,
 			      struct nonce nonce,
 			      void *buf, size_t len)
 {
@@ -155,8 +155,8 @@ static inline int do_encrypt(struct crypto_sync_skcipher *tfm,
 int bch2_chacha_encrypt_key(struct bch_key *key, struct nonce nonce,
 			    void *buf, size_t len)
 {
-	struct crypto_sync_skcipher *chacha20 =
-		crypto_alloc_sync_skcipher("chacha20", 0, 0);
+	struct crypto_skcipher *chacha20 =
+		crypto_alloc_skcipher("chacha20", 0, 0);
 	int ret;
 
 	if (!chacha20) {
@@ -164,8 +164,7 @@ int bch2_chacha_encrypt_key(struct bch_key *key, struct nonce nonce,
 		return PTR_ERR(chacha20);
 	}
 
-	ret = crypto_skcipher_setkey(&chacha20->base,
-				     (void *) key, sizeof(*key));
+	ret = crypto_skcipher_setkey(chacha20, (void *) key, sizeof(*key));
 	if (ret) {
 		pr_err("crypto_skcipher_setkey() error: %i", ret);
 		goto err;
@@ -173,7 +172,7 @@ int bch2_chacha_encrypt_key(struct bch_key *key, struct nonce nonce,
 
 	ret = do_encrypt(chacha20, nonce, buf, len);
 err:
-	crypto_free_sync_skcipher(chacha20);
+	crypto_free_skcipher(chacha20);
 	return ret;
 }
 
@@ -270,7 +269,7 @@ static struct bch_csum __bch2_checksum_bio(struct bch_fs *c, unsigned type,
 			kunmap_atomic(p);
 		}
 #else
-		__bio_for_each_bvec(bv, bio, *iter, *iter)
+		__bio_for_each_contig_segment(bv, bio, *iter, *iter)
 			bch2_checksum_update(&state, page_address(bv.bv_page) + bv.bv_offset,
 				bv.bv_len);
 #endif
@@ -293,7 +292,7 @@ static struct bch_csum __bch2_checksum_bio(struct bch_fs *c, unsigned type,
 			kunmap_atomic(p);
 		}
 #else
-		__bio_for_each_bvec(bv, bio, *iter, *iter)
+		__bio_for_each_contig_segment(bv, bio, *iter, *iter)
 			crypto_shash_update(desc,
 				page_address(bv.bv_page) + bv.bv_offset,
 				bv.bv_len);
@@ -556,7 +555,7 @@ static int bch2_alloc_ciphers(struct bch_fs *c)
 	int ret;
 
 	if (!c->chacha20)
-		c->chacha20 = crypto_alloc_sync_skcipher("chacha20", 0, 0);
+		c->chacha20 = crypto_alloc_skcipher("chacha20", 0, 0);
 	ret = PTR_ERR_OR_ZERO(c->chacha20);
 
 	if (ret) {
@@ -641,7 +640,7 @@ int bch2_enable_encryption(struct bch_fs *c, bool keyed)
 			goto err;
 	}
 
-	ret = crypto_skcipher_setkey(&c->chacha20->base,
+	ret = crypto_skcipher_setkey(c->chacha20,
 			(void *) &key.key, sizeof(key.key));
 	if (ret)
 		goto err;
@@ -669,7 +668,7 @@ 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_sync_skcipher(c->chacha20);
+		crypto_free_skcipher(c->chacha20);
 	if (!IS_ERR_OR_NULL(c->sha256))
 		crypto_free_shash(c->sha256);
 }
@@ -701,7 +700,7 @@ int bch2_fs_encryption_init(struct bch_fs *c)
 	if (ret)
 		goto out;
 
-	ret = crypto_skcipher_setkey(&c->chacha20->base,
+	ret = crypto_skcipher_setkey(c->chacha20,
 			(void *) &key.key, sizeof(key.key));
 	if (ret)
 		goto out;
diff --git a/fs/bcachefs/checksum.h b/fs/bcachefs/checksum.h
index c86c3c05d620..0fd4c19cd7b6 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/chacha.h>
+#include <crypto/chacha20.h>
 
 static inline bool bch2_checksum_mergeable(unsigned type)
 {
@@ -140,9 +140,9 @@ 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 & (CHACHA_BLOCK_SIZE - 1));
+	EBUG_ON(offset & (CHACHA20_BLOCK_SIZE - 1));
 
-	le32_add_cpu(&nonce.d[0], offset / CHACHA_BLOCK_SIZE);
+	le32_add_cpu(&nonce.d[0], offset / CHACHA20_BLOCK_SIZE);
 	return nonce;
 }
 
diff --git a/fs/bcachefs/compress.c b/fs/bcachefs/compress.c
index f692f35a6a98..691009fc2431 100644
--- a/fs/bcachefs/compress.c
+++ b/fs/bcachefs/compress.c
@@ -45,7 +45,7 @@ static bool bio_phys_contig(struct bio *bio, struct bvec_iter start)
 	struct bvec_iter iter;
 	void *expected_start = NULL;
 
-	__bio_for_each_bvec(bv, bio, iter, start) {
+	__bio_for_each_segment(bv, bio, iter, start) {
 		if (expected_start &&
 		    expected_start != page_address(bv.bv_page) + bv.bv_offset)
 			return false;
@@ -197,9 +197,9 @@ static int __bio_uncompress(struct bch_fs *c, struct bio *src,
 			goto err;
 
 		workspace = mempool_alloc(&c->decompress_workspace, GFP_NOIO);
-		ctx = zstd_init_dctx(workspace, zstd_dctx_workspace_bound());
+		ctx = ZSTD_initDCtx(workspace, ZSTD_DCtxWorkspaceBound());
 
-		ret = zstd_decompress_dctx(ctx,
+		ret = ZSTD_decompressDCtx(ctx,
 				dst_data,	dst_len,
 				src_data.b + 4, real_src_len);
 
@@ -333,8 +333,8 @@ static int attempt_compress(struct bch_fs *c,
 		return strm.total_out;
 	}
 	case BCH_COMPRESSION_TYPE_zstd: {
-		ZSTD_CCtx *ctx = zstd_init_cctx(workspace,
-			zstd_cctx_workspace_bound(&c->zstd_params.cParams));
+		ZSTD_CCtx *ctx = ZSTD_initCCtx(workspace,
+			ZSTD_CCtxWorkspaceBound(c->zstd_params.cParams));
 
 		/*
 		 * ZSTD requires that when we decompress we pass in the exact
@@ -347,11 +347,11 @@ static int attempt_compress(struct bch_fs *c,
 		 * factor (7 bytes) from the dst buffer size to account for
 		 * that.
 		 */
-		size_t len = zstd_compress_cctx(ctx,
+		size_t len = ZSTD_compressCCtx(ctx,
 				dst + 4,	dst_len - 4 - 7,
 				src,		src_len,
-				&c->zstd_params);
-		if (zstd_is_error(len))
+				c->zstd_params);
+		if (ZSTD_isError(len))
 			return 0;
 
 		*((__le32 *) dst) = cpu_to_le32(len);
@@ -546,7 +546,7 @@ static int __bch2_fs_compress_init(struct bch_fs *c, u64 features)
 {
 	size_t decompress_workspace_size = 0;
 	bool decompress_workspace_needed;
-	ZSTD_parameters params = zstd_get_params(0, c->opts.encoded_extent_max);
+	ZSTD_parameters params = ZSTD_getParams(0, c->opts.encoded_extent_max, 0);
 	struct {
 		unsigned	feature;
 		unsigned	type;
@@ -558,8 +558,8 @@ static int __bch2_fs_compress_init(struct bch_fs *c, u64 features)
 			zlib_deflate_workspacesize(MAX_WBITS, DEF_MEM_LEVEL),
 			zlib_inflate_workspacesize(), },
 		{ BCH_FEATURE_zstd, BCH_COMPRESSION_TYPE_zstd,
-			zstd_cctx_workspace_bound(&params.cParams),
-			zstd_dctx_workspace_bound() },
+			ZSTD_CCtxWorkspaceBound(params.cParams),
+			ZSTD_DCtxWorkspaceBound() },
 	}, *i;
 	int ret = 0;
 
diff --git a/fs/bcachefs/darray.h b/fs/bcachefs/darray.h
index 519ab9b96e67..6de0c1c8ae36 100644
--- a/fs/bcachefs/darray.h
+++ b/fs/bcachefs/darray.h
@@ -11,6 +11,19 @@
 #include "util.h"
 #include <linux/slab.h>
 
+static inline void * __must_check krealloc_array(void *p,
+						 size_t new_n,
+						 size_t new_size,
+						 gfp_t flags)
+{
+	size_t bytes;
+
+	if (unlikely(check_mul_overflow(new_n, new_size, &bytes)))
+		return NULL;
+
+	return krealloc(p, bytes, flags);
+}
+
 #define DARRAY(type)							\
 struct {								\
 	size_t nr, size;						\
diff --git a/fs/bcachefs/debug.c b/fs/bcachefs/debug.c
index d87131f58c80..485f8df64730 100644
--- a/fs/bcachefs/debug.c
+++ b/fs/bcachefs/debug.c
@@ -45,11 +45,11 @@ static bool bch2_btree_verify_replica(struct bch_fs *c, struct btree *b,
 	if (!bch2_dev_get_ioref(ca, READ))
 		return false;
 
-	bio = bio_alloc_bioset(ca->disk_sb.bdev,
-			       buf_pages(n_sorted, btree_bytes(c)),
-			       REQ_OP_READ|REQ_META,
-			       GFP_NOIO,
-			       &c->btree_bio);
+	bio = bio_alloc_bioset(GFP_NOIO,
+			buf_pages(n_sorted, btree_bytes(c)),
+			&c->btree_bio);
+	bio_set_dev(bio, ca->disk_sb.bdev);
+	bio->bi_opf		= REQ_OP_READ|REQ_META;
 	bio->bi_iter.bi_sector	= pick.ptr.offset;
 	bch2_bio_map(bio, n_sorted, btree_bytes(c));
 
diff --git a/fs/bcachefs/ec.c b/fs/bcachefs/ec.c
index d3fa2d7ae1a2..ab8fdbb61e1c 100644
--- a/fs/bcachefs/ec.c
+++ b/fs/bcachefs/ec.c
@@ -407,16 +407,13 @@ static void ec_block_io(struct bch_fs *c, struct ec_stripe_buf *buf,
 	this_cpu_add(ca->io_done->sectors[rw][data_type], buf->size);
 
 	while (offset < bytes) {
-		unsigned nr_iovecs = min_t(size_t, BIO_MAX_VECS,
+		unsigned nr_iovecs = min_t(size_t, BIO_MAX_PAGES,
 					   DIV_ROUND_UP(bytes, PAGE_SIZE));
 		unsigned b = min_t(size_t, bytes - offset,
 				   nr_iovecs << PAGE_SHIFT);
 		struct ec_bio *ec_bio;
 
-		ec_bio = container_of(bio_alloc_bioset(ca->disk_sb.bdev,
-						       nr_iovecs,
-						       rw,
-						       GFP_KERNEL,
+		ec_bio = container_of(bio_alloc_bioset(GFP_KERNEL, nr_iovecs,
 						       &c->ec_bioset),
 				      struct ec_bio, bio);
 
@@ -424,6 +421,9 @@ static void ec_block_io(struct bch_fs *c, struct ec_stripe_buf *buf,
 		ec_bio->buf			= buf;
 		ec_bio->idx			= idx;
 
+		bio_set_dev(&ec_bio->bio, ca->disk_sb.bdev);
+		bio_set_op_attrs(&ec_bio->bio, rw, 0);
+
 		ec_bio->bio.bi_iter.bi_sector	= ptr->offset + buf->offset + (offset >> 9);
 		ec_bio->bio.bi_end_io		= ec_block_endio;
 		ec_bio->bio.bi_private		= cl;
diff --git a/fs/bcachefs/errcode.c b/fs/bcachefs/errcode.c
index cc9ce0be356e..0db8aa75ee28 100644
--- a/fs/bcachefs/errcode.c
+++ b/fs/bcachefs/errcode.c
@@ -3,8 +3,6 @@
 #include "bcachefs.h"
 #include "errcode.h"
 
-#include <linux/errname.h>
-
 static const char * const bch2_errcode_strs[] = {
 #define x(class, err) [BCH_ERR_##err - BCH_ERR_START] = #err,
 	BCH_ERRCODES()
@@ -20,6 +18,223 @@ static unsigned bch2_errcode_parents[] = {
 #undef x
 };
 
+/*
+ * Ensure these tables do not accidentally become gigantic if some
+ * huge errno makes it in. On most architectures, the first table will
+ * only have about 140 entries, but mips and parisc have more sparsely
+ * allocated errnos (with EHWPOISON = 257 on parisc, and EDQUOT = 1133
+ * on mips), so this wastes a bit of space on those - though we
+ * special case the EDQUOT case.
+ */
+#define E(err) [err + BUILD_BUG_ON_ZERO(err <= 0 || err > 300)] = "-" #err
+static const char *names_0[] = {
+	E(E2BIG),
+	E(EACCES),
+	E(EADDRINUSE),
+	E(EADDRNOTAVAIL),
+	E(EADV),
+	E(EAFNOSUPPORT),
+	E(EALREADY),
+	E(EBADE),
+	E(EBADF),
+	E(EBADFD),
+	E(EBADMSG),
+	E(EBADR),
+	E(EBADRQC),
+	E(EBADSLT),
+	E(EBFONT),
+	E(EBUSY),
+#ifdef ECANCELLED
+	E(ECANCELLED),
+#endif
+	E(ECHILD),
+	E(ECHRNG),
+	E(ECOMM),
+	E(ECONNABORTED),
+	E(ECONNRESET),
+	E(EDEADLOCK),
+	E(EDESTADDRREQ),
+	E(EDOM),
+	E(EDOTDOT),
+#ifndef CONFIG_MIPS
+	E(EDQUOT),
+#endif
+	E(EEXIST),
+	E(EFAULT),
+	E(EFBIG),
+	E(EHOSTDOWN),
+	E(EHOSTUNREACH),
+	E(EHWPOISON),
+	E(EIDRM),
+	E(EILSEQ),
+#ifdef EINIT
+	E(EINIT),
+#endif
+	E(EINPROGRESS),
+	E(EINTR),
+	E(EINVAL),
+	E(EIO),
+	E(EISCONN),
+	E(EISDIR),
+	E(EISNAM),
+	E(EKEYEXPIRED),
+	E(EKEYREJECTED),
+	E(EKEYREVOKED),
+	E(EL2HLT),
+	E(EL2NSYNC),
+	E(EL3HLT),
+	E(EL3RST),
+	E(ELIBACC),
+	E(ELIBBAD),
+	E(ELIBEXEC),
+	E(ELIBMAX),
+	E(ELIBSCN),
+	E(ELNRNG),
+	E(ELOOP),
+	E(EMEDIUMTYPE),
+	E(EMFILE),
+	E(EMLINK),
+	E(EMSGSIZE),
+	E(EMULTIHOP),
+	E(ENAMETOOLONG),
+	E(ENAVAIL),
+	E(ENETDOWN),
+	E(ENETRESET),
+	E(ENETUNREACH),
+	E(ENFILE),
+	E(ENOANO),
+	E(ENOBUFS),
+	E(ENOCSI),
+	E(ENODATA),
+	E(ENODEV),
+	E(ENOENT),
+	E(ENOEXEC),
+	E(ENOKEY),
+	E(ENOLCK),
+	E(ENOLINK),
+	E(ENOMEDIUM),
+	E(ENOMEM),
+	E(ENOMSG),
+	E(ENONET),
+	E(ENOPKG),
+	E(ENOPROTOOPT),
+	E(ENOSPC),
+	E(ENOSR),
+	E(ENOSTR),
+#ifdef ENOSYM
+	E(ENOSYM),
+#endif
+	E(ENOSYS),
+	E(ENOTBLK),
+	E(ENOTCONN),
+	E(ENOTDIR),
+	E(ENOTEMPTY),
+	E(ENOTNAM),
+	E(ENOTRECOVERABLE),
+	E(ENOTSOCK),
+	E(ENOTTY),
+	E(ENOTUNIQ),
+	E(ENXIO),
+	E(EOPNOTSUPP),
+	E(EOVERFLOW),
+	E(EOWNERDEAD),
+	E(EPERM),
+	E(EPFNOSUPPORT),
+	E(EPIPE),
+#ifdef EPROCLIM
+	E(EPROCLIM),
+#endif
+	E(EPROTO),
+	E(EPROTONOSUPPORT),
+	E(EPROTOTYPE),
+	E(ERANGE),
+	E(EREMCHG),
+#ifdef EREMDEV
+	E(EREMDEV),
+#endif
+	E(EREMOTE),
+	E(EREMOTEIO),
+#ifdef EREMOTERELEASE
+	E(EREMOTERELEASE),
+#endif
+	E(ERESTART),
+	E(ERFKILL),
+	E(EROFS),
+#ifdef ERREMOTE
+	E(ERREMOTE),
+#endif
+	E(ESHUTDOWN),
+	E(ESOCKTNOSUPPORT),
+	E(ESPIPE),
+	E(ESRCH),
+	E(ESRMNT),
+	E(ESTALE),
+	E(ESTRPIPE),
+	E(ETIME),
+	E(ETIMEDOUT),
+	E(ETOOMANYREFS),
+	E(ETXTBSY),
+	E(EUCLEAN),
+	E(EUNATCH),
+	E(EUSERS),
+	E(EXDEV),
+	E(EXFULL),
+
+	E(ECANCELED), /* ECANCELLED */
+	E(EAGAIN), /* EWOULDBLOCK */
+	E(ECONNREFUSED), /* EREFUSED */
+	E(EDEADLK), /* EDEADLOCK */
+};
+#undef E
+
+#define E(err) [err - 512 + BUILD_BUG_ON_ZERO(err < 512 || err > 550)] = "-" #err
+static const char *names_512[] = {
+	E(ERESTARTSYS),
+	E(ERESTARTNOINTR),
+	E(ERESTARTNOHAND),
+	E(ENOIOCTLCMD),
+	E(ERESTART_RESTARTBLOCK),
+	E(EPROBE_DEFER),
+	E(EOPENSTALE),
+
+	E(EBADHANDLE),
+	E(ENOTSYNC),
+	E(EBADCOOKIE),
+	E(ENOTSUPP),
+	E(ETOOSMALL),
+	E(ESERVERFAULT),
+	E(EBADTYPE),
+	E(EJUKEBOX),
+	E(EIOCBQUEUED),
+	E(ERECALLCONFLICT),
+};
+#undef E
+
+static const char *__errname(unsigned err)
+{
+	if (err < ARRAY_SIZE(names_0))
+		return names_0[err];
+	if (err >= 512 && err - 512 < ARRAY_SIZE(names_512))
+		return names_512[err - 512];
+	/* But why? */
+	if (IS_ENABLED(CONFIG_MIPS) && err == EDQUOT) /* 1133 */
+		return "-EDQUOT";
+	return NULL;
+}
+
+/*
+ * errname(EIO) -> "EIO"
+ * errname(-EIO) -> "-EIO"
+ */
+static const char *errname(int err)
+{
+	const char *name = __errname(abs(err));
+	if (!name)
+		return NULL;
+
+	return err > 0 ? name + 1 : name;
+}
+
 const char *bch2_err_str(int err)
 {
 	const char *errstr;
diff --git a/fs/bcachefs/fs-io.c b/fs/bcachefs/fs-io.c
index 02ef3430a30b..70374fa463b4 100644
--- a/fs/bcachefs/fs-io.c
+++ b/fs/bcachefs/fs-io.c
@@ -35,15 +35,6 @@
 #include <trace/events/bcachefs.h>
 #include <trace/events/writeback.h>
 
-static inline bool bio_full(struct bio *bio, unsigned len)
-{
-	if (bio->bi_vcnt >= bio->bi_max_vecs)
-		return true;
-	if (bio->bi_iter.bi_size > UINT_MAX - len)
-		return true;
-	return false;
-}
-
 static inline struct address_space *faults_disabled_mapping(void)
 {
 	return (void *) (((unsigned long) current->faults_disabled_mapping) & ~1UL);
@@ -297,13 +288,28 @@ static inline struct bch_page_state *bch2_page_state(struct page *page)
 /* for newly allocated pages: */
 static void __bch2_page_state_release(struct page *page)
 {
-	kfree(detach_page_private(page));
+	struct bch_page_state *s = __bch2_page_state(page);
+
+	if (!s)
+		return;
+
+	ClearPagePrivate(page);
+	set_page_private(page, 0);
+	put_page(page);
+	kfree(s);
 }
 
 static void bch2_page_state_release(struct page *page)
 {
-	EBUG_ON(!PageLocked(page));
-	__bch2_page_state_release(page);
+	struct bch_page_state *s = bch2_page_state(page);
+
+	if (!s)
+		return;
+
+	ClearPagePrivate(page);
+	set_page_private(page, 0);
+	put_page(page);
+	kfree(s);
 }
 
 /* for newly allocated pages: */
@@ -317,7 +323,13 @@ static struct bch_page_state *__bch2_page_state_create(struct page *page,
 		return NULL;
 
 	spin_lock_init(&s->lock);
-	attach_page_private(page, s);
+	/*
+	 * migrate_page_move_mapping() assumes that pages with private data
+	 * have their count elevated by 1.
+	 */
+	get_page(page);
+	set_page_private(page, (unsigned long) s);
+	SetPagePrivate(page);
 	return s;
 }
 
@@ -434,20 +446,22 @@ static void mark_pagecache_unallocated(struct bch_inode_info *inode,
 {
 	pgoff_t index = start >> PAGE_SECTORS_SHIFT;
 	pgoff_t end_index = (end - 1) >> PAGE_SECTORS_SHIFT;
-	struct folio_batch fbatch;
-	unsigned i, j;
+	struct pagevec pvec;
 
 	if (end <= start)
 		return;
 
-	folio_batch_init(&fbatch);
+	pagevec_init(&pvec);
 
-	while (filemap_get_folios(inode->v.i_mapping,
-				  &index, end_index, &fbatch)) {
-		for (i = 0; i < folio_batch_count(&fbatch); i++) {
-			struct folio *folio = fbatch.folios[i];
-			u64 pg_start = folio->index << PAGE_SECTORS_SHIFT;
-			u64 pg_end = (folio->index + 1) << PAGE_SECTORS_SHIFT;
+	do {
+		unsigned nr_pages, i, j;
+
+		nr_pages = pagevec_lookup_range(&pvec, inode->v.i_mapping,
+						&index, end_index);
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+			u64 pg_start = page->index << PAGE_SECTORS_SHIFT;
+			u64 pg_end = (page->index + 1) << PAGE_SECTORS_SHIFT;
 			unsigned pg_offset = max(start, pg_start) - pg_start;
 			unsigned pg_len = min(end, pg_end) - pg_offset - pg_start;
 			struct bch_page_state *s;
@@ -456,8 +470,8 @@ static void mark_pagecache_unallocated(struct bch_inode_info *inode,
 			BUG_ON(pg_offset >= PAGE_SECTORS);
 			BUG_ON(pg_offset + pg_len > PAGE_SECTORS);
 
-			folio_lock(folio);
-			s = bch2_page_state(&folio->page);
+			lock_page(page);
+			s = bch2_page_state(page);
 
 			if (s) {
 				spin_lock(&s->lock);
@@ -466,11 +480,10 @@ static void mark_pagecache_unallocated(struct bch_inode_info *inode,
 				spin_unlock(&s->lock);
 			}
 
-			folio_unlock(folio);
+			unlock_page(page);
 		}
-		folio_batch_release(&fbatch);
-		cond_resched();
-	}
+		pagevec_release(&pvec);
+	} while (index <= end_index);
 }
 
 static void mark_pagecache_reserved(struct bch_inode_info *inode,
@@ -479,21 +492,23 @@ static void mark_pagecache_reserved(struct bch_inode_info *inode,
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
 	pgoff_t index = start >> PAGE_SECTORS_SHIFT;
 	pgoff_t end_index = (end - 1) >> PAGE_SECTORS_SHIFT;
-	struct folio_batch fbatch;
+	struct pagevec pvec;
 	s64 i_sectors_delta = 0;
-	unsigned i, j;
 
 	if (end <= start)
 		return;
 
-	folio_batch_init(&fbatch);
+	pagevec_init(&pvec);
+
+	do {
+		unsigned nr_pages, i, j;
 
-	while (filemap_get_folios(inode->v.i_mapping,
-				  &index, end_index, &fbatch)) {
-		for (i = 0; i < folio_batch_count(&fbatch); i++) {
-			struct folio *folio = fbatch.folios[i];
-			u64 pg_start = folio->index << PAGE_SECTORS_SHIFT;
-			u64 pg_end = (folio->index + 1) << PAGE_SECTORS_SHIFT;
+		nr_pages = pagevec_lookup_range(&pvec, inode->v.i_mapping,
+						&index, end_index);
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+			u64 pg_start = page->index << PAGE_SECTORS_SHIFT;
+			u64 pg_end = (page->index + 1) << PAGE_SECTORS_SHIFT;
 			unsigned pg_offset = max(start, pg_start) - pg_start;
 			unsigned pg_len = min(end, pg_end) - pg_offset - pg_start;
 			struct bch_page_state *s;
@@ -502,8 +517,8 @@ static void mark_pagecache_reserved(struct bch_inode_info *inode,
 			BUG_ON(pg_offset >= PAGE_SECTORS);
 			BUG_ON(pg_offset + pg_len > PAGE_SECTORS);
 
-			folio_lock(folio);
-			s = bch2_page_state(&folio->page);
+			lock_page(page);
+			s = bch2_page_state(page);
 
 			if (s) {
 				spin_lock(&s->lock);
@@ -522,11 +537,10 @@ static void mark_pagecache_reserved(struct bch_inode_info *inode,
 				spin_unlock(&s->lock);
 			}
 
-			folio_unlock(folio);
+			unlock_page(page);
 		}
-		folio_batch_release(&fbatch);
-		cond_resched();
-	}
+		pagevec_release(&pvec);
+	} while (index <= end_index);
 
 	i_sectors_acct(c, inode, NULL, i_sectors_delta);
 }
@@ -840,31 +854,62 @@ out:
 	return ret;
 }
 
-void bch2_invalidate_folio(struct folio *folio, size_t offset, size_t length)
+void bch2_invalidatepage(struct page *page, unsigned int offset,
+			 unsigned int length)
 {
-	if (offset || length < folio_size(folio))
+	if (offset || length < PAGE_SIZE)
 		return;
 
-	bch2_clear_page_bits(&folio->page);
+	bch2_clear_page_bits(page);
+}
+
+int bch2_releasepage(struct page *page, gfp_t gfp_mask)
+{
+	if (PageDirty(page))
+		return 0;
+
+	bch2_clear_page_bits(page);
+	return 1;
 }
 
-bool bch2_release_folio(struct folio *folio, gfp_t gfp_mask)
+#ifdef CONFIG_MIGRATION
+int bch2_migrate_page(struct address_space *mapping, struct page *newpage,
+		      struct page *page, enum migrate_mode mode)
 {
-	if (folio_test_dirty(folio) || folio_test_writeback(folio))
-		return false;
+	int ret;
+
+	EBUG_ON(!PageLocked(page));
+	EBUG_ON(!PageLocked(newpage));
+
+	ret = migrate_page_move_mapping(mapping, newpage, page, NULL, mode, 0);
+	if (ret != MIGRATEPAGE_SUCCESS)
+		return ret;
+
+	if (PagePrivate(page)) {
+		ClearPagePrivate(page);
+		get_page(newpage);
+		set_page_private(newpage, page_private(page));
+		set_page_private(page, 0);
+		put_page(page);
+		SetPagePrivate(newpage);
+	}
 
-	bch2_clear_page_bits(&folio->page);
-	return true;
+	if (mode != MIGRATE_SYNC_NO_COPY)
+		migrate_page_copy(newpage, page);
+	else
+		migrate_page_states(newpage, page);
+	return MIGRATEPAGE_SUCCESS;
 }
+#endif
 
 /* readpage(s): */
 
 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, iter) {
+	bio_for_each_segment_all(bv, bio, i) {
 		struct page *page = bv->bv_page;
 
 		if (!bio->bi_status) {
@@ -883,29 +928,31 @@ struct readpages_iter {
 	struct address_space	*mapping;
 	struct page		**pages;
 	unsigned		nr_pages;
+	unsigned		nr_added;
 	unsigned		idx;
 	pgoff_t			offset;
 };
 
 static int readpages_iter_init(struct readpages_iter *iter,
-			       struct readahead_control *ractl)
+			       struct address_space *mapping,
+			       struct list_head *pages, unsigned nr_pages)
 {
-	unsigned i, nr_pages = readahead_count(ractl);
-
 	memset(iter, 0, sizeof(*iter));
 
-	iter->mapping	= ractl->mapping;
-	iter->offset	= readahead_index(ractl);
-	iter->nr_pages	= nr_pages;
+	iter->mapping	= mapping;
+	iter->offset	= list_last_entry(pages, struct page, lru)->index;
 
 	iter->pages = kmalloc_array(nr_pages, sizeof(struct page *), GFP_NOFS);
 	if (!iter->pages)
 		return -ENOMEM;
 
-	nr_pages = __readahead_batch(ractl, iter->pages, nr_pages);
-	for (i = 0; i < nr_pages; i++) {
-		__bch2_page_state_create(iter->pages[i], __GFP_NOFAIL);
-		put_page(iter->pages[i]);
+	while (!list_empty(pages)) {
+		struct page *page = list_last_entry(pages, struct page, lru);
+
+		__bch2_page_state_create(page, __GFP_NOFAIL);
+
+		iter->pages[iter->nr_pages++] = page;
+		list_del(&page->lru);
 	}
 
 	return 0;
@@ -913,9 +960,41 @@ static int readpages_iter_init(struct readpages_iter *iter,
 
 static inline struct page *readpage_iter_next(struct readpages_iter *iter)
 {
-	if (iter->idx >= iter->nr_pages)
-		return NULL;
+	struct page *page;
+	unsigned i;
+	int ret;
+
+	BUG_ON(iter->idx > iter->nr_added);
+	BUG_ON(iter->nr_added > iter->nr_pages);
+
+	if (iter->idx < iter->nr_added)
+		goto out;
+
+	while (1) {
+		if (iter->idx == iter->nr_pages)
+			return NULL;
+
+		ret = add_to_page_cache_lru_vec(iter->mapping,
+				iter->pages	+ iter->nr_added,
+				iter->nr_pages	- iter->nr_added,
+				iter->offset	+ iter->nr_added,
+				GFP_NOFS);
+		if (ret > 0)
+			break;
+
+		page = iter->pages[iter->nr_added];
+		iter->idx++;
+		iter->nr_added++;
 
+		__bch2_page_state_release(page);
+		put_page(page);
+	}
+
+	iter->nr_added += ret;
+
+	for (i = iter->idx; i < iter->nr_added; i++)
+		put_page(iter->pages[i]);
+out:
 	EBUG_ON(iter->pages[iter->idx]->index != iter->offset + iter->idx);
 
 	return iter->pages[iter->idx];
@@ -953,8 +1032,11 @@ static void readpage_bio_extend(struct readpages_iter *iter,
 			if (!get_more)
 				break;
 
-			page = xa_load(&iter->mapping->i_pages, page_offset);
-			if (page && !xa_is_value(page))
+			rcu_read_lock();
+			page = radix_tree_lookup(&iter->mapping->i_pages, page_offset);
+			rcu_read_unlock();
+
+			if (page && !radix_tree_exceptional_entry(page))
 				break;
 
 			page = __page_cache_alloc(readahead_gfp_mask(iter->mapping));
@@ -1087,9 +1169,10 @@ err:
 	bch2_bkey_buf_exit(&sk, c);
 }
 
-void bch2_readahead(struct readahead_control *ractl)
+int bch2_readpages(struct file *file, struct address_space *mapping,
+		   struct list_head *pages, unsigned nr_pages)
 {
-	struct bch_inode_info *inode = to_bch_ei(ractl->mapping->host);
+	struct bch_inode_info *inode = to_bch_ei(mapping->host);
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
 	struct bch_io_opts opts = io_opts(c, &inode->ei_inode);
 	struct btree_trans trans;
@@ -1097,7 +1180,7 @@ void bch2_readahead(struct readahead_control *ractl)
 	struct readpages_iter readpages_iter;
 	int ret;
 
-	ret = readpages_iter_init(&readpages_iter, ractl);
+	ret = readpages_iter_init(&readpages_iter, mapping, pages, nr_pages);
 	BUG_ON(ret);
 
 	bch2_trans_init(&trans, c, 0, 0);
@@ -1109,14 +1192,14 @@ void bch2_readahead(struct readahead_control *ractl)
 		unsigned n = min_t(unsigned,
 				   readpages_iter.nr_pages -
 				   readpages_iter.idx,
-				   BIO_MAX_VECS);
+				   BIO_MAX_PAGES);
 		struct bch_read_bio *rbio =
-			rbio_init(bio_alloc_bioset(NULL, n, REQ_OP_READ,
-						   GFP_NOFS, &c->bio_read),
+			rbio_init(bio_alloc_bioset(GFP_NOFS, n, &c->bio_read),
 				  opts);
 
 		readpages_iter.idx++;
 
+		bio_set_op_attrs(&rbio->bio, REQ_OP_READ, 0);
 		rbio->bio.bi_iter.bi_sector = (sector_t) index << PAGE_SECTORS_SHIFT;
 		rbio->bio.bi_end_io = bch2_readpages_end_io;
 		BUG_ON(!bio_add_page(&rbio->bio, page, PAGE_SIZE, 0));
@@ -1129,6 +1212,8 @@ void bch2_readahead(struct readahead_control *ractl)
 
 	bch2_trans_exit(&trans);
 	kfree(readpages_iter.pages);
+
+	return 0;
 }
 
 static void __bchfs_readpage(struct bch_fs *c, struct bch_read_bio *rbio,
@@ -1148,6 +1233,20 @@ static void __bchfs_readpage(struct bch_fs *c, struct bch_read_bio *rbio,
 	bch2_trans_exit(&trans);
 }
 
+int bch2_readpage(struct file *file, struct page *page)
+{
+	struct bch_inode_info *inode = to_bch_ei(page->mapping->host);
+	struct bch_fs *c = inode->v.i_sb->s_fs_info;
+	struct bch_io_opts opts = io_opts(c, &inode->ei_inode);
+	struct bch_read_bio *rbio;
+
+	rbio = rbio_init(bio_alloc_bioset(GFP_NOFS, 1, &c->bio_read), opts);
+	rbio->bio.bi_end_io = bch2_readpages_end_io;
+
+	__bchfs_readpage(c, rbio, inode_inum(inode), page);
+	return 0;
+}
+
 static void bch2_read_single_page_end_io(struct bio *bio)
 {
 	complete(bio->bi_private);
@@ -1162,7 +1261,7 @@ static int bch2_read_single_page(struct page *page,
 	int ret;
 	DECLARE_COMPLETION_ONSTACK(done);
 
-	rbio = rbio_init(bio_alloc_bioset(NULL, 1, REQ_OP_READ, GFP_NOFS, &c->bio_read),
+	rbio = rbio_init(bio_alloc_bioset(GFP_NOFS, 1, &c->bio_read),
 			 io_opts(c, &inode->ei_inode));
 	rbio->bio.bi_private = &done;
 	rbio->bio.bi_end_io = bch2_read_single_page_end_io;
@@ -1180,16 +1279,6 @@ static int bch2_read_single_page(struct page *page,
 	return 0;
 }
 
-int bch2_read_folio(struct file *file, struct folio *folio)
-{
-	struct page *page = &folio->page;
-	int ret;
-
-	ret = bch2_read_single_page(page, page->mapping);
-	folio_unlock(folio);
-	return bch2_err_class(ret);
-}
-
 /* writepages: */
 
 struct bch_writepage_state {
@@ -1219,35 +1308,34 @@ 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;
+	unsigned i, j;
 
 	if (io->op.error) {
 		set_bit(EI_INODE_ERROR, &io->inode->ei_flags);
 
-		bio_for_each_segment_all(bvec, bio, iter) {
+		bio_for_each_segment_all(bvec, bio, i) {
 			struct bch_page_state *s;
 
 			SetPageError(bvec->bv_page);
-			mapping_set_error(bvec->bv_page->mapping, -EIO);
+			mapping_set_error(io->inode->v.i_mapping, -EIO);
 
 			s = __bch2_page_state(bvec->bv_page);
 			spin_lock(&s->lock);
-			for (i = 0; i < PAGE_SECTORS; i++)
-				s->s[i].nr_replicas = 0;
+			for (j = 0; j < PAGE_SECTORS; j++)
+				s->s[j].nr_replicas = 0;
 			spin_unlock(&s->lock);
 		}
 	}
 
 	if (io->op.flags & BCH_WRITE_WROTE_DATA_INLINE) {
-		bio_for_each_segment_all(bvec, bio, iter) {
+		bio_for_each_segment_all(bvec, bio, i) {
 			struct bch_page_state *s;
 
 			s = __bch2_page_state(bvec->bv_page);
 			spin_lock(&s->lock);
-			for (i = 0; i < PAGE_SECTORS; i++)
-				s->s[i].nr_replicas = 0;
+			for (j = 0; j < PAGE_SECTORS; j++)
+				s->s[j].nr_replicas = 0;
 			spin_unlock(&s->lock);
 		}
 	}
@@ -1271,7 +1359,7 @@ 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, iter) {
+	bio_for_each_segment_all(bvec, bio, i) {
 		struct bch_page_state *s = __bch2_page_state(bvec->bv_page);
 
 		if (atomic_dec_and_test(&s->write_count))
@@ -1303,9 +1391,7 @@ static void bch2_writepage_io_alloc(struct bch_fs *c,
 {
 	struct bch_write_op *op;
 
-	w->io = container_of(bio_alloc_bioset(NULL, BIO_MAX_VECS,
-					      REQ_OP_WRITE,
-					      GFP_NOFS,
+	w->io = container_of(bio_alloc_bioset(GFP_NOFS, BIO_MAX_PAGES,
 					      &c->writepage_bioset),
 			     struct bch_writepage_io, op.wbio.bio);
 
@@ -1425,9 +1511,9 @@ do_io:
 
 		if (w->io &&
 		    (w->io->op.res.nr_replicas != nr_replicas_this_write ||
-		     bio_full(&w->io->op.wbio.bio, PAGE_SIZE) ||
+		     bio_full(&w->io->op.wbio.bio) ||
 		     w->io->op.wbio.bio.bi_iter.bi_size + (sectors << 9) >=
-		     (BIO_MAX_VECS * PAGE_SIZE) ||
+		     (BIO_MAX_PAGES * PAGE_SIZE) ||
 		     bio_end_sector(&w->io->op.wbio.bio) != sector))
 			bch2_writepage_do_io(w);
 
@@ -1477,7 +1563,7 @@ int bch2_writepages(struct address_space *mapping, struct writeback_control *wbc
 /* buffered writes: */
 
 int bch2_write_begin(struct file *file, struct address_space *mapping,
-		     loff_t pos, unsigned len,
+		     loff_t pos, unsigned len, unsigned flags,
 		     struct page **pagep, void **fsdata)
 {
 	struct bch_inode_info *inode = to_bch_ei(mapping->host);
@@ -1497,7 +1583,7 @@ int bch2_write_begin(struct file *file, struct address_space *mapping,
 
 	bch2_pagecache_add_get(&inode->ei_pagecache_lock);
 
-	page = grab_cache_page_write_begin(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		goto err_unlock;
 
@@ -1628,7 +1714,7 @@ static int __bch2_buffered_write(struct bch_inode_info *inode,
 	bch2_page_reservation_init(c, inode, &res);
 
 	for (i = 0; i < nr_pages; i++) {
-		pages[i] = grab_cache_page_write_begin(mapping, index + i);
+		pages[i] = grab_cache_page_write_begin(mapping, index + i, 0);
 		if (!pages[i]) {
 			nr_pages = i;
 			if (!i) {
@@ -1689,8 +1775,8 @@ static int __bch2_buffered_write(struct bch_inode_info *inode,
 		unsigned pg_offset = (offset + copied) & (PAGE_SIZE - 1);
 		unsigned pg_len = min_t(unsigned, len - copied,
 					PAGE_SIZE - pg_offset);
-		unsigned pg_copied = copy_page_from_iter_atomic(page,
-						pg_offset, pg_len,iter);
+		unsigned pg_copied = iov_iter_copy_from_user_atomic(page,
+						iter, pg_offset, pg_len);
 
 		if (!pg_copied)
 			break;
@@ -1703,6 +1789,7 @@ static int __bch2_buffered_write(struct bch_inode_info *inode,
 		}
 
 		flush_dcache_page(page);
+		iov_iter_advance(iter, pg_copied);
 		copied += pg_copied;
 
 		if (pg_copied != pg_len)
@@ -1772,11 +1859,11 @@ again:
 		 * to check that the address is actually valid, when atomic
 		 * usercopies are used, below.
 		 */
-		if (unlikely(fault_in_iov_iter_readable(iter, bytes))) {
+		if (unlikely(iov_iter_fault_in_readable(iter, bytes))) {
 			bytes = min_t(unsigned long, iov_iter_count(iter),
 				      PAGE_SIZE - offset);
 
-			if (unlikely(fault_in_iov_iter_readable(iter, bytes))) {
+			if (unlikely(iov_iter_fault_in_readable(iter, bytes))) {
 				ret = -EFAULT;
 				break;
 			}
@@ -1820,6 +1907,18 @@ again:
 
 /* O_DIRECT reads */
 
+static void bio_release_pages(struct bio *bio, bool mark_dirty)
+{
+	struct bio_vec *bvec;
+	unsigned i;
+
+	bio_for_each_segment_all(bvec, bio, i) {
+		if (mark_dirty && !PageCompound(bvec->bv_page))
+			set_page_dirty_lock(bvec->bv_page);
+		put_page(bvec->bv_page);
+	}
+}
+
 static void bio_check_or_release(struct bio *bio, bool check_dirty)
 {
 	if (check_dirty) {
@@ -1834,7 +1933,7 @@ static void bch2_dio_read_complete(struct closure *cl)
 {
 	struct dio_read *dio = container_of(cl, struct dio_read, cl);
 
-	dio->req->ki_complete(dio->req, dio->ret);
+	dio->req->ki_complete(dio->req, dio->ret, 0);
 	bio_check_or_release(&dio->rbio.bio, dio->should_dirty);
 }
 
@@ -1882,10 +1981,8 @@ static int bch2_direct_IO_read(struct kiocb *req, struct iov_iter *iter)
 	shorten = iov_iter_count(iter) - round_up(ret, block_bytes(c));
 	iter->count -= shorten;
 
-	bio = bio_alloc_bioset(NULL,
-			       bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS),
-			       REQ_OP_READ,
-			       GFP_KERNEL,
+	bio = bio_alloc_bioset(GFP_KERNEL,
+			       iov_iter_npages(iter, BIO_MAX_PAGES),
 			       &c->dio_read_bioset);
 
 	bio->bi_end_io = bch2_direct_IO_read_endio;
@@ -1919,10 +2016,8 @@ static int bch2_direct_IO_read(struct kiocb *req, struct iov_iter *iter)
 
 	goto start;
 	while (iter->count) {
-		bio = bio_alloc_bioset(NULL,
-				       bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS),
-				       REQ_OP_READ,
-				       GFP_KERNEL,
+		bio = bio_alloc_bioset(GFP_KERNEL,
+				       iov_iter_npages(iter, BIO_MAX_PAGES),
 				       &c->bio_read);
 		bio->bi_end_io		= bch2_direct_IO_read_split_endio;
 start:
@@ -2055,9 +2150,8 @@ static long bch2_dio_write_loop(struct dio_write *dio)
 	struct bch_inode_info *inode = file_bch_inode(req->ki_filp);
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
 	struct bio *bio = &dio->op.wbio.bio;
-	struct bvec_iter_all iter;
 	struct bio_vec *bv;
-	unsigned unaligned, iter_count;
+	unsigned i, unaligned, iter_count;
 	bool sync = dio->sync, dropped_locks;
 	long ret;
 
@@ -2068,7 +2162,7 @@ static long bch2_dio_write_loop(struct dio_write *dio)
 		iter_count = dio->iter.count;
 
 		if (kthread && dio->mm)
-			kthread_use_mm(dio->mm);
+			use_mm(dio->mm);
 		BUG_ON(current->faults_disabled_mapping);
 		current->faults_disabled_mapping = mapping;
 
@@ -2078,7 +2172,7 @@ static long bch2_dio_write_loop(struct dio_write *dio)
 
 		current->faults_disabled_mapping = NULL;
 		if (kthread && dio->mm)
-			kthread_unuse_mm(dio->mm);
+			unuse_mm(dio->mm);
 
 		/*
 		 * If the fault handler returned an error but also signalled
@@ -2175,9 +2269,8 @@ loop:
 			i_size_write(&inode->v, req->ki_pos);
 		spin_unlock(&inode->v.i_lock);
 
-		if (likely(!bio_flagged(bio, BIO_NO_PAGE_REF)))
-			bio_for_each_segment_all(bv, bio, iter)
-				put_page(bv->bv_page);
+		bio_for_each_segment_all(bv, bio, i)
+			put_page(bv->bv_page);
 		bio->bi_vcnt = 0;
 
 		if (dio->op.error) {
@@ -2188,7 +2281,7 @@ loop:
 		if (!dio->iter.count)
 			break;
 
-		bio_reset(bio, NULL, REQ_OP_WRITE);
+		bio_reset(bio);
 		reinit_completion(&dio->done);
 	}
 
@@ -2200,9 +2293,8 @@ err:
 	if (dio->free_iov)
 		kfree(dio->iter.iov);
 
-	if (likely(!bio_flagged(bio, BIO_NO_PAGE_REF)))
-		bio_for_each_segment_all(bv, bio, iter)
-			put_page(bv->bv_page);
+	bio_for_each_segment_all(bv, bio, i)
+		put_page(bv->bv_page);
 	bio_put(bio);
 
 	/* inode->i_dio_count is our ref on inode and thus bch_fs */
@@ -2212,7 +2304,7 @@ err:
 		ret = bch2_err_class(ret);
 
 	if (!sync) {
-		req->ki_complete(req, ret);
+		req->ki_complete(req, ret, 0);
 		ret = -EIOCBQUEUED;
 	}
 	return ret;
@@ -2271,10 +2363,8 @@ ssize_t bch2_direct_write(struct kiocb *req, struct iov_iter *iter)
 		locked = false;
 	}
 
-	bio = bio_alloc_bioset(NULL,
-			       bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS),
-			       REQ_OP_WRITE,
-			       GFP_KERNEL,
+	bio = bio_alloc_bioset(GFP_KERNEL,
+			       iov_iter_npages(iter, BIO_MAX_PAGES),
 			       &c->dio_write_bioset);
 	dio = container_of(bio, struct dio_write, op.wbio.bio);
 	init_completion(&dio->done);
@@ -2563,7 +2653,7 @@ static int bch2_extend(struct user_namespace *mnt_userns,
 
 	truncate_setsize(&inode->v, iattr->ia_size);
 
-	return bch2_setattr_nonsize(mnt_userns, inode, iattr);
+	return bch2_setattr_nonsize(inode, iattr);
 }
 
 static int bch2_truncate_finish_fn(struct bch_inode_info *inode,
@@ -2685,7 +2775,7 @@ int bch2_truncate(struct user_namespace *mnt_userns,
 	ret = bch2_write_inode(c, inode, bch2_truncate_finish_fn, NULL, 0);
 	mutex_unlock(&inode->ei_update_lock);
 
-	ret = bch2_setattr_nonsize(mnt_userns, inode, iattr);
+	ret = bch2_setattr_nonsize(inode, iattr);
 err:
 	bch2_pagecache_block_put(&inode->ei_pagecache_lock);
 	return bch2_err_class(ret);
@@ -3102,10 +3192,6 @@ long bch2_fallocate_dispatch(struct file *file, int mode,
 	inode_dio_wait(&inode->v);
 	bch2_pagecache_block_get(&inode->ei_pagecache_lock);
 
-	ret = file_modified(file);
-	if (ret)
-		goto err;
-
 	if (!(mode & ~(FALLOC_FL_KEEP_SIZE|FALLOC_FL_ZERO_RANGE)))
 		ret = bchfs_fallocate(inode, mode, offset, len);
 	else if (mode == (FALLOC_FL_PUNCH_HOLE|FALLOC_FL_KEEP_SIZE))
@@ -3116,7 +3202,7 @@ long bch2_fallocate_dispatch(struct file *file, int mode,
 		ret = bchfs_fcollapse_finsert(inode, offset, len, false);
 	else
 		ret = -EOPNOTSUPP;
-err:
+
 	bch2_pagecache_block_put(&inode->ei_pagecache_lock);
 	inode_unlock(&inode->v);
 	percpu_ref_put(&c->writes);
@@ -3173,6 +3259,235 @@ err:
 	return bch2_quota_reservation_add(c, inode, res, sectors, true);
 }
 
+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)
@@ -3276,40 +3591,36 @@ static loff_t bch2_seek_pagecache_data(struct inode *vinode,
 				       loff_t start_offset,
 				       loff_t end_offset)
 {
-	struct folio_batch fbatch;
+	struct address_space *mapping = vinode->i_mapping;
+	struct page *page;
 	pgoff_t start_index	= start_offset >> PAGE_SHIFT;
 	pgoff_t end_index	= end_offset >> PAGE_SHIFT;
 	pgoff_t index		= start_index;
-	unsigned i;
 	loff_t ret;
 	int offset;
 
-	folio_batch_init(&fbatch);
-
-	while (filemap_get_folios(vinode->i_mapping,
-				  &index, end_index, &fbatch)) {
-		for (i = 0; i < folio_batch_count(&fbatch); i++) {
-			struct folio *folio = fbatch.folios[i];
+	while (index <= end_index) {
+		if (find_get_pages_range(mapping, &index, end_index, 1, &page)) {
+			lock_page(page);
 
-			folio_lock(folio);
-
-			offset = page_data_offset(&folio->page,
-					folio->index == start_index
+			offset = page_data_offset(page,
+					page->index == start_index
 					? start_offset & (PAGE_SIZE - 1)
 					: 0);
 			if (offset >= 0) {
-				ret = clamp(((loff_t) folio->index << PAGE_SHIFT) +
+				ret = clamp(((loff_t) page->index << PAGE_SHIFT) +
 					    offset,
 					    start_offset, end_offset);
-				folio_unlock(folio);
-				folio_batch_release(&fbatch);
+				unlock_page(page);
+				put_page(page);
 				return ret;
 			}
 
-			folio_unlock(folio);
+			unlock_page(page);
+			put_page(page);
+		} else {
+			break;
 		}
-		folio_batch_release(&fbatch);
-		cond_resched();
 	}
 
 	return end_offset;
diff --git a/fs/bcachefs/fs-io.h b/fs/bcachefs/fs-io.h
index a8835298613a..f9e7f49b13c7 100644
--- a/fs/bcachefs/fs-io.h
+++ b/fs/bcachefs/fs-io.h
@@ -15,13 +15,15 @@ int __must_check bch2_write_inode_size(struct bch_fs *,
 				       struct bch_inode_info *,
 				       loff_t, unsigned);
 
-int bch2_read_folio(struct file *, struct folio *);
+int bch2_writepage(struct page *, struct writeback_control *);
+int bch2_readpage(struct file *, struct page *);
 
 int bch2_writepages(struct address_space *, struct writeback_control *);
-void bch2_readahead(struct readahead_control *);
+int bch2_readpages(struct file *, struct address_space *,
+		   struct list_head *, unsigned);
 
 int bch2_write_begin(struct file *, struct address_space *, loff_t,
-		     unsigned, struct page **, void **);
+		     unsigned, unsigned, struct page **, void **);
 int bch2_write_end(struct file *, struct address_space *, loff_t,
 		   unsigned, unsigned, struct page *, void *);
 
@@ -34,6 +36,10 @@ int bch2_truncate(struct user_namespace *,
 		  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);
 
@@ -41,8 +47,10 @@ loff_t bch2_llseek(struct file *, loff_t, int);
 
 vm_fault_t bch2_page_fault(struct vm_fault *);
 vm_fault_t bch2_page_mkwrite(struct vm_fault *);
-void bch2_invalidate_folio(struct folio *, size_t, size_t);
-bool bch2_release_folio(struct folio *, gfp_t);
+void bch2_invalidatepage(struct page *, unsigned int, unsigned int);
+int bch2_releasepage(struct page *, gfp_t);
+int bch2_migrate_page(struct address_space *, struct page *,
+		      struct page *, enum migrate_mode);
 
 void bch2_fs_fsio_exit(struct bch_fs *);
 int bch2_fs_fsio_init(struct bch_fs *);
diff --git a/fs/bcachefs/fs-ioctl.c b/fs/bcachefs/fs-ioctl.c
index 2bb680827b44..2ff9e344e2d6 100644
--- a/fs/bcachefs/fs-ioctl.c
+++ b/fs/bcachefs/fs-ioctl.c
@@ -93,7 +93,7 @@ static int bch2_ioc_setflags(struct bch_fs *c,
 		return ret;
 
 	inode_lock(&inode->v);
-	if (!inode_owner_or_capable(file_mnt_user_ns(file), &inode->v)) {
+	if (!inode_owner_or_capable(&inode->v)) {
 		ret = -EACCES;
 		goto setflags_out;
 	}
@@ -172,7 +172,7 @@ static int bch2_ioc_fssetxattr(struct bch_fs *c,
 		return ret;
 
 	inode_lock(&inode->v);
-	if (!inode_owner_or_capable(file_mnt_user_ns(file), &inode->v)) {
+	if (!inode_owner_or_capable(&inode->v)) {
 		ret = -EACCES;
 		goto err;
 	}
@@ -284,20 +284,22 @@ static int bch2_ioc_goingdown(struct bch_fs *c, u32 __user *arg)
 	down_write(&c->vfs_sb->s_umount);
 
 	switch (flags) {
-	case FSOP_GOING_FLAGS_DEFAULT:
-		ret = freeze_bdev(c->vfs_sb->s_bdev);
+	case FSOP_GOING_FLAGS_DEFAULT: {
+		struct super_block *sb = freeze_bdev(c->vfs_sb->s_bdev);
 		if (ret)
 			goto err;
 
-		bch2_journal_flush(&c->journal);
-		c->vfs_sb->s_flags |= SB_RDONLY;
-		bch2_fs_emergency_read_only(c);
-		thaw_bdev(c->vfs_sb->s_bdev);
+		if (sb && !IS_ERR(sb)) {
+			bch2_journal_flush(&c->journal);
+			c->vfs_sb->s_flags |= SB_RDONLY;
+			bch2_fs_emergency_read_only(c);
+			thaw_bdev(c->vfs_sb->s_bdev, sb);
+		}
 		break;
+	}
 
 	case FSOP_GOING_FLAGS_LOGFLUSH:
 		bch2_journal_flush(&c->journal);
-		fallthrough;
 
 	case FSOP_GOING_FLAGS_NOLOGFLUSH:
 		c->vfs_sb->s_flags |= SB_RDONLY;
@@ -393,8 +395,7 @@ retry:
 		goto err3;
 	}
 
-	error = inode_permission(file_mnt_user_ns(filp),
-				 dir, MAY_WRITE | MAY_EXEC);
+	error = inode_permission(dir, MAY_WRITE | MAY_EXEC);
 	if (error)
 		goto err3;
 
@@ -409,7 +410,7 @@ retry:
 	    !arg.src_ptr)
 		snapshot_src.subvol = to_bch_ei(dir)->ei_inode.bi_subvol;
 
-	inode = __bch2_create(file_mnt_user_ns(filp), to_bch_ei(dir),
+	inode = __bch2_create(NULL, to_bch_ei(dir),
 			      dst_dentry, arg.mode|S_IFDIR,
 			      0, snapshot_src, create_flags);
 	error = PTR_ERR_OR_ZERO(inode);
@@ -458,10 +459,8 @@ static long bch2_ioctl_subvolume_destroy(struct bch_fs *c, struct file *filp,
 	dir = path.dentry->d_parent->d_inode;
 
 	ret = __bch2_unlink(dir, path.dentry, true);
-	if (!ret) {
-		fsnotify_rmdir(dir, path.dentry);
+	if (!ret)
 		d_delete(path.dentry);
-	}
 	path_put(&path);
 
 	return ret;
diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c
index 57e6e21896e1..6e539594e7e5 100644
--- a/fs/bcachefs/fs.c
+++ b/fs/bcachefs/fs.c
@@ -265,6 +265,7 @@ __bch2_create(struct user_namespace *mnt_userns,
 	      unsigned flags)
 {
 	struct bch_fs *c = dir->v.i_sb->s_fs_info;
+	struct user_namespace *ns = dir->v.i_sb->s_user_ns;
 	struct btree_trans trans;
 	struct bch_inode_unpacked dir_u;
 	struct bch_inode_info *inode, *old;
@@ -305,8 +306,8 @@ retry:
 				  inode_inum(dir), &dir_u, &inode_u,
 				  !(flags & BCH_CREATE_TMPFILE)
 				  ? &dentry->d_name : NULL,
-				  from_kuid(mnt_userns, current_fsuid()),
-				  from_kgid(mnt_userns, current_fsgid()),
+				  from_kuid(ns, current_fsuid()),
+				  from_kgid(ns, current_fsgid()),
 				  mode, rdev,
 				  default_acl, acl, snapshot_src, flags) ?:
 		bch2_quota_acct(c, bch_qid(&inode_u), Q_INO, 1,
@@ -410,12 +411,11 @@ static struct dentry *bch2_lookup(struct inode *vdir, struct dentry *dentry,
 	return d_splice_alias(vinode, dentry);
 }
 
-static int bch2_mknod(struct user_namespace *mnt_userns,
-		      struct inode *vdir, struct dentry *dentry,
+static int bch2_mknod(struct inode *vdir, struct dentry *dentry,
 		      umode_t mode, dev_t rdev)
 {
 	struct bch_inode_info *inode =
-		__bch2_create(mnt_userns, to_bch_ei(vdir), dentry, mode, rdev,
+		__bch2_create(NULL, to_bch_ei(vdir), dentry, mode, rdev,
 			      (subvol_inum) { 0 }, 0);
 
 	if (IS_ERR(inode))
@@ -425,11 +425,10 @@ static int bch2_mknod(struct user_namespace *mnt_userns,
 	return 0;
 }
 
-static int bch2_create(struct user_namespace *mnt_userns,
-		       struct inode *vdir, struct dentry *dentry,
+static int bch2_create(struct inode *vdir, struct dentry *dentry,
 		       umode_t mode, bool excl)
 {
-	return bch2_mknod(mnt_userns, vdir, dentry, mode|S_IFREG, 0);
+	return bch2_mknod(vdir, dentry, mode|S_IFREG, 0);
 }
 
 static int __bch2_link(struct bch_fs *c,
@@ -518,15 +517,14 @@ static int bch2_unlink(struct inode *vdir, struct dentry *dentry)
 	return __bch2_unlink(vdir, dentry, false);
 }
 
-static int bch2_symlink(struct user_namespace *mnt_userns,
-			struct inode *vdir, struct dentry *dentry,
+static int bch2_symlink(struct inode *vdir, struct dentry *dentry,
 			const char *symname)
 {
 	struct bch_fs *c = vdir->i_sb->s_fs_info;
 	struct bch_inode_info *dir = to_bch_ei(vdir), *inode;
 	int ret;
 
-	inode = __bch2_create(mnt_userns, dir, dentry, S_IFLNK|S_IRWXUGO, 0,
+	inode = __bch2_create(NULL, dir, dentry, S_IFLNK|S_IRWXUGO, 0,
 			      (subvol_inum) { 0 }, BCH_CREATE_TMPFILE);
 	if (unlikely(IS_ERR(inode)))
 		return PTR_ERR(inode);
@@ -553,14 +551,12 @@ err:
 	return ret;
 }
 
-static int bch2_mkdir(struct user_namespace *mnt_userns,
-		      struct inode *vdir, struct dentry *dentry, umode_t mode)
+static int bch2_mkdir(struct inode *vdir, struct dentry *dentry, umode_t mode)
 {
-	return bch2_mknod(mnt_userns, vdir, dentry, mode|S_IFDIR, 0);
+	return bch2_mknod(vdir, dentry, mode|S_IFDIR, 0);
 }
 
-static int bch2_rename2(struct user_namespace *mnt_userns,
-			struct inode *src_vdir, struct dentry *src_dentry,
+static int bch2_rename2(struct inode *src_vdir, struct dentry *src_dentry,
 			struct inode *dst_vdir, struct dentry *dst_dentry,
 			unsigned flags)
 {
@@ -666,8 +662,7 @@ err:
 	return ret;
 }
 
-static void bch2_setattr_copy(struct user_namespace *mnt_userns,
-			      struct bch_inode_info *inode,
+static void bch2_setattr_copy(struct bch_inode_info *inode,
 			      struct bch_inode_unpacked *bi,
 			      struct iattr *attr)
 {
@@ -675,9 +670,9 @@ static void bch2_setattr_copy(struct user_namespace *mnt_userns,
 	unsigned int ia_valid = attr->ia_valid;
 
 	if (ia_valid & ATTR_UID)
-		bi->bi_uid = from_kuid(mnt_userns, attr->ia_uid);
+		bi->bi_uid = from_kuid(c->vfs_sb->s_user_ns, attr->ia_uid);
 	if (ia_valid & ATTR_GID)
-		bi->bi_gid = from_kgid(mnt_userns, attr->ia_gid);
+		bi->bi_gid = from_kgid(c->vfs_sb->s_user_ns, attr->ia_gid);
 
 	if (ia_valid & ATTR_SIZE)
 		bi->bi_size = attr->ia_size;
@@ -696,14 +691,13 @@ static void bch2_setattr_copy(struct user_namespace *mnt_userns,
 			: inode->v.i_gid;
 
 		if (!in_group_p(gid) &&
-		    !capable_wrt_inode_uidgid(mnt_userns, &inode->v, CAP_FSETID))
+		    !capable_wrt_inode_uidgid(&inode->v, CAP_FSETID))
 			mode &= ~S_ISGID;
 		bi->bi_mode = mode;
 	}
 }
 
-int bch2_setattr_nonsize(struct user_namespace *mnt_userns,
-			 struct bch_inode_info *inode,
+int bch2_setattr_nonsize(struct bch_inode_info *inode,
 			 struct iattr *attr)
 {
 	struct bch_fs *c = inode->v.i_sb->s_fs_info;
@@ -740,7 +734,7 @@ retry:
 	if (ret)
 		goto btree_err;
 
-	bch2_setattr_copy(mnt_userns, inode, &inode_u, attr);
+	bch2_setattr_copy(inode, &inode_u, attr);
 
 	if (attr->ia_valid & ATTR_MODE) {
 		ret = bch2_acl_chmod(&trans, inode_inum(inode), &inode_u,
@@ -772,8 +766,7 @@ err:
 	return bch2_err_class(ret);
 }
 
-static int bch2_getattr(struct user_namespace *mnt_userns,
-			const struct path *path, struct kstat *stat,
+static int bch2_getattr(const struct path *path, struct kstat *stat,
 			u32 request_mask, unsigned query_flags)
 {
 	struct bch_inode_info *inode = to_bch_ei(d_inode(path->dentry));
@@ -813,28 +806,26 @@ static int bch2_getattr(struct user_namespace *mnt_userns,
 	return 0;
 }
 
-static int bch2_setattr(struct user_namespace *mnt_userns,
-			struct dentry *dentry, struct iattr *iattr)
+static int bch2_setattr(struct dentry *dentry, struct iattr *iattr)
 {
 	struct bch_inode_info *inode = to_bch_ei(dentry->d_inode);
 	int ret;
 
 	lockdep_assert_held(&inode->v.i_rwsem);
 
-	ret = setattr_prepare(mnt_userns, dentry, iattr);
+	ret = setattr_prepare(dentry, iattr);
 	if (ret)
 		return ret;
 
 	return iattr->ia_valid & ATTR_SIZE
-		? bch2_truncate(mnt_userns, inode, iattr)
-		: bch2_setattr_nonsize(mnt_userns, inode, iattr);
+		? bch2_truncate(NULL, inode, iattr)
+		: bch2_setattr_nonsize(inode, iattr);
 }
 
-static int bch2_tmpfile(struct user_namespace *mnt_userns,
-			struct inode *vdir, struct dentry *dentry, umode_t mode)
+static int bch2_tmpfile(struct inode *vdir, struct dentry *dentry, umode_t mode)
 {
 	struct bch_inode_info *inode =
-		__bch2_create(mnt_userns, to_bch_ei(vdir), dentry, mode, 0,
+		__bch2_create(NULL, to_bch_ei(vdir), dentry, mode, 0,
 			      (subvol_inum) { 0 }, BCH_CREATE_TMPFILE);
 
 	if (IS_ERR(inode))
@@ -913,10 +904,6 @@ static int bch2_fiemap(struct inode *vinode, struct fiemap_extent_info *info,
 	u32 snapshot;
 	int ret = 0;
 
-	ret = fiemap_prep(&ei->v, info, start, &len, FIEMAP_FLAG_SYNC);
-	if (ret)
-		return ret;
-
 	if (start + len < start)
 		return -EINVAL;
 
@@ -1032,6 +1019,15 @@ static int bch2_vfs_readdir(struct file *file, struct dir_context *ctx)
 	return bch2_readdir(c, inode_inum(inode), 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,
@@ -1046,7 +1042,7 @@ static const struct file_operations bch_file_operations = {
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= bch2_compat_fs_ioctl,
 #endif
-	.remap_file_range = bch2_remap_file_range,
+	.clone_file_range = bch2_clone_file_range,
 };
 
 static const struct inode_operations bch_file_inode_operations = {
@@ -1113,17 +1109,17 @@ static const struct inode_operations bch_special_inode_operations = {
 };
 
 static const struct address_space_operations bch_address_space_operations = {
-	.read_folio	= bch2_read_folio,
+	.readpage	= bch2_readpage,
 	.writepages	= bch2_writepages,
-	.readahead	= bch2_readahead,
-	.dirty_folio	= filemap_dirty_folio,
+	.readpages	= bch2_readpages,
+	.set_page_dirty	= __set_page_dirty_nobuffers,
 	.write_begin	= bch2_write_begin,
 	.write_end	= bch2_write_end,
-	.invalidate_folio = bch2_invalidate_folio,
-	.release_folio	= bch2_release_folio,
+	.invalidatepage	= bch2_invalidatepage,
+	.releasepage	= bch2_releasepage,
 	.direct_IO	= noop_direct_IO,
 #ifdef CONFIG_MIGRATION
-	.migrate_folio	= filemap_migrate_folio,
+	.migratepage	= bch2_migrate_page,
 #endif
 	.error_remove_page = generic_error_remove_page,
 };
@@ -1574,14 +1570,13 @@ static int bch2_sync_fs(struct super_block *sb, int wait)
 static struct bch_fs *bch2_path_to_fs(const char *path)
 {
 	struct bch_fs *c;
-	dev_t dev;
-	int ret;
+	struct block_device *bdev = lookup_bdev(path);
 
-	ret = lookup_bdev(path, &dev);
-	if (ret)
-		return ERR_PTR(ret);
+	if (IS_ERR(bdev))
+		return ERR_CAST(bdev);
 
-	c = bch2_dev_to_fs(dev);
+	c = bch2_dev_to_fs(bdev->bd_dev);
+	bdput(bdev);
 	if (c)
 		closure_put(&c->cl);
 	return c ?: ERR_PTR(-ENOENT);
@@ -1843,8 +1838,6 @@ got_sb:
 	sb->s_xattr		= bch2_xattr_handlers;
 	sb->s_magic		= BCACHEFS_STATFS_MAGIC;
 	sb->s_time_gran		= c->sb.nsec_per_time_unit;
-	sb->s_time_min		= div_s64(S64_MIN, c->sb.time_units_per_sec) + 1;
-	sb->s_time_max		= div_s64(S64_MAX, c->sb.time_units_per_sec);
 	c->vfs_sb		= sb;
 	strlcpy(sb->s_id, c->name, sizeof(sb->s_id));
 
@@ -1852,7 +1845,9 @@ got_sb:
 	if (ret)
 		goto err_put_super;
 
-	sb->s_bdi->ra_pages		= VM_READAHEAD_PAGES;
+	sb->s_bdi->congested_fn		= bch2_congested;
+	sb->s_bdi->congested_data	= c;
+	sb->s_bdi->ra_pages		= VM_MAX_READAHEAD * 1024 / PAGE_SIZE;
 
 	for_each_online_member(ca, c, i) {
 		struct block_device *bdev = ca->disk_sb.bdev;
@@ -1871,7 +1866,7 @@ got_sb:
 		sb->s_flags	|= SB_POSIXACL;
 #endif
 
-	sb->s_shrink.seeks = 0;
+	sb->s_shrink.seeks = 1;
 
 	vinode = bch2_vfs_inode_get(c, BCACHEFS_ROOT_SUBVOL_INUM);
 	ret = PTR_ERR_OR_ZERO(vinode);
diff --git a/fs/bcachefs/fs.h b/fs/bcachefs/fs.h
index 9f4b57e30e2a..bb005f768a54 100644
--- a/fs/bcachefs/fs.h
+++ b/fs/bcachefs/fs.h
@@ -186,8 +186,7 @@ void bch2_inode_update_after_write(struct btree_trans *,
 int __must_check bch2_write_inode(struct bch_fs *, struct bch_inode_info *,
 				  inode_set_fn, void *, unsigned);
 
-int bch2_setattr_nonsize(struct user_namespace *,
-			 struct bch_inode_info *,
+int bch2_setattr_nonsize(struct bch_inode_info *,
 			 struct iattr *);
 int __bch2_unlink(struct inode *, struct dentry *, bool);
 
diff --git a/fs/bcachefs/fsck.c b/fs/bcachefs/fsck.c
index 12f2ef4417cb..fbcbf2953230 100644
--- a/fs/bcachefs/fsck.c
+++ b/fs/bcachefs/fsck.c
@@ -2087,8 +2087,8 @@ static void inc_link(struct bch_fs *c, struct snapshots_seen *s,
 	if (inum < range_start || inum >= range_end)
 		return;
 
-	link = __inline_bsearch(&key, links->d, links->nr,
-				sizeof(links->d[0]), nlink_cmp);
+	link = bsearch(&key, links->d, links->nr,
+		       sizeof(links->d[0]), nlink_cmp);
 	if (!link)
 		return;
 
diff --git a/fs/bcachefs/io.c b/fs/bcachefs/io.c
index 558d0c232816..02a7153adbba 100644
--- a/fs/bcachefs/io.c
+++ b/fs/bcachefs/io.c
@@ -136,10 +136,10 @@ 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, iter)
+	bio_for_each_segment_all(bv, bio, i)
 		if (bv->bv_page != ZERO_PAGE(0))
 			mempool_free(bv->bv_page, &c->bio_bounce_pages);
 	bio->bi_vcnt = 0;
@@ -470,8 +470,8 @@ void bch2_submit_wbio_replicas(struct bch_write_bio *wbio, struct bch_fs *c,
 		ca = bch_dev_bkey_exists(c, ptr->dev);
 
 		if (to_entry(ptr + 1) < ptrs.end) {
-			n = to_wbio(bio_alloc_clone(NULL, &wbio->bio,
-						GFP_NOIO, &ca->replica_set));
+			n = to_wbio(bio_clone_fast(&wbio->bio, GFP_NOIO,
+						   &ca->replica_set));
 
 			n->bio.bi_end_io	= wbio->bio.bi_end_io;
 			n->bio.bi_private	= wbio->bio.bi_private;
@@ -712,10 +712,9 @@ static struct bio *bch2_write_bio_alloc(struct bch_fs *c,
 				       ? ((unsigned long) buf & (PAGE_SIZE - 1))
 				       : 0), PAGE_SIZE);
 
-	pages = min(pages, BIO_MAX_VECS);
+	pages = min_t(unsigned, pages, BIO_MAX_PAGES);
 
-	bio = bio_alloc_bioset(NULL, pages, 0,
-			       GFP_NOIO, &c->bio_write);
+	bio = bio_alloc_bioset(GFP_NOIO, pages, &c->bio_write);
 	wbio			= wbio_init(bio);
 	wbio->put_bio		= true;
 	/* copy WRITE_SYNC flag */
@@ -1446,7 +1445,7 @@ static struct promote_op *__promote_alloc(struct bch_fs *c,
 		goto err;
 
 	rbio_init(&(*rbio)->bio, opts);
-	bio_init(&(*rbio)->bio, NULL, (*rbio)->bio.bi_inline_vecs, pages, 0);
+	bio_init(&(*rbio)->bio, (*rbio)->bio.bi_inline_vecs, pages);
 
 	if (bch2_bio_alloc_pages(&(*rbio)->bio, sectors << 9,
 				 GFP_NOIO))
@@ -1461,7 +1460,7 @@ static struct promote_op *__promote_alloc(struct bch_fs *c,
 		goto err;
 
 	bio = &op->write.op.wbio.bio;
-	bio_init(bio, NULL, bio->bi_inline_vecs, pages, 0);
+	bio_init(bio, bio->bi_inline_vecs, pages);
 
 	ret = bch2_data_update_init(c, &op->write,
 			writepoint_hashed((unsigned long) current),
@@ -2144,10 +2143,8 @@ get_bio:
 	} else if (bounce) {
 		unsigned sectors = pick.crc.compressed_size;
 
-		rbio = rbio_init(bio_alloc_bioset(NULL,
+		rbio = rbio_init(bio_alloc_bioset(GFP_NOIO,
 						  DIV_ROUND_UP(sectors, PAGE_SECTORS),
-						  0,
-						  GFP_NOIO,
 						  &c->bio_read_split),
 				 orig->opts);
 
@@ -2163,8 +2160,8 @@ get_bio:
 		 * from the whole bio, in which case we don't want to retry and
 		 * lose the error)
 		 */
-		rbio = rbio_init(bio_alloc_clone(NULL, &orig->bio, GFP_NOIO,
-						 &c->bio_read_split),
+		rbio = rbio_init(bio_clone_fast(&orig->bio, GFP_NOIO,
+						&c->bio_read_split),
 				 orig->opts);
 		rbio->bio.bi_iter = iter;
 		rbio->split	= true;
diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c
index ab594623341f..1e53ba70a448 100644
--- a/fs/bcachefs/journal.c
+++ b/fs/bcachefs/journal.c
@@ -1168,12 +1168,10 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb)
 
 	nr_bvecs = DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE);
 
-	ca->journal.bio = bio_kmalloc(nr_bvecs, GFP_KERNEL);
+	ca->journal.bio = bio_kmalloc(GFP_KERNEL, nr_bvecs);
 	if (!ca->journal.bio)
 		return -ENOMEM;
 
-	bio_init(ca->journal.bio, NULL, ca->journal.bio->bi_inline_vecs, nr_bvecs, 0);
-
 	ja->buckets = kcalloc(ja->nr, sizeof(u64), GFP_KERNEL);
 	if (!ja->buckets)
 		return -ENOMEM;
diff --git a/fs/bcachefs/journal.h b/fs/bcachefs/journal.h
index d3caa7ea7ce9..5c31fe81db9d 100644
--- a/fs/bcachefs/journal.h
+++ b/fs/bcachefs/journal.h
@@ -279,7 +279,7 @@ static inline void bch2_journal_res_put(struct journal *j,
 	if (!res->ref)
 		return;
 
-	lock_release(&j->res_map, _THIS_IP_);
+	lock_release(&j->res_map, 0, _THIS_IP_);
 
 	while (res->u64s)
 		bch2_journal_add_entry(j, res,
diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c
index 68113a08f1b9..047fcf341388 100644
--- a/fs/bcachefs/journal_io.c
+++ b/fs/bcachefs/journal_io.c
@@ -873,10 +873,11 @@ reread:
 				end - offset, buf->size >> 9);
 			nr_bvecs = buf_pages(buf->data, sectors_read << 9);
 
-			bio = bio_kmalloc(nr_bvecs, GFP_KERNEL);
-			bio_init(bio, ca->disk_sb.bdev, bio->bi_inline_vecs, nr_bvecs, REQ_OP_READ);
+			bio = bio_kmalloc(GFP_KERNEL, nr_bvecs);
 
+			bio_set_dev(bio, ca->disk_sb.bdev);
 			bio->bi_iter.bi_sector = offset;
+			bio_set_op_attrs(bio, REQ_OP_READ, 0);
 			bch2_bio_map(bio, buf->data, sectors_read << 9);
 
 			ret = submit_bio_wait(bio);
@@ -1613,10 +1614,12 @@ static void do_journal_write(struct closure *cl)
 			     sectors);
 
 		bio = ca->journal.bio;
-		bio_reset(bio, ca->disk_sb.bdev, REQ_OP_WRITE|REQ_SYNC|REQ_META);
+		bio_reset(bio);
+		bio_set_dev(bio, ca->disk_sb.bdev);
 		bio->bi_iter.bi_sector	= ptr->offset;
 		bio->bi_end_io		= journal_write_endio;
 		bio->bi_private		= ca;
+		bio->bi_opf		= REQ_OP_WRITE|REQ_SYNC|REQ_META;
 
 		BUG_ON(bio->bi_iter.bi_sector == ca->prev_journal_sector);
 		ca->prev_journal_sector = bio->bi_iter.bi_sector;
@@ -1794,7 +1797,9 @@ retry_alloc:
 			percpu_ref_get(&ca->io_ref);
 
 			bio = ca->journal.bio;
-			bio_reset(bio, ca->disk_sb.bdev, REQ_OP_FLUSH);
+			bio_reset(bio);
+			bio_set_dev(bio, ca->disk_sb.bdev);
+			bio->bi_opf		= REQ_OP_FLUSH;
 			bio->bi_end_io		= journal_write_endio;
 			bio->bi_private		= ca;
 			closure_bio_submit(bio, cl);
diff --git a/fs/bcachefs/journal_reclaim.c b/fs/bcachefs/journal_reclaim.c
index e69595bd1359..b11e50759fa5 100644
--- a/fs/bcachefs/journal_reclaim.c
+++ b/fs/bcachefs/journal_reclaim.c
@@ -12,6 +12,7 @@
 
 #include <linux/kthread.h>
 #include <linux/sched/mm.h>
+#include <linux/sched/task.h>
 #include <trace/events/bcachefs.h>
 
 /* Free space calculations: */
@@ -283,11 +284,11 @@ void bch2_journal_do_discards(struct journal *j)
 		while (should_discard_bucket(j, ja)) {
 			if (!c->opts.nochanges &&
 			    ca->mi.discard &&
-			    bdev_max_discard_sectors(ca->disk_sb.bdev))
+			    blk_queue_discard(bdev_get_queue(ca->disk_sb.bdev)))
 				blkdev_issue_discard(ca->disk_sb.bdev,
 					bucket_to_sector(ca,
 						ja->buckets[ja->discard_idx]),
-					ca->mi.bucket_size, GFP_NOIO);
+					ca->mi.bucket_size, GFP_NOIO, 0);
 
 			spin_lock(&j->lock);
 			ja->discard_idx = (ja->discard_idx + 1) % ja->nr;
diff --git a/fs/bcachefs/move.c b/fs/bcachefs/move.c
index 4f4dfaa7bfb7..0ebb1c1773be 100644
--- a/fs/bcachefs/move.c
+++ b/fs/bcachefs/move.c
@@ -277,7 +277,7 @@ static int bch2_move_extent(struct btree_trans *trans,
 	io->read_sectors	= k.k->size;
 	io->write_sectors	= k.k->size;
 
-	bio_init(&io->write.op.wbio.bio, NULL, io->bi_inline_vecs, pages, 0);
+	bio_init(&io->write.op.wbio.bio, io->bi_inline_vecs, pages);
 	bio_set_prio(&io->write.op.wbio.bio,
 		     IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0));
 
@@ -287,7 +287,7 @@ static int bch2_move_extent(struct btree_trans *trans,
 
 	io->rbio.c		= c;
 	io->rbio.opts		= io_opts;
-	bio_init(&io->rbio.bio, NULL, io->bi_inline_vecs, pages, 0);
+	bio_init(&io->rbio.bio, io->bi_inline_vecs, pages);
 	io->rbio.bio.bi_vcnt = pages;
 	bio_set_prio(&io->rbio.bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0));
 	io->rbio.bio.bi_iter.bi_size = sectors << 9;
diff --git a/fs/bcachefs/move.h b/fs/bcachefs/move.h
index c0fec69bbb6a..16b2dad30e64 100644
--- a/fs/bcachefs/move.h
+++ b/fs/bcachefs/move.h
@@ -60,8 +60,6 @@ int bch2_data_job(struct bch_fs *,
 		  struct bch_move_stats *,
 		  struct bch_ioctl_data);
 
-inline void bch_move_stats_init(struct bch_move_stats *stats,
-				char *name);
-
+void bch_move_stats_init(struct bch_move_stats *, char *);
 
 #endif /* _BCACHEFS_MOVE_H */
diff --git a/fs/bcachefs/str_hash.h b/fs/bcachefs/str_hash.h
index 6178ae620ff1..fa411a3a10d5 100644
--- a/fs/bcachefs/str_hash.h
+++ b/fs/bcachefs/str_hash.h
@@ -13,7 +13,7 @@
 
 #include <linux/crc32c.h>
 #include <crypto/hash.h>
-#include <crypto/sha2.h>
+#include <crypto/sha.h>
 
 static inline enum bch_str_hash_type
 bch2_str_hash_opt_to_type(struct bch_fs *c, enum bch_str_hash_opts opt)
diff --git a/fs/bcachefs/super-io.c b/fs/bcachefs/super-io.c
index cbc5979a5181..75770453ba08 100644
--- a/fs/bcachefs/super-io.c
+++ b/fs/bcachefs/super-io.c
@@ -143,12 +143,10 @@ int bch2_sb_realloc(struct bch_sb_handle *sb, unsigned u64s)
 	if (sb->have_bio) {
 		unsigned nr_bvecs = DIV_ROUND_UP(new_buffer_size, PAGE_SIZE);
 
-		bio = bio_kmalloc(nr_bvecs, GFP_KERNEL);
+		bio = bio_kmalloc(GFP_KERNEL, nr_bvecs);
 		if (!bio)
 			return -ENOMEM;
 
-		bio_init(bio, NULL, bio->bi_inline_vecs, nr_bvecs, 0);
-
 		if (sb->bio)
 			kfree(sb->bio);
 		sb->bio = bio;
@@ -296,7 +294,7 @@ static int bch2_sb_validate(struct bch_sb_handle *disk_sb, struct printbuf *out,
 	block_size = le16_to_cpu(sb->block_size);
 
 	if (block_size > PAGE_SECTORS) {
-		prt_printf(out, "Block size too big (got %u, max %u)",
+		prt_printf(out, "Block size too big (got %u, max %lu)",
 		       block_size, PAGE_SECTORS);
 		return -EINVAL;
 	}
@@ -530,8 +528,10 @@ static int read_one_super(struct bch_sb_handle *sb, u64 offset, struct printbuf
 	size_t bytes;
 	int ret;
 reread:
-	bio_reset(sb->bio, sb->bdev, REQ_OP_READ|REQ_SYNC|REQ_META);
+	bio_reset(sb->bio);
+	bio_set_dev(sb->bio, sb->bdev);
 	sb->bio->bi_iter.bi_sector = offset;
+	bio_set_op_attrs(sb->bio, REQ_OP_READ, REQ_SYNC|REQ_META);
 	bch2_bio_map(sb->bio, sb->sb, sb->buffer_size);
 
 	ret = submit_bio_wait(sb->bio);
@@ -659,8 +659,10 @@ int bch2_read_super(const char *path, struct bch_opts *opts,
 	 * Error reading primary superblock - read location of backup
 	 * superblocks:
 	 */
-	bio_reset(sb->bio, sb->bdev, REQ_OP_READ|REQ_SYNC|REQ_META);
+	bio_reset(sb->bio);
+	bio_set_dev(sb->bio, sb->bdev);
 	sb->bio->bi_iter.bi_sector = BCH_SB_LAYOUT_SECTOR;
+	bio_set_op_attrs(sb->bio, REQ_OP_READ, REQ_SYNC|REQ_META);
 	/*
 	 * use sb buffer to read layout, since sb buffer is page aligned but
 	 * layout won't be:
@@ -744,10 +746,12 @@ static void read_back_super(struct bch_fs *c, struct bch_dev *ca)
 	struct bch_sb *sb = ca->disk_sb.sb;
 	struct bio *bio = ca->disk_sb.bio;
 
-	bio_reset(bio, ca->disk_sb.bdev, REQ_OP_READ|REQ_SYNC|REQ_META);
+	bio_reset(bio);
+	bio_set_dev(bio, ca->disk_sb.bdev);
 	bio->bi_iter.bi_sector	= le64_to_cpu(sb->layout.sb_offset[0]);
 	bio->bi_end_io		= write_super_endio;
 	bio->bi_private		= ca;
+	bio_set_op_attrs(bio, REQ_OP_READ, REQ_SYNC|REQ_META);
 	bch2_bio_map(bio, ca->sb_read_scratch, PAGE_SIZE);
 
 	this_cpu_add(ca->io_done->sectors[READ][BCH_DATA_sb],
@@ -768,10 +772,12 @@ static void write_one_super(struct bch_fs *c, struct bch_dev *ca, unsigned idx)
 	sb->csum = csum_vstruct(c, BCH_SB_CSUM_TYPE(sb),
 				null_nonce(), sb);
 
-	bio_reset(bio, ca->disk_sb.bdev, REQ_OP_WRITE|REQ_SYNC|REQ_META);
+	bio_reset(bio);
+	bio_set_dev(bio, ca->disk_sb.bdev);
 	bio->bi_iter.bi_sector	= le64_to_cpu(sb->offset);
 	bio->bi_end_io		= write_super_endio;
 	bio->bi_private		= ca;
+	bio_set_op_attrs(bio, REQ_OP_WRITE, REQ_SYNC|REQ_META);
 	bch2_bio_map(bio, sb,
 		     roundup((size_t) vstruct_bytes(sb),
 			     bdev_logical_block_size(ca->disk_sb.bdev)));
diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c
index 3f674bf061ff..84fc22b3d2c8 100644
--- a/fs/bcachefs/super.c
+++ b/fs/bcachefs/super.c
@@ -65,19 +65,10 @@ MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Kent Overstreet <kent.overstreet@gmail.com>");
 
 #define KTYPE(type)							\
-static const struct attribute_group type ## _group = {			\
-	.attrs = type ## _files						\
-};									\
-									\
-static const struct attribute_group *type ## _groups[] = {		\
-	&type ## _group,						\
-	NULL								\
-};									\
-									\
-static const struct kobj_type type ## _ktype = {			\
+struct kobj_type type ## _ktype = {					\
 	.release	= type ## _release,				\
 	.sysfs_ops	= &type ## _sysfs_ops,				\
-	.default_groups = type ## _groups				\
+	.default_attrs	= type ## _files				\
 }
 
 static void bch2_fs_release(struct kobject *);
@@ -182,6 +173,44 @@ static void bch2_dev_usage_journal_reserve(struct bch_fs *c)
 			&c->dev_usage_journal_res, u64s * nr);
 }
 
+int bch2_congested(void *data, int bdi_bits)
+{
+	struct bch_fs *c = data;
+	struct backing_dev_info *bdi;
+	struct bch_dev *ca;
+	unsigned i;
+	int ret = 0;
+
+	rcu_read_lock();
+	if (bdi_bits & (1 << WB_sync_congested)) {
+		/* Reads - check all devices: */
+		for_each_readable_member(ca, c, i) {
+			bdi = ca->disk_sb.bdev->bd_bdi;
+
+			if (bdi_congested(bdi, bdi_bits)) {
+				ret = 1;
+				break;
+			}
+		}
+	} else {
+		const struct bch_devs_mask *devs =
+			bch2_target_to_mask(c, c->opts.foreground_target) ?:
+			&c->rw_devs[BCH_DATA_user];
+
+		for_each_member_device_rcu(ca, c, i, devs) {
+			bdi = ca->disk_sb.bdev->bd_bdi;
+
+			if (bdi_congested(bdi, bdi_bits)) {
+				ret = 1;
+				break;
+			}
+		}
+	}
+	rcu_read_unlock();
+
+	return ret;
+}
+
 /* Filesystem RO/RW: */
 
 /*
@@ -502,7 +531,8 @@ void __bch2_fs_stop(struct bch_fs *c)
 	for_each_member_device(ca, c, i)
 		if (ca->kobj.state_in_sysfs &&
 		    ca->disk_sb.bdev)
-			sysfs_remove_link(bdev_kobj(ca->disk_sb.bdev), "bcachefs");
+			sysfs_remove_link(&part_to_dev(ca->disk_sb.bdev->bd_part)->kobj,
+					  "bcachefs");
 
 	if (c->kobj.state_in_sysfs)
 		kobject_del(&c->kobj);
@@ -993,7 +1023,8 @@ static void bch2_dev_free(struct bch_dev *ca)
 
 	if (ca->kobj.state_in_sysfs &&
 	    ca->disk_sb.bdev)
-		sysfs_remove_link(bdev_kobj(ca->disk_sb.bdev), "bcachefs");
+		sysfs_remove_link(&part_to_dev(ca->disk_sb.bdev->bd_part)->kobj,
+				  "bcachefs");
 
 	if (ca->kobj.state_in_sysfs)
 		kobject_del(&ca->kobj);
@@ -1029,7 +1060,10 @@ static void __bch2_dev_offline(struct bch_fs *c, struct bch_dev *ca)
 	wait_for_completion(&ca->io_ref_completion);
 
 	if (ca->kobj.state_in_sysfs) {
-		sysfs_remove_link(bdev_kobj(ca->disk_sb.bdev), "bcachefs");
+		struct kobject *block =
+			&part_to_dev(ca->disk_sb.bdev->bd_part)->kobj;
+
+		sysfs_remove_link(block, "bcachefs");
 		sysfs_remove_link(&ca->kobj, "block");
 	}
 
@@ -1066,12 +1100,12 @@ static int bch2_dev_sysfs_online(struct bch_fs *c, struct bch_dev *ca)
 	}
 
 	if (ca->disk_sb.bdev) {
-		struct kobject *block = bdev_kobj(ca->disk_sb.bdev);
+		struct kobject *block =
+			&part_to_dev(ca->disk_sb.bdev->bd_part)->kobj;
 
 		ret = sysfs_create_link(block, &ca->kobj, "bcachefs");
 		if (ret)
 			return ret;
-
 		ret = sysfs_create_link(&ca->kobj, block, "block");
 		if (ret)
 			return ret;
diff --git a/fs/bcachefs/super.h b/fs/bcachefs/super.h
index 8501adaff4c2..de7b40a886ec 100644
--- a/fs/bcachefs/super.h
+++ b/fs/bcachefs/super.h
@@ -223,6 +223,7 @@ static inline bool is_superblock_bucket(struct bch_dev *ca, u64 b)
 
 struct bch_fs *bch2_dev_to_fs(dev_t);
 struct bch_fs *bch2_uuid_to_fs(uuid_le);
+int bch2_congested(void *, int);
 
 bool bch2_dev_state_allowed(struct bch_fs *, struct bch_dev *,
 			   enum bch_member_state, int);
diff --git a/fs/bcachefs/sysfs.c b/fs/bcachefs/sysfs.c
index 103fde97231f..db7abe32517a 100644
--- a/fs/bcachefs/sysfs.c
+++ b/fs/bcachefs/sysfs.c
@@ -290,7 +290,7 @@ static int bch2_compression_stats_to_text(struct printbuf *out, struct bch_fs *c
 	    incompressible_sectors = 0,
 	    compressed_sectors_compressed = 0,
 	    compressed_sectors_uncompressed = 0;
-	int ret;
+	int ret = 0;
 
 	if (!test_bit(BCH_FS_STARTED, &c->flags))
 		return -EPERM;
diff --git a/fs/bcachefs/util.c b/fs/bcachefs/util.c
index f08215af359f..04a8b4b188bd 100644
--- a/fs/bcachefs/util.c
+++ b/fs/bcachefs/util.c
@@ -299,21 +299,6 @@ void bch2_print_string_as_lines(const char *prefix, const char *lines)
 
 int bch2_prt_backtrace(struct printbuf *out, struct task_struct *task)
 {
-	unsigned long entries[32];
-	unsigned i, nr_entries;
-	int ret;
-
-	ret = down_read_killable(&task->signal->exec_update_lock);
-	if (ret)
-		return ret;
-
-	nr_entries = stack_trace_save_tsk(task, entries, ARRAY_SIZE(entries), 0);
-	for (i = 0; i < nr_entries; i++) {
-		prt_printf(out, "[<0>] %pB", (void *)entries[i]);
-		prt_newline(out);
-	}
-
-	up_read(&task->signal->exec_update_lock);
 	return 0;
 }
 
diff --git a/fs/bcachefs/util.h b/fs/bcachefs/util.h
index 846e6024a80b..4a6968aa1cc1 100644
--- a/fs/bcachefs/util.h
+++ b/fs/bcachefs/util.h
@@ -20,6 +20,9 @@
 #include <linux/workqueue.h>
 #include <linux/mean_and_variance.h>
 
+#define PAGE_SECTORS_SHIFT	(PAGE_SHIFT - 9)
+#define PAGE_SECTORS		(1UL << PAGE_SECTORS_SHIFT)
+
 struct closure;
 
 #ifdef CONFIG_BCACHEFS_DEBUG
@@ -87,7 +90,7 @@ static inline void *vpmalloc(size_t size, gfp_t gfp_mask)
 {
 	return (void *) __get_free_pages(gfp_mask|__GFP_NOWARN,
 					 get_order(size)) ?:
-		__vmalloc(size, gfp_mask);
+		__vmalloc(size, gfp_mask, PAGE_KERNEL);
 }
 
 static inline void kvpfree(void *p, size_t size)
@@ -669,6 +672,35 @@ 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));
diff --git a/fs/bcachefs/varint.c b/fs/bcachefs/varint.c
index 5143b603bf67..a2d6bb7136c7 100644
--- a/fs/bcachefs/varint.c
+++ b/fs/bcachefs/varint.c
@@ -1,7 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include <linux/bitops.h>
-#include <linux/math.h>
 #include <linux/string.h>
 #include <asm/unaligned.h>
 
diff --git a/fs/bcachefs/xattr.c b/fs/bcachefs/xattr.c
index 4fc1c3afab69..d409596523d3 100644
--- a/fs/bcachefs/xattr.c
+++ b/fs/bcachefs/xattr.c
@@ -378,7 +378,6 @@ static int bch2_xattr_get_handler(const struct xattr_handler *handler,
 }
 
 static int bch2_xattr_set_handler(const struct xattr_handler *handler,
-				  struct user_namespace *mnt_userns,
 				  struct dentry *dentry, struct inode *vinode,
 				  const char *name, const void *value,
 				  size_t size, int flags)
@@ -517,7 +516,6 @@ static int inode_opt_set_fn(struct bch_inode_info *inode,
 }
 
 static int bch2_xattr_bcachefs_set(const struct xattr_handler *handler,
-				   struct user_namespace *mnt_userns,
 				   struct dentry *dentry, struct inode *vinode,
 				   const char *name, const void *value,
 				   size_t size, int flags)
diff --git a/include/linux/generic-radix-tree.h b/include/linux/generic-radix-tree.h
index c74b7376990d..5aa1986128df 100644
--- a/include/linux/generic-radix-tree.h
+++ b/include/linux/generic-radix-tree.h
@@ -40,7 +40,6 @@
 #include <linux/bug.h>
 #include <linux/limits.h>
 #include <linux/log2.h>
-#include <linux/math.h>
 #include <linux/types.h>
 
 struct genradix_root;
diff --git a/include/trace/events/bcachefs.h b/include/trace/events/bcachefs.h
index d3d9e965e702..907649033621 100644
--- a/include/trace/events/bcachefs.h
+++ b/include/trace/events/bcachefs.h
@@ -104,10 +104,10 @@ DECLARE_EVENT_CLASS(bio,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev ? bio_dev(bio) : 0;
+		__entry->dev		= bio->bi_disk ? bio_dev(bio) : 0;
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio->bi_iter.bi_size >> 9;
-		blk_fill_rwbs(__entry->rwbs, bio->bi_opf);
+		blk_fill_rwbs(__entry->rwbs, bio->bi_opf, bio->bi_iter.bi_size);
 	),
 
 	TP_printk("%d,%d  %s %llu + %u",
diff --git a/kernel/locking/six.c b/kernel/locking/six.c
index b11660af245b..e2e0acdaef5e 100644
--- a/kernel/locking/six.c
+++ b/kernel/locking/six.c
@@ -18,7 +18,7 @@
 #endif
 
 #define six_acquire(l, t, r)	lock_acquire(l, 0, t, r, 1, NULL, _RET_IP_)
-#define six_release(l)		lock_release(l, _RET_IP_)
+#define six_release(l)		lock_release(l, 0, _RET_IP_)
 
 static void do_six_unlock_type(struct six_lock *lock, enum six_lock_type type);
 
diff --git a/lib/mean_and_variance.c b/lib/mean_and_variance.c
index 643e3113500b..75fe68e13d97 100644
--- a/lib/mean_and_variance.c
+++ b/lib/mean_and_variance.c
@@ -38,7 +38,6 @@
 #include <linux/compiler.h>
 #include <linux/export.h>
 #include <linux/limits.h>
-#include <linux/math.h>
 #include <linux/math64.h>
 #include <linux/mean_and_variance.h>
 #include <linux/module.h>