Merge tag 'loongarch-kvm-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/chenhuacai/linux-loongson into HEAD

LoongArch KVM changes for v6.10 1. Add ParaVirt IPI support. 2. Add software breakpoint support. 3. Add mmio trace events support.
author: Paolo Bonzini <pbonzini@redhat.com> 2024-05-10 10:49:02 -0400
committer: Paolo Bonzini <pbonzini@redhat.com> 2024-05-10 13:20:18 -0400
commit: 4232da23d75d173195c6766729e51947b64f83cd (patch)
tree: 7eb1420b118d452c0b497bf6267ecdfb0a6d5931 /fs
parent: Merge branch 'kvm-sev-es-ghcbv2' into HEAD (diff)
parent: LoongArch: KVM: Add mmio trace events support (diff)
download: wireguard-linux-4232da23d75d173195c6766729e51947b64f83cd.tar.xz
wireguard-linux-4232da23d75d173195c6766729e51947b64f83cd.zip
140 files changed, 1984 insertions, 949 deletions
diff --git a/fs/9p/fid.h b/fs/9p/fid.h
index 29281b7c3887..0d6138bee2a3 100644
--- a/fs/9p/fid.h
+++ b/fs/9p/fid.h
@@ -49,9 +49,6 @@ static inline struct p9_fid *v9fs_fid_clone(struct dentry *dentry)
 static inline void v9fs_fid_add_modes(struct p9_fid *fid, unsigned int s_flags,
 	unsigned int s_cache, unsigned int f_flags)
 {
-	if (fid->qid.type != P9_QTFILE)
-		return;
-
 	if ((!s_cache) ||
 	   ((fid->qid.version == 0) && !(s_flags & V9FS_IGNORE_QV)) ||
 	   (s_flags & V9FS_DIRECT_IO) || (f_flags & O_DIRECT)) {
diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h
index 9defa12208f9..1775fcc7f0e8 100644
--- a/fs/9p/v9fs.h
+++ b/fs/9p/v9fs.h
@@ -179,13 +179,14 @@ extern int v9fs_vfs_rename(struct mnt_idmap *idmap,
 			   struct inode *old_dir, struct dentry *old_dentry,
 			   struct inode *new_dir, struct dentry *new_dentry,
 			   unsigned int flags);
-extern struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid);
+extern struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid,
+						bool new);
 extern const struct inode_operations v9fs_dir_inode_operations_dotl;
 extern const struct inode_operations v9fs_file_inode_operations_dotl;
 extern const struct inode_operations v9fs_symlink_inode_operations_dotl;
 extern const struct netfs_request_ops v9fs_req_ops;
 extern struct inode *v9fs_fid_iget_dotl(struct super_block *sb,
-					struct p9_fid *fid);
+						struct p9_fid *fid, bool new);
 
 /* other default globals */
 #define V9FS_PORT	564
@@ -224,12 +225,12 @@ static inline int v9fs_proto_dotl(struct v9fs_session_info *v9ses)
  */
 static inline struct inode *
 v9fs_get_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-			struct super_block *sb)
+			struct super_block *sb, bool new)
 {
 	if (v9fs_proto_dotl(v9ses))
-		return v9fs_fid_iget_dotl(sb, fid);
+		return v9fs_fid_iget_dotl(sb, fid, new);
 	else
-		return v9fs_fid_iget(sb, fid);
+		return v9fs_fid_iget(sb, fid, new);
 }
 
 #endif
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index abdbbaee5184..348cc90bf9c5 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -520,6 +520,7 @@ const struct file_operations v9fs_file_operations = {
 	.splice_read = v9fs_file_splice_read,
 	.splice_write = iter_file_splice_write,
 	.fsync = v9fs_file_fsync,
+	.setlease = simple_nosetlease,
 };
 
 const struct file_operations v9fs_file_operations_dotl = {
@@ -534,4 +535,5 @@ const struct file_operations v9fs_file_operations_dotl = {
 	.splice_read = v9fs_file_splice_read,
 	.splice_write = iter_file_splice_write,
 	.fsync = v9fs_file_fsync_dotl,
+	.setlease = simple_nosetlease,
 };
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index b01b1bbf2493..7a3308d77606 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -83,7 +83,7 @@ static int p9mode2perm(struct v9fs_session_info *v9ses,
 	int res;
 	int mode = stat->mode;
 
-	res = mode & S_IALLUGO;
+	res = mode & 0777; /* S_IRWXUGO */
 	if (v9fs_proto_dotu(v9ses)) {
 		if ((mode & P9_DMSETUID) == P9_DMSETUID)
 			res |= S_ISUID;
@@ -178,6 +178,9 @@ int v9fs_uflags2omode(int uflags, int extended)
 		break;
 	}
 
+	if (uflags & O_TRUNC)
+		ret |= P9_OTRUNC;
+
 	if (extended) {
 		if (uflags & O_EXCL)
 			ret |= P9_OEXCL;
@@ -361,7 +364,8 @@ void v9fs_evict_inode(struct inode *inode)
 		clear_inode(inode);
 }
 
-struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid)
+struct inode *
+v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid, bool new)
 {
 	dev_t rdev;
 	int retval;
@@ -373,8 +377,18 @@ struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid)
 	inode = iget_locked(sb, QID2INO(&fid->qid));
 	if (unlikely(!inode))
 		return ERR_PTR(-ENOMEM);
-	if (!(inode->i_state & I_NEW))
-		return inode;
+	if (!(inode->i_state & I_NEW)) {
+		if (!new) {
+			goto done;
+		} else {
+			p9_debug(P9_DEBUG_VFS, "WARNING: Inode collision %ld\n",
+						inode->i_ino);
+			iput(inode);
+			remove_inode_hash(inode);
+			inode = iget_locked(sb, QID2INO(&fid->qid));
+			WARN_ON(!(inode->i_state & I_NEW));
+		}
+	}
 
 	/*
 	 * initialize the inode with the stat info
@@ -398,11 +412,11 @@ struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid)
 	v9fs_set_netfs_context(inode);
 	v9fs_cache_inode_get_cookie(inode);
 	unlock_new_inode(inode);
+done:
 	return inode;
 error:
 	iget_failed(inode);
 	return ERR_PTR(retval);
-
 }
 
 /**
@@ -434,8 +448,15 @@ static int v9fs_at_to_dotl_flags(int flags)
  */
 static void v9fs_dec_count(struct inode *inode)
 {
-	if (!S_ISDIR(inode->i_mode) || inode->i_nlink > 2)
-		drop_nlink(inode);
+	if (!S_ISDIR(inode->i_mode) || inode->i_nlink > 2) {
+		if (inode->i_nlink) {
+			drop_nlink(inode);
+		} else {
+			p9_debug(P9_DEBUG_VFS,
+						"WARNING: unexpected i_nlink zero %d inode %ld\n",
+						inode->i_nlink, inode->i_ino);
+		}
+	}
 }
 
 /**
@@ -486,6 +507,9 @@ static int v9fs_remove(struct inode *dir, struct dentry *dentry, int flags)
 		} else
 			v9fs_dec_count(inode);
 
+		if (inode->i_nlink <= 0)	/* no more refs unhash it */
+			remove_inode_hash(inode);
+
 		v9fs_invalidate_inode_attr(inode);
 		v9fs_invalidate_inode_attr(dir);
 
@@ -551,7 +575,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 		/*
 		 * instantiate inode and assign the unopened fid to the dentry
 		 */
-		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb);
+		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb, true);
 		if (IS_ERR(inode)) {
 			err = PTR_ERR(inode);
 			p9_debug(P9_DEBUG_VFS,
@@ -680,7 +704,7 @@ struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 	else if (IS_ERR(fid))
 		inode = ERR_CAST(fid);
 	else
-		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb);
+		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb, false);
 	/*
 	 * If we had a rename on the server and a parallel lookup
 	 * for the new name, then make sure we instantiate with
@@ -1061,8 +1085,6 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
 	struct v9fs_session_info *v9ses = sb->s_fs_info;
 	struct v9fs_inode *v9inode = V9FS_I(inode);
 
-	set_nlink(inode, 1);
-
 	inode_set_atime(inode, stat->atime, 0);
 	inode_set_mtime(inode, stat->mtime, 0);
 	inode_set_ctime(inode, stat->mtime, 0);
diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c
index 55dde186041a..c61b97bd13b9 100644
--- a/fs/9p/vfs_inode_dotl.c
+++ b/fs/9p/vfs_inode_dotl.c
@@ -52,7 +52,10 @@ static kgid_t v9fs_get_fsgid_for_create(struct inode *dir_inode)
 	return current_fsgid();
 }
 
-struct inode *v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid)
+
+
+struct inode *
+v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid, bool new)
 {
 	int retval;
 	struct inode *inode;
@@ -62,8 +65,18 @@ struct inode *v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid)
 	inode = iget_locked(sb, QID2INO(&fid->qid));
 	if (unlikely(!inode))
 		return ERR_PTR(-ENOMEM);
-	if (!(inode->i_state & I_NEW))
-		return inode;
+	if (!(inode->i_state & I_NEW)) {
+		if (!new) {
+			goto done;
+		} else { /* deal with race condition in inode number reuse */
+			p9_debug(P9_DEBUG_ERROR, "WARNING: Inode collision %lx\n",
+						inode->i_ino);
+			iput(inode);
+			remove_inode_hash(inode);
+			inode = iget_locked(sb, QID2INO(&fid->qid));
+			WARN_ON(!(inode->i_state & I_NEW));
+		}
+	}
 
 	/*
 	 * initialize the inode with the stat info
@@ -90,12 +103,11 @@ struct inode *v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid)
 		goto error;
 
 	unlock_new_inode(inode);
-
+done:
 	return inode;
 error:
 	iget_failed(inode);
 	return ERR_PTR(retval);
-
 }
 
 struct dotl_openflag_map {
@@ -247,7 +259,7 @@ v9fs_vfs_atomic_open_dotl(struct inode *dir, struct dentry *dentry,
 		p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
 		goto out;
 	}
-	inode = v9fs_fid_iget_dotl(dir->i_sb, fid);
+	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", err);
@@ -340,7 +352,7 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 	}
 
 	/* instantiate inode and assign the unopened fid to the dentry */
-	inode = v9fs_fid_iget_dotl(dir->i_sb, fid);
+	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
@@ -776,7 +788,7 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 			 err);
 		goto error;
 	}
-	inode = v9fs_fid_iget_dotl(dir->i_sb, fid);
+	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index 4236058c7bbd..f52fdf42945c 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -139,7 +139,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
 	else
 		sb->s_d_op = &v9fs_dentry_operations;
 
-	inode = v9fs_get_inode_from_fid(v9ses, fid, sb);
+	inode = v9fs_get_inode_from_fid(v9ses, fid, sb, true);
 	if (IS_ERR(inode)) {
 		retval = PTR_ERR(inode);
 		goto release_sb;
@@ -244,6 +244,21 @@ done:
 	return res;
 }
 
+static int v9fs_drop_inode(struct inode *inode)
+{
+	struct v9fs_session_info *v9ses;
+
+	v9ses = v9fs_inode2v9ses(inode);
+	if (v9ses->cache & (CACHE_META|CACHE_LOOSE))
+		return generic_drop_inode(inode);
+	/*
+	 * in case of non cached mode always drop the
+	 * inode because we want the inode attribute
+	 * to always match that on the server.
+	 */
+	return 1;
+}
+
 static int v9fs_write_inode(struct inode *inode,
 			    struct writeback_control *wbc)
 {
@@ -268,6 +283,7 @@ static const struct super_operations v9fs_super_ops = {
 	.alloc_inode = v9fs_alloc_inode,
 	.free_inode = v9fs_free_inode,
 	.statfs = simple_statfs,
+	.drop_inode = v9fs_drop_inode,
 	.evict_inode = v9fs_evict_inode,
 	.show_options = v9fs_show_options,
 	.umount_begin = v9fs_umount_begin,
@@ -278,6 +294,7 @@ static const struct super_operations v9fs_super_ops_dotl = {
 	.alloc_inode = v9fs_alloc_inode,
 	.free_inode = v9fs_free_inode,
 	.statfs = v9fs_statfs,
+	.drop_inode = v9fs_drop_inode,
 	.evict_inode = v9fs_evict_inode,
 	.show_options = v9fs_show_options,
 	.umount_begin = v9fs_umount_begin,
diff --git a/fs/bcachefs/acl.c b/fs/bcachefs/acl.c
index 3640f417cce1..5c180fdc3efb 100644
--- a/fs/bcachefs/acl.c
+++ b/fs/bcachefs/acl.c
@@ -281,7 +281,6 @@ struct posix_acl *bch2_get_acl(struct mnt_idmap *idmap,
 	struct xattr_search_key search = X_SEARCH(acl_to_xattr_type(type), "", 0);
 	struct btree_trans *trans = bch2_trans_get(c);
 	struct btree_iter iter = { NULL };
-	struct bkey_s_c_xattr xattr;
 	struct posix_acl *acl = NULL;
 	struct bkey_s_c k;
 	int ret;
@@ -290,28 +289,27 @@ retry:
 
 	ret = bch2_hash_lookup(trans, &iter, bch2_xattr_hash_desc,
 			&hash, inode_inum(inode), &search, 0);
-	if (ret) {
-		if (!bch2_err_matches(ret, ENOENT))
-			acl = ERR_PTR(ret);
-		goto out;
-	}
+	if (ret)
+		goto err;
 
 	k = bch2_btree_iter_peek_slot(&iter);
 	ret = bkey_err(k);
-	if (ret) {
-		acl = ERR_PTR(ret);
-		goto out;
-	}
+	if (ret)
+		goto err;
 
-	xattr = bkey_s_c_to_xattr(k);
+	struct bkey_s_c_xattr xattr = bkey_s_c_to_xattr(k);
 	acl = bch2_acl_from_disk(trans, xattr_val(xattr.v),
-			le16_to_cpu(xattr.v->x_val_len));
+				 le16_to_cpu(xattr.v->x_val_len));
+	ret = PTR_ERR_OR_ZERO(acl);
+err:
+	if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+		goto retry;
 
-	if (!IS_ERR(acl))
+	if (ret)
+		acl = !bch2_err_matches(ret, ENOENT) ? ERR_PTR(ret) : NULL;
+
+	if (!IS_ERR_OR_NULL(acl))
 		set_cached_acl(&inode->v, type, acl);
-out:
-	if (bch2_err_matches(PTR_ERR_OR_ZERO(acl), BCH_ERR_transaction_restart))
-		goto retry;
 
 	bch2_trans_iter_exit(trans, &iter);
 	bch2_trans_put(trans);
diff --git a/fs/bcachefs/backpointers.c b/fs/bcachefs/backpointers.c
index 114328acde72..a20044201002 100644
--- a/fs/bcachefs/backpointers.c
+++ b/fs/bcachefs/backpointers.c
@@ -49,13 +49,15 @@ int bch2_backpointer_invalid(struct bch_fs *c, struct bkey_s_c k,
 	if (!bch2_dev_exists2(c, bp.k->p.inode))
 		return 0;
 
+	struct bch_dev *ca = bch_dev_bkey_exists(c, bp.k->p.inode);
 	struct bpos bucket = bp_pos_to_bucket(c, bp.k->p);
 	int ret = 0;
 
-	bkey_fsck_err_on(!bpos_eq(bp.k->p, bucket_pos_to_bp(c, bucket, bp.v->bucket_offset)),
+	bkey_fsck_err_on((bp.v->bucket_offset >> MAX_EXTENT_COMPRESS_RATIO_SHIFT) >= ca->mi.bucket_size ||
+			 !bpos_eq(bp.k->p, bucket_pos_to_bp(c, bucket, bp.v->bucket_offset)),
 			 c, err,
-			 backpointer_pos_wrong,
-			 "backpointer at wrong pos");
+			 backpointer_bucket_offset_wrong,
+			 "backpointer bucket_offset wrong");
 fsck_err:
 	return ret;
 }
@@ -468,7 +470,7 @@ found:
 		goto err;
 	}
 
-	bio = bio_alloc(ca->disk_sb.bdev, 1, REQ_OP_READ, GFP_KERNEL);
+	bio = bio_alloc(ca->disk_sb.bdev, buf_pages(data_buf, bytes), REQ_OP_READ, GFP_KERNEL);
 	bio->bi_iter.bi_sector = p.ptr.offset;
 	bch2_bio_map(bio, data_buf, bytes);
 	ret = submit_bio_wait(bio);
diff --git a/fs/bcachefs/backpointers.h b/fs/bcachefs/backpointers.h
index da012ca7daee..85949b9fd880 100644
--- a/fs/bcachefs/backpointers.h
+++ b/fs/bcachefs/backpointers.h
@@ -53,14 +53,11 @@ static inline struct bpos bucket_pos_to_bp(const struct bch_fs *c,
 					   u64 bucket_offset)
 {
 	struct bch_dev *ca = bch_dev_bkey_exists(c, bucket.inode);
-	struct bpos ret;
-
-	ret = POS(bucket.inode,
-		  (bucket_to_sector(ca, bucket.offset) <<
-		   MAX_EXTENT_COMPRESS_RATIO_SHIFT) + bucket_offset);
+	struct bpos ret = POS(bucket.inode,
+			      (bucket_to_sector(ca, bucket.offset) <<
+			       MAX_EXTENT_COMPRESS_RATIO_SHIFT) + bucket_offset);
 
 	EBUG_ON(!bkey_eq(bucket, bp_pos_to_bucket(c, ret)));
-
 	return ret;
 }
 
diff --git a/fs/bcachefs/bcachefs.h b/fs/bcachefs/bcachefs.h
index a31a5f706929..91c3c1fef233 100644
--- a/fs/bcachefs/bcachefs.h
+++ b/fs/bcachefs/bcachefs.h
@@ -709,6 +709,8 @@ struct btree_trans_buf {
 	x(stripe_delete)						\
 	x(reflink)							\
 	x(fallocate)							\
+	x(fsync)							\
+	x(dio_write)							\
 	x(discard)							\
 	x(discard_fast)							\
 	x(invalidate)							\
diff --git a/fs/bcachefs/bcachefs_format.h b/fs/bcachefs/bcachefs_format.h
index 63102992d955..f7fbfccd2b1e 100644
--- a/fs/bcachefs/bcachefs_format.h
+++ b/fs/bcachefs/bcachefs_format.h
@@ -578,7 +578,8 @@ struct bch_member {
 	__le64			nbuckets;	/* device size */
 	__le16			first_bucket;   /* index of first bucket used */
 	__le16			bucket_size;	/* sectors */
-	__le32			pad;
+	__u8			btree_bitmap_shift;
+	__u8			pad[3];
 	__le64			last_mount;	/* time_t */
 
 	__le64			flags;
@@ -587,6 +588,7 @@ struct bch_member {
 	__le64			errors_at_reset[BCH_MEMBER_ERROR_NR];
 	__le64			errors_reset_time;
 	__le64			seq;
+	__le64			btree_allocated_bitmap;
 };
 
 #define BCH_MEMBER_V1_BYTES	56
@@ -876,7 +878,8 @@ struct bch_sb_field_downgrade {
 	x(rebalance_work,		BCH_VERSION(1,  3))		\
 	x(member_seq,			BCH_VERSION(1,  4))		\
 	x(subvolume_fs_parent,		BCH_VERSION(1,  5))		\
-	x(btree_subvolume_children,	BCH_VERSION(1,  6))
+	x(btree_subvolume_children,	BCH_VERSION(1,  6))		\
+	x(mi_btree_bitmap,		BCH_VERSION(1,  7))
 
 enum bcachefs_metadata_version {
 	bcachefs_metadata_version_min = 9,
@@ -1314,7 +1317,7 @@ static inline __u64 __bset_magic(struct bch_sb *sb)
 	x(write_buffer_keys,	11)		\
 	x(datetime,		12)
 
-enum {
+enum bch_jset_entry_type {
 #define x(f, nr)	BCH_JSET_ENTRY_##f	= nr,
 	BCH_JSET_ENTRY_TYPES()
 #undef x
@@ -1360,7 +1363,7 @@ struct jset_entry_blacklist_v2 {
 	x(inodes,		1)		\
 	x(key_version,		2)
 
-enum {
+enum bch_fs_usage_type {
 #define x(f, nr)	BCH_FS_USAGE_##f	= nr,
 	BCH_FS_USAGE_TYPES()
 #undef x
@@ -1501,7 +1504,8 @@ enum btree_id_flags {
 	  BIT_ULL(KEY_TYPE_stripe))						\
 	x(reflink,		7,	BTREE_ID_EXTENTS|BTREE_ID_DATA,		\
 	  BIT_ULL(KEY_TYPE_reflink_v)|						\
-	  BIT_ULL(KEY_TYPE_indirect_inline_data))				\
+	  BIT_ULL(KEY_TYPE_indirect_inline_data)|				\
+	  BIT_ULL(KEY_TYPE_error))						\
 	x(subvolumes,		8,	0,					\
 	  BIT_ULL(KEY_TYPE_subvolume))						\
 	x(snapshots,		9,	0,					\
@@ -1535,6 +1539,20 @@ enum btree_id {
 	BTREE_ID_NR
 };
 
+static inline bool btree_id_is_alloc(enum btree_id id)
+{
+	switch (id) {
+	case BTREE_ID_alloc:
+	case BTREE_ID_backpointers:
+	case BTREE_ID_need_discard:
+	case BTREE_ID_freespace:
+	case BTREE_ID_bucket_gens:
+		return true;
+	default:
+		return false;
+	}
+}
+
 #define BTREE_MAX_DEPTH		4U
 
 /* Btree nodes */
diff --git a/fs/bcachefs/bkey.h b/fs/bcachefs/bkey.h
index cf23ff47bed8..3a45d128f608 100644
--- a/fs/bcachefs/bkey.h
+++ b/fs/bcachefs/bkey.h
@@ -314,6 +314,12 @@ static inline unsigned bkeyp_key_u64s(const struct bkey_format *format,
 	return bkey_packed(k) ? format->key_u64s : BKEY_U64s;
 }
 
+static inline bool bkeyp_u64s_valid(const struct bkey_format *f,
+				    const struct bkey_packed *k)
+{
+	return ((unsigned) k->u64s - bkeyp_key_u64s(f, k) <= U8_MAX - BKEY_U64s);
+}
+
 static inline unsigned bkeyp_key_bytes(const struct bkey_format *format,
 				       const struct bkey_packed *k)
 {
diff --git a/fs/bcachefs/bkey_methods.c b/fs/bcachefs/bkey_methods.c
index 5e52684764eb..db336a43fc08 100644
--- a/fs/bcachefs/bkey_methods.c
+++ b/fs/bcachefs/bkey_methods.c
@@ -171,11 +171,15 @@ int __bch2_bkey_invalid(struct bch_fs *c, struct bkey_s_c k,
 	if (type >= BKEY_TYPE_NR)
 		return 0;
 
-	bkey_fsck_err_on((flags & BKEY_INVALID_COMMIT) &&
+	bkey_fsck_err_on((type == BKEY_TYPE_btree ||
+			  (flags & BKEY_INVALID_COMMIT)) &&
 			 !(bch2_key_types_allowed[type] & BIT_ULL(k.k->type)), c, err,
 			 bkey_invalid_type_for_btree,
 			 "invalid key type for btree %s (%s)",
-			 bch2_btree_node_type_str(type), bch2_bkey_types[k.k->type]);
+			 bch2_btree_node_type_str(type),
+			 k.k->type < KEY_TYPE_MAX
+			 ? bch2_bkey_types[k.k->type]
+			 : "(unknown)");
 
 	if (btree_node_type_is_extents(type) && !bkey_whiteout(k.k)) {
 		bkey_fsck_err_on(k.k->size == 0, c, err,
diff --git a/fs/bcachefs/btree_cache.c b/fs/bcachefs/btree_cache.c
index 84474324dba9..02c70e813fac 100644
--- a/fs/bcachefs/btree_cache.c
+++ b/fs/bcachefs/btree_cache.c
@@ -709,9 +709,31 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans,
 	struct bch_fs *c = trans->c;
 	struct btree_cache *bc = &c->btree_cache;
 	struct btree *b;
-	u32 seq;
 
-	BUG_ON(level + 1 >= BTREE_MAX_DEPTH);
+	if (unlikely(level >= BTREE_MAX_DEPTH)) {
+		int ret = bch2_fs_topology_error(c, "attempting to get btree node at level %u, >= max depth %u",
+						 level, BTREE_MAX_DEPTH);
+		return ERR_PTR(ret);
+	}
+
+	if (unlikely(!bkey_is_btree_ptr(&k->k))) {
+		struct printbuf buf = PRINTBUF;
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(k));
+
+		int ret = bch2_fs_topology_error(c, "attempting to get btree node with non-btree key %s", buf.buf);
+		printbuf_exit(&buf);
+		return ERR_PTR(ret);
+	}
+
+	if (unlikely(k->k.u64s > BKEY_BTREE_PTR_U64s_MAX)) {
+		struct printbuf buf = PRINTBUF;
+		bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(k));
+
+		int ret = bch2_fs_topology_error(c, "attempting to get btree node with too big key %s", buf.buf);
+		printbuf_exit(&buf);
+		return ERR_PTR(ret);
+	}
+
 	/*
 	 * Parent node must be locked, else we could read in a btree node that's
 	 * been freed:
@@ -752,34 +774,26 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans,
 	}
 
 	set_btree_node_read_in_flight(b);
-
 	six_unlock_write(&b->c.lock);
-	seq = six_lock_seq(&b->c.lock);
-	six_unlock_intent(&b->c.lock);
 
-	/* Unlock before doing IO: */
-	if (path && sync)
-		bch2_trans_unlock_noassert(trans);
-
-	bch2_btree_node_read(trans, b, sync);
+	if (path) {
+		u32 seq = six_lock_seq(&b->c.lock);
 
-	if (!sync)
-		return NULL;
+		/* Unlock before doing IO: */
+		six_unlock_intent(&b->c.lock);
+		bch2_trans_unlock_noassert(trans);
 
-	if (path) {
-		int ret = bch2_trans_relock(trans) ?:
-			bch2_btree_path_relock_intent(trans, path);
-		if (ret) {
-			BUG_ON(!trans->restarted);
-			return ERR_PTR(ret);
-		}
-	}
+		bch2_btree_node_read(trans, b, sync);
 
-	if (!six_relock_type(&b->c.lock, lock_type, seq)) {
-		BUG_ON(!path);
+		if (!sync)
+			return NULL;
 
-		trace_and_count(c, trans_restart_relock_after_fill, trans, _THIS_IP_, path);
-		return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_relock_after_fill));
+		if (!six_relock_type(&b->c.lock, lock_type, seq))
+			b = NULL;
+	} else {
+		bch2_btree_node_read(trans, b, sync);
+		if (lock_type == SIX_LOCK_read)
+			six_lock_downgrade(&b->c.lock);
 	}
 
 	return b;
@@ -1112,18 +1126,19 @@ int bch2_btree_node_prefetch(struct btree_trans *trans,
 {
 	struct bch_fs *c = trans->c;
 	struct btree_cache *bc = &c->btree_cache;
-	struct btree *b;
 
 	BUG_ON(path && !btree_node_locked(path, level + 1));
 	BUG_ON(level >= BTREE_MAX_DEPTH);
 
-	b = btree_cache_find(bc, k);
+	struct btree *b = btree_cache_find(bc, k);
 	if (b)
 		return 0;
 
 	b = bch2_btree_node_fill(trans, path, k, btree_id,
 				 level, SIX_LOCK_read, false);
-	return PTR_ERR_OR_ZERO(b);
+	if (!IS_ERR_OR_NULL(b))
+		six_unlock_read(&b->c.lock);
+	return bch2_trans_relock(trans) ?: PTR_ERR_OR_ZERO(b);
 }
 
 void bch2_btree_node_evict(struct btree_trans *trans, const struct bkey_i *k)
@@ -1148,6 +1163,8 @@ wait_on_io:
 
 	btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_intent);
 	btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_write);
+	if (unlikely(b->hash_val != btree_ptr_hash_val(k)))
+		goto out;
 
 	if (btree_node_dirty(b)) {
 		__bch2_btree_node_write(c, b, BTREE_WRITE_cache_reclaim);
@@ -1162,7 +1179,7 @@ wait_on_io:
 	btree_node_data_free(c, b);
 	bch2_btree_node_hash_remove(bc, b);
 	mutex_unlock(&bc->lock);
-
+out:
 	six_unlock_write(&b->c.lock);
 	six_unlock_intent(&b->c.lock);
 }
diff --git a/fs/bcachefs/btree_gc.c b/fs/bcachefs/btree_gc.c
index 6280da1244b5..791470b0c654 100644
--- a/fs/bcachefs/btree_gc.c
+++ b/fs/bcachefs/btree_gc.c
@@ -368,11 +368,16 @@ again:
 				buf.buf)) {
 			bch2_btree_node_evict(trans, cur_k.k);
 			cur = NULL;
-			ret =   bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes) ?:
-				bch2_journal_key_delete(c, b->c.btree_id,
-							b->c.level, cur_k.k->k.p);
+			ret = bch2_journal_key_delete(c, b->c.btree_id,
+						      b->c.level, cur_k.k->k.p);
 			if (ret)
 				break;
+
+			if (!btree_id_is_alloc(b->c.btree_id)) {
+				ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes);
+				if (ret)
+					break;
+			}
 			continue;
 		}
 
@@ -544,12 +549,12 @@ reconstruct_root:
 				bch2_btree_root_alloc_fake(c, i, 0);
 			} else {
 				bch2_btree_root_alloc_fake(c, i, 1);
+				bch2_shoot_down_journal_keys(c, i, 1, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
 				ret = bch2_get_scanned_nodes(c, i, 0, POS_MIN, SPOS_MAX);
 				if (ret)
 					break;
 			}
 
-			bch2_shoot_down_journal_keys(c, i, 1, BTREE_MAX_DEPTH, POS_MIN, SPOS_MAX);
 			reconstructed_root = true;
 		}
 
@@ -823,6 +828,7 @@ static int bch2_gc_mark_key(struct btree_trans *trans, enum btree_id btree_id,
 	struct bch_fs *c = trans->c;
 	struct bkey deleted = KEY(0, 0, 0);
 	struct bkey_s_c old = (struct bkey_s_c) { &deleted, NULL };
+	struct printbuf buf = PRINTBUF;
 	int ret = 0;
 
 	deleted.p = k->k->p;
@@ -843,11 +849,23 @@ static int bch2_gc_mark_key(struct btree_trans *trans, enum btree_id btree_id,
 	if (ret)
 		goto err;
 
+	if (mustfix_fsck_err_on(level && !bch2_dev_btree_bitmap_marked(c, *k),
+				c, btree_bitmap_not_marked,
+				"btree ptr not marked in member info btree allocated bitmap\n  %s",
+				(bch2_bkey_val_to_text(&buf, c, *k),
+				 buf.buf))) {
+		mutex_lock(&c->sb_lock);
+		bch2_dev_btree_bitmap_mark(c, *k);
+		bch2_write_super(c);
+		mutex_unlock(&c->sb_lock);
+	}
+
 	ret = commit_do(trans, NULL, NULL, 0,
 			bch2_key_trigger(trans, btree_id, level, old,
 					 unsafe_bkey_s_c_to_s(*k), BTREE_TRIGGER_GC));
 fsck_err:
 err:
+	printbuf_exit(&buf);
 	bch_err_fn(c, ret);
 	return ret;
 }
@@ -1569,7 +1587,7 @@ static int bch2_gc_write_reflink_key(struct btree_trans *trans,
 		struct bkey_i *new = bch2_bkey_make_mut_noupdate(trans, k);
 		ret = PTR_ERR_OR_ZERO(new);
 		if (ret)
-			return ret;
+			goto out;
 
 		if (!r->refcount)
 			new->k.type = KEY_TYPE_deleted;
@@ -1577,6 +1595,7 @@ static int bch2_gc_write_reflink_key(struct btree_trans *trans,
 			*bkey_refcount(bkey_i_to_s(new)) = cpu_to_le64(r->refcount);
 		ret = bch2_trans_update(trans, iter, new, 0);
 	}
+out:
 fsck_err:
 	printbuf_exit(&buf);
 	return ret;
diff --git a/fs/bcachefs/btree_io.c b/fs/bcachefs/btree_io.c
index d7de82ac3893..debb0edc3455 100644
--- a/fs/bcachefs/btree_io.c
+++ b/fs/bcachefs/btree_io.c
@@ -831,7 +831,7 @@ static int bset_key_invalid(struct bch_fs *c, struct btree *b,
 		(rw == WRITE ? bch2_bkey_val_invalid(c, k, READ, err) : 0);
 }
 
-static bool __bkey_valid(struct bch_fs *c, struct btree *b,
+static bool bkey_packed_valid(struct bch_fs *c, struct btree *b,
 			 struct bset *i, struct bkey_packed *k)
 {
 	if (bkey_p_next(k) > vstruct_last(i))
@@ -840,7 +840,7 @@ static bool __bkey_valid(struct bch_fs *c, struct btree *b,
 	if (k->format > KEY_FORMAT_CURRENT)
 		return false;
 
-	if (k->u64s < bkeyp_key_u64s(&b->format, k))
+	if (!bkeyp_u64s_valid(&b->format, k))
 		return false;
 
 	struct printbuf buf = PRINTBUF;
@@ -884,11 +884,13 @@ static int validate_bset_keys(struct bch_fs *c, struct btree *b,
 				 "invalid bkey format %u", k->format))
 			goto drop_this_key;
 
-		if (btree_err_on(k->u64s < bkeyp_key_u64s(&b->format, k),
+		if (btree_err_on(!bkeyp_u64s_valid(&b->format, k),
 				 -BCH_ERR_btree_node_read_err_fixable,
 				 c, NULL, b, i,
 				 btree_node_bkey_bad_u64s,
-				 "k->u64s too small (%u < %u)", k->u64s, bkeyp_key_u64s(&b->format, k)))
+				 "bad k->u64s %u (min %u max %zu)", k->u64s,
+				 bkeyp_key_u64s(&b->format, k),
+				 U8_MAX - BKEY_U64s + bkeyp_key_u64s(&b->format, k)))
 			goto drop_this_key;
 
 		if (!write)
@@ -947,13 +949,12 @@ drop_this_key:
 			 * do
 			 */
 
-			if (!__bkey_valid(c, b, i, (void *) ((u64 *) k + next_good_key))) {
+			if (!bkey_packed_valid(c, b, i, (void *) ((u64 *) k + next_good_key))) {
 				for (next_good_key = 1;
 				     next_good_key < (u64 *) vstruct_last(i) - (u64 *) k;
 				     next_good_key++)
-					if (__bkey_valid(c, b, i, (void *) ((u64 *) k + next_good_key)))
+					if (bkey_packed_valid(c, b, i, (void *) ((u64 *) k + next_good_key)))
 						goto got_good_key;
-
 			}
 
 			/*
@@ -1339,7 +1340,9 @@ start:
 			       rb->start_time);
 	bio_put(&rb->bio);
 
-	if (saw_error && !btree_node_read_error(b)) {
+	if (saw_error &&
+	    !btree_node_read_error(b) &&
+	    c->curr_recovery_pass != BCH_RECOVERY_PASS_scan_for_btree_nodes) {
 		printbuf_reset(&buf);
 		bch2_bpos_to_text(&buf, b->key.k.p);
 		bch_err_ratelimited(c, "%s: rewriting btree node at btree=%s level=%u %s due to error",
diff --git a/fs/bcachefs/btree_iter.h b/fs/bcachefs/btree_iter.h
index 24772538e4cc..1c70836dd7cc 100644
--- a/fs/bcachefs/btree_iter.h
+++ b/fs/bcachefs/btree_iter.h
@@ -498,8 +498,13 @@ static inline void set_btree_iter_dontneed(struct btree_iter *iter)
 {
 	struct btree_trans *trans = iter->trans;
 
-	if (!trans->restarted)
-		btree_iter_path(trans, iter)->preserve = false;
+	if (!iter->path || trans->restarted)
+		return;
+
+	struct btree_path *path = btree_iter_path(trans, iter);
+	path->preserve		= false;
+	if (path->ref == 1)
+		path->should_be_locked	= false;
 }
 
 void *__bch2_trans_kmalloc(struct btree_trans *, size_t);
@@ -642,7 +647,7 @@ int __bch2_btree_trans_too_many_iters(struct btree_trans *);
 
 static inline int btree_trans_too_many_iters(struct btree_trans *trans)
 {
-	if (bitmap_weight(trans->paths_allocated, trans->nr_paths) > BTREE_ITER_INITIAL - 8)
+	if (bitmap_weight(trans->paths_allocated, trans->nr_paths) > BTREE_ITER_NORMAL_LIMIT - 8)
 		return __bch2_btree_trans_too_many_iters(trans);
 
 	return 0;
diff --git a/fs/bcachefs/btree_journal_iter.c b/fs/bcachefs/btree_journal_iter.c
index 5cbcbfe85235..1e8cf49a6935 100644
--- a/fs/bcachefs/btree_journal_iter.c
+++ b/fs/bcachefs/btree_journal_iter.c
@@ -130,12 +130,30 @@ struct bkey_i *bch2_journal_keys_peek_slot(struct bch_fs *c, enum btree_id btree
 	return bch2_journal_keys_peek_upto(c, btree_id, level, pos, pos, &idx);
 }
 
+static void journal_iter_verify(struct journal_iter *iter)
+{
+	struct journal_keys *keys = iter->keys;
+	size_t gap_size = keys->size - keys->nr;
+
+	BUG_ON(iter->idx >= keys->gap &&
+	       iter->idx <  keys->gap + gap_size);
+
+	if (iter->idx < keys->size) {
+		struct journal_key *k = keys->data + iter->idx;
+
+		int cmp = cmp_int(k->btree_id,	iter->btree_id) ?:
+			  cmp_int(k->level,	iter->level);
+		BUG_ON(cmp < 0);
+	}
+}
+
 static void journal_iters_fix(struct bch_fs *c)
 {
 	struct journal_keys *keys = &c->journal_keys;
 	/* The key we just inserted is immediately before the gap: */
 	size_t gap_end = keys->gap + (keys->size - keys->nr);
-	struct btree_and_journal_iter *iter;
+	struct journal_key *new_key = &keys->data[keys->gap - 1];
+	struct journal_iter *iter;
 
 	/*
 	 * If an iterator points one after the key we just inserted, decrement
@@ -143,9 +161,14 @@ static void journal_iters_fix(struct bch_fs *c)
 	 * decrement was unnecessary, bch2_btree_and_journal_iter_peek() will
 	 * handle that:
 	 */
-	list_for_each_entry(iter, &c->journal_iters, journal.list)
-		if (iter->journal.idx == gap_end)
-			iter->journal.idx = keys->gap - 1;
+	list_for_each_entry(iter, &c->journal_iters, list) {
+		journal_iter_verify(iter);
+		if (iter->idx		== gap_end &&
+		    new_key->btree_id	== iter->btree_id &&
+		    new_key->level	== iter->level)
+			iter->idx = keys->gap - 1;
+		journal_iter_verify(iter);
+	}
 }
 
 static void journal_iters_move_gap(struct bch_fs *c, size_t old_gap, size_t new_gap)
@@ -192,7 +215,12 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id,
 	if (idx > keys->gap)
 		idx -= keys->size - keys->nr;
 
+	size_t old_gap = keys->gap;
+
 	if (keys->nr == keys->size) {
+		journal_iters_move_gap(c, old_gap, keys->size);
+		old_gap = keys->size;
+
 		struct journal_keys new_keys = {
 			.nr			= keys->nr,
 			.size			= max_t(size_t, keys->size, 8) * 2,
@@ -216,7 +244,7 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id,
 		keys->gap	= keys->nr;
 	}
 
-	journal_iters_move_gap(c, keys->gap, idx);
+	journal_iters_move_gap(c, old_gap, idx);
 
 	move_gap(keys, idx);
 
@@ -301,16 +329,21 @@ static void bch2_journal_iter_advance(struct journal_iter *iter)
 
 static struct bkey_s_c bch2_journal_iter_peek(struct journal_iter *iter)
 {
-	struct journal_key *k = iter->keys->data + iter->idx;
+	journal_iter_verify(iter);
+
+	while (iter->idx < iter->keys->size) {
+		struct journal_key *k = iter->keys->data + iter->idx;
+
+		int cmp = cmp_int(k->btree_id,	iter->btree_id) ?:
+			  cmp_int(k->level,	iter->level);
+		if (cmp > 0)
+			break;
+		BUG_ON(cmp);
 
-	while (k < iter->keys->data + iter->keys->size &&
-	       k->btree_id	== iter->btree_id &&
-	       k->level		== iter->level) {
 		if (!k->overwritten)
 			return bkey_i_to_s_c(k->k);
 
 		bch2_journal_iter_advance(iter);
-		k = iter->keys->data + iter->idx;
 	}
 
 	return bkey_s_c_null;
@@ -330,6 +363,8 @@ static void bch2_journal_iter_init(struct bch_fs *c,
 	iter->level	= level;
 	iter->keys	= &c->journal_keys;
 	iter->idx	= bch2_journal_key_search(&c->journal_keys, id, level, pos);
+
+	journal_iter_verify(iter);
 }
 
 static struct bkey_s_c bch2_journal_iter_peek_btree(struct btree_and_journal_iter *iter)
@@ -434,10 +469,15 @@ void __bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *trans,
 	iter->trans = trans;
 	iter->b = b;
 	iter->node_iter = node_iter;
-	bch2_journal_iter_init(trans->c, &iter->journal, b->c.btree_id, b->c.level, pos);
-	INIT_LIST_HEAD(&iter->journal.list);
 	iter->pos = b->data->min_key;
 	iter->at_end = false;
+	INIT_LIST_HEAD(&iter->journal.list);
+
+	if (trans->journal_replay_not_finished) {
+		bch2_journal_iter_init(trans->c, &iter->journal, b->c.btree_id, b->c.level, pos);
+		if (!test_bit(BCH_FS_may_go_rw, &trans->c->flags))
+			list_add(&iter->journal.list, &trans->c->journal_iters);
+	}
 }
 
 /*
@@ -452,9 +492,6 @@ void bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *trans,
 
 	bch2_btree_node_iter_init_from_start(&node_iter, b);
 	__bch2_btree_and_journal_iter_init_node_iter(trans, iter, b, node_iter, b->data->min_key);
-	if (trans->journal_replay_not_finished &&
-	    !test_bit(BCH_FS_may_go_rw, &trans->c->flags))
-		list_add(&iter->journal.list, &trans->c->journal_iters);
 }
 
 /* sort and dedup all keys in the journal: */
diff --git a/fs/bcachefs/btree_key_cache.c b/fs/bcachefs/btree_key_cache.c
index 581edcb0911b..e8c1c530cd95 100644
--- a/fs/bcachefs/btree_key_cache.c
+++ b/fs/bcachefs/btree_key_cache.c
@@ -169,6 +169,7 @@ static void bkey_cached_move_to_freelist(struct btree_key_cache *bc,
 	} else {
 		mutex_lock(&bc->lock);
 		list_move_tail(&ck->list, &bc->freed_pcpu);
+		bc->nr_freed_pcpu++;
 		mutex_unlock(&bc->lock);
 	}
 }
@@ -245,6 +246,7 @@ bkey_cached_alloc(struct btree_trans *trans, struct btree_path *path,
 		if (!list_empty(&bc->freed_pcpu)) {
 			ck = list_last_entry(&bc->freed_pcpu, struct bkey_cached, list);
 			list_del_init(&ck->list);
+			bc->nr_freed_pcpu--;
 		}
 		mutex_unlock(&bc->lock);
 	}
@@ -659,7 +661,7 @@ static int btree_key_cache_flush_pos(struct btree_trans *trans,
 		commit_flags |= BCH_WATERMARK_reclaim;
 
 	if (ck->journal.seq != journal_last_seq(j) ||
-	    j->watermark == BCH_WATERMARK_stripe)
+	    !test_bit(JOURNAL_SPACE_LOW, &c->journal.flags))
 		commit_flags |= BCH_TRANS_COMMIT_no_journal_res;
 
 	ret   = bch2_btree_iter_traverse(&b_iter) ?:
@@ -840,8 +842,6 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 	 * Newest freed entries are at the end of the list - once we hit one
 	 * that's too new to be freed, we can bail out:
 	 */
-	scanned += bc->nr_freed_nonpcpu;
-
 	list_for_each_entry_safe(ck, t, &bc->freed_nonpcpu, list) {
 		if (!poll_state_synchronize_srcu(&c->btree_trans_barrier,
 						 ck->btree_trans_barrier_seq))
@@ -855,11 +855,6 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 		bc->nr_freed_nonpcpu--;
 	}
 
-	if (scanned >= nr)
-		goto out;
-
-	scanned += bc->nr_freed_pcpu;
-
 	list_for_each_entry_safe(ck, t, &bc->freed_pcpu, list) {
 		if (!poll_state_synchronize_srcu(&c->btree_trans_barrier,
 						 ck->btree_trans_barrier_seq))
@@ -873,9 +868,6 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 		bc->nr_freed_pcpu--;
 	}
 
-	if (scanned >= nr)
-		goto out;
-
 	rcu_read_lock();
 	tbl = rht_dereference_rcu(bc->table.tbl, &bc->table);
 	if (bc->shrink_iter >= tbl->size)
@@ -891,12 +883,12 @@ static unsigned long bch2_btree_key_cache_scan(struct shrinker *shrink,
 			next = rht_dereference_bucket_rcu(pos->next, tbl, bc->shrink_iter);
 			ck = container_of(pos, struct bkey_cached, hash);
 
-			if (test_bit(BKEY_CACHED_DIRTY, &ck->flags))
+			if (test_bit(BKEY_CACHED_DIRTY, &ck->flags)) {
 				goto next;
-
-			if (test_bit(BKEY_CACHED_ACCESSED, &ck->flags))
+			} else if (test_bit(BKEY_CACHED_ACCESSED, &ck->flags)) {
 				clear_bit(BKEY_CACHED_ACCESSED, &ck->flags);
-			else if (bkey_cached_lock_for_evict(ck)) {
+				goto next;
+			} else if (bkey_cached_lock_for_evict(ck)) {
 				bkey_cached_evict(bc, ck);
 				bkey_cached_free(bc, ck);
 			}
@@ -914,7 +906,6 @@ next:
 	} while (scanned < nr && bc->shrink_iter != start);
 
 	rcu_read_unlock();
-out:
 	memalloc_nofs_restore(flags);
 	srcu_read_unlock(&c->btree_trans_barrier, srcu_idx);
 	mutex_unlock(&bc->lock);
diff --git a/fs/bcachefs/btree_locking.c b/fs/bcachefs/btree_locking.c
index b9b151e693ed..f2caf491957e 100644
--- a/fs/bcachefs/btree_locking.c
+++ b/fs/bcachefs/btree_locking.c
@@ -440,33 +440,7 @@ void bch2_btree_node_lock_write_nofail(struct btree_trans *trans,
 				       struct btree_path *path,
 				       struct btree_bkey_cached_common *b)
 {
-	struct btree_path *linked;
-	unsigned i, iter;
-	int ret;
-
-	/*
-	 * XXX BIG FAT NOTICE
-	 *
-	 * Drop all read locks before taking a write lock:
-	 *
-	 * This is a hack, because bch2_btree_node_lock_write_nofail() is a
-	 * hack - but by dropping read locks first, this should never fail, and
-	 * we only use this in code paths where whatever read locks we've
-	 * already taken are no longer needed:
-	 */
-
-	trans_for_each_path(trans, linked, iter) {
-		if (!linked->nodes_locked)
-			continue;
-
-		for (i = 0; i < BTREE_MAX_DEPTH; i++)
-			if (btree_node_read_locked(linked, i)) {
-				btree_node_unlock(trans, linked, i);
-				btree_path_set_dirty(linked, BTREE_ITER_NEED_RELOCK);
-			}
-	}
-
-	ret = __btree_node_lock_write(trans, path, b, true);
+	int ret = __btree_node_lock_write(trans, path, b, true);
 	BUG_ON(ret);
 }
 
diff --git a/fs/bcachefs/btree_node_scan.c b/fs/bcachefs/btree_node_scan.c
index 3f33be7e5e5c..45cb8149d374 100644
--- a/fs/bcachefs/btree_node_scan.c
+++ b/fs/bcachefs/btree_node_scan.c
@@ -57,13 +57,14 @@ static void found_btree_node_to_key(struct bkey_i *k, const struct found_btree_n
 	bp->v.seq		= cpu_to_le64(f->cookie);
 	bp->v.sectors_written	= 0;
 	bp->v.flags		= 0;
+	bp->v.sectors_written	= cpu_to_le16(f->sectors_written);
 	bp->v.min_key		= f->min_key;
 	SET_BTREE_PTR_RANGE_UPDATED(&bp->v, f->range_updated);
 	memcpy(bp->v.start, f->ptrs, sizeof(struct bch_extent_ptr) * f->nr_ptrs);
 }
 
 static bool found_btree_node_is_readable(struct btree_trans *trans,
-					 const struct found_btree_node *f)
+					 struct found_btree_node *f)
 {
 	struct { __BKEY_PADDED(k, BKEY_BTREE_PTR_VAL_U64s_MAX); } k;
 
@@ -71,8 +72,10 @@ static bool found_btree_node_is_readable(struct btree_trans *trans,
 
 	struct btree *b = bch2_btree_node_get_noiter(trans, &k.k, f->btree_id, f->level, false);
 	bool ret = !IS_ERR_OR_NULL(b);
-	if (ret)
+	if (ret) {
+		f->sectors_written = b->written;
 		six_unlock_read(&b->c.lock);
+	}
 
 	/*
 	 * We might update this node's range; if that happens, we need the node
@@ -133,6 +136,19 @@ static void try_read_btree_node(struct find_btree_nodes *f, struct bch_dev *ca,
 	if (le64_to_cpu(bn->magic) != bset_magic(c))
 		return;
 
+	if (bch2_csum_type_is_encryption(BSET_CSUM_TYPE(&bn->keys))) {
+		struct nonce nonce = btree_nonce(&bn->keys, 0);
+		unsigned bytes = (void *) &bn->keys - (void *) &bn->flags;
+
+		bch2_encrypt(c, BSET_CSUM_TYPE(&bn->keys), nonce, &bn->flags, bytes);
+	}
+
+	if (btree_id_is_alloc(BTREE_NODE_ID(bn)))
+		return;
+
+	if (BTREE_NODE_LEVEL(bn) >= BTREE_MAX_DEPTH)
+		return;
+
 	rcu_read_lock();
 	struct found_btree_node n = {
 		.btree_id	= BTREE_NODE_ID(bn),
@@ -192,8 +208,13 @@ static int read_btree_nodes_worker(void *p)
 				last_print = jiffies;
 			}
 
-			try_read_btree_node(w->f, ca, bio, buf,
-					    bucket * ca->mi.bucket_size + bucket_offset);
+			u64 sector = bucket * ca->mi.bucket_size + bucket_offset;
+
+			if (c->sb.version_upgrade_complete >= bcachefs_metadata_version_mi_btree_bitmap &&
+			    !bch2_dev_btree_bitmap_marked_sectors(ca, sector, btree_sectors(c)))
+				continue;
+
+			try_read_btree_node(w->f, ca, bio, buf, sector);
 		}
 err:
 	bio_put(bio);
@@ -213,6 +234,9 @@ static int read_btree_nodes(struct find_btree_nodes *f)
 	closure_init_stack(&cl);
 
 	for_each_online_member(c, ca) {
+		if (!(ca->mi.data_allowed & BIT(BCH_DATA_btree)))
+			continue;
+
 		struct find_btree_nodes_worker *w = kmalloc(sizeof(*w), GFP_KERNEL);
 		struct task_struct *t;
 
@@ -281,6 +305,8 @@ again:
 
 			start->max_key = bpos_predecessor(n->min_key);
 			start->range_updated = true;
+		} else if (n->level) {
+			n->overwritten = true;
 		} else {
 			struct printbuf buf = PRINTBUF;
 
@@ -290,7 +316,7 @@ again:
 			found_btree_node_to_text(&buf, c, n);
 			bch_err(c, "%s", buf.buf);
 			printbuf_exit(&buf);
-			return -1;
+			return -BCH_ERR_fsck_repair_unimplemented;
 		}
 	}
 
@@ -436,6 +462,9 @@ bool bch2_btree_has_scanned_nodes(struct bch_fs *c, enum btree_id btree)
 int bch2_get_scanned_nodes(struct bch_fs *c, enum btree_id btree,
 			   unsigned level, struct bpos node_min, struct bpos node_max)
 {
+	if (btree_id_is_alloc(btree))
+		return 0;
+
 	struct find_btree_nodes *f = &c->found_btree_nodes;
 
 	int ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_scan_for_btree_nodes);
diff --git a/fs/bcachefs/btree_node_scan_types.h b/fs/bcachefs/btree_node_scan_types.h
index abb7b27d556a..5cfaeb5ac831 100644
--- a/fs/bcachefs/btree_node_scan_types.h
+++ b/fs/bcachefs/btree_node_scan_types.h
@@ -9,6 +9,7 @@ struct found_btree_node {
 	bool			overwritten:1;
 	u8			btree_id;
 	u8			level;
+	unsigned		sectors_written;
 	u32			seq;
 	u64			cookie;
 
diff --git a/fs/bcachefs/btree_trans_commit.c b/fs/bcachefs/btree_trans_commit.c
index aa9da4970740..bbec91e8e650 100644
--- a/fs/bcachefs/btree_trans_commit.c
+++ b/fs/bcachefs/btree_trans_commit.c
@@ -397,12 +397,13 @@ static int btree_key_can_insert_cached(struct btree_trans *trans, unsigned flags
 	struct bkey_cached *ck = (void *) path->l[0].b;
 	unsigned new_u64s;
 	struct bkey_i *new_k;
+	unsigned watermark = flags & BCH_WATERMARK_MASK;
 
 	EBUG_ON(path->level);
 
-	if (!test_bit(BKEY_CACHED_DIRTY, &ck->flags) &&
-	    bch2_btree_key_cache_must_wait(c) &&
-	    !(flags & BCH_TRANS_COMMIT_journal_reclaim))
+	if (watermark < BCH_WATERMARK_reclaim &&
+	    !test_bit(BKEY_CACHED_DIRTY, &ck->flags) &&
+	    bch2_btree_key_cache_must_wait(c))
 		return -BCH_ERR_btree_insert_need_journal_reclaim;
 
 	/*
@@ -499,9 +500,8 @@ static int run_one_trans_trigger(struct btree_trans *trans, struct btree_insert_
 }
 
 static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
-			      struct btree_insert_entry *btree_id_start)
+			      unsigned btree_id_start)
 {
-	struct btree_insert_entry *i;
 	bool trans_trigger_run;
 	int ret, overwrite;
 
@@ -514,13 +514,13 @@ static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
 		do {
 			trans_trigger_run = false;
 
-			for (i = btree_id_start;
-			     i < trans->updates + trans->nr_updates && i->btree_id <= btree_id;
+			for (unsigned i = btree_id_start;
+			     i < trans->nr_updates && trans->updates[i].btree_id <= btree_id;
 			     i++) {
-				if (i->btree_id != btree_id)
+				if (trans->updates[i].btree_id != btree_id)
 					continue;
 
-				ret = run_one_trans_trigger(trans, i, overwrite);
+				ret = run_one_trans_trigger(trans, trans->updates + i, overwrite);
 				if (ret < 0)
 					return ret;
 				if (ret)
@@ -534,8 +534,7 @@ static int run_btree_triggers(struct btree_trans *trans, enum btree_id btree_id,
 
 static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 {
-	struct btree_insert_entry *btree_id_start = trans->updates;
-	unsigned btree_id = 0;
+	unsigned btree_id = 0, btree_id_start = 0;
 	int ret = 0;
 
 	/*
@@ -549,8 +548,8 @@ static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 		if (btree_id == BTREE_ID_alloc)
 			continue;
 
-		while (btree_id_start < trans->updates + trans->nr_updates &&
-		       btree_id_start->btree_id < btree_id)
+		while (btree_id_start < trans->nr_updates &&
+		       trans->updates[btree_id_start].btree_id < btree_id)
 			btree_id_start++;
 
 		ret = run_btree_triggers(trans, btree_id, btree_id_start);
@@ -558,11 +557,13 @@ static int bch2_trans_commit_run_triggers(struct btree_trans *trans)
 			return ret;
 	}
 
-	trans_for_each_update(trans, i) {
+	for (unsigned idx = 0; idx < trans->nr_updates; idx++) {
+		struct btree_insert_entry *i = trans->updates + idx;
+
 		if (i->btree_id > BTREE_ID_alloc)
 			break;
 		if (i->btree_id == BTREE_ID_alloc) {
-			ret = run_btree_triggers(trans, BTREE_ID_alloc, i);
+			ret = run_btree_triggers(trans, BTREE_ID_alloc, idx);
 			if (ret)
 				return ret;
 			break;
@@ -826,7 +827,8 @@ static inline int do_bch2_trans_commit(struct btree_trans *trans, unsigned flags
 	struct bch_fs *c = trans->c;
 	int ret = 0, u64s_delta = 0;
 
-	trans_for_each_update(trans, i) {
+	for (unsigned idx = 0; idx < trans->nr_updates; idx++) {
+		struct btree_insert_entry *i = trans->updates + idx;
 		if (i->cached)
 			continue;
 
diff --git a/fs/bcachefs/btree_types.h b/fs/bcachefs/btree_types.h
index 9404d96c38f3..c69b233c41bb 100644
--- a/fs/bcachefs/btree_types.h
+++ b/fs/bcachefs/btree_types.h
@@ -321,9 +321,9 @@ struct bkey_cached {
 	struct btree_bkey_cached_common c;
 
 	unsigned long		flags;
+	unsigned long		btree_trans_barrier_seq;
 	u16			u64s;
 	bool			valid;
-	u32			btree_trans_barrier_seq;
 	struct bkey_cached_key	key;
 
 	struct rhash_head	hash;
@@ -364,7 +364,21 @@ struct btree_insert_entry {
 	unsigned long		ip_allocated;
 };
 
+/* Number of btree paths we preallocate, usually enough */
 #define BTREE_ITER_INITIAL		64
+/*
+ * Lmiit for btree_trans_too_many_iters(); this is enough that almost all code
+ * paths should run inside this limit, and if they don't it usually indicates a
+ * bug (leaking/duplicated btree paths).
+ *
+ * exception: some fsck paths
+ *
+ * bugs with excessive path usage seem to have possibly been eliminated now, so
+ * we might consider eliminating this (and btree_trans_too_many_iter()) at some
+ * point.
+ */
+#define BTREE_ITER_NORMAL_LIMIT		256
+/* never exceed limit */
 #define BTREE_ITER_MAX			(1U << 10)
 
 struct btree_trans_commit_hook;
diff --git a/fs/bcachefs/btree_update_interior.c b/fs/bcachefs/btree_update_interior.c
index 32397b99752f..b4efd8cc4d1a 100644
--- a/fs/bcachefs/btree_update_interior.c
+++ b/fs/bcachefs/btree_update_interior.c
@@ -21,14 +21,15 @@
 #include "keylist.h"
 #include "recovery_passes.h"
 #include "replicas.h"
+#include "sb-members.h"
 #include "super-io.h"
 #include "trace.h"
 
 #include <linux/random.h>
 
-const char * const bch2_btree_update_modes[] = {
+static const char * const bch2_btree_update_modes[] = {
 #define x(t) #t,
-	BCH_WATERMARKS()
+	BTREE_UPDATE_MODES()
 #undef x
 	NULL
 };
@@ -605,6 +606,26 @@ static void btree_update_add_key(struct btree_update *as,
 	bch2_keylist_push(keys);
 }
 
+static bool btree_update_new_nodes_marked_sb(struct btree_update *as)
+{
+	for_each_keylist_key(&as->new_keys, k)
+		if (!bch2_dev_btree_bitmap_marked(as->c, bkey_i_to_s_c(k)))
+			return false;
+	return true;
+}
+
+static void btree_update_new_nodes_mark_sb(struct btree_update *as)
+{
+	struct bch_fs *c = as->c;
+
+	mutex_lock(&c->sb_lock);
+	for_each_keylist_key(&as->new_keys, k)
+		bch2_dev_btree_bitmap_mark(c, bkey_i_to_s_c(k));
+
+	bch2_write_super(c);
+	mutex_unlock(&c->sb_lock);
+}
+
 /*
  * The transactional part of an interior btree node update, where we journal the
  * update we did to the interior node and update alloc info:
@@ -662,6 +683,9 @@ static void btree_update_nodes_written(struct btree_update *as)
 	if (ret)
 		goto err;
 
+	if (!btree_update_new_nodes_marked_sb(as))
+		btree_update_new_nodes_mark_sb(as);
+
 	/*
 	 * Wait for any in flight writes to finish before we free the old nodes
 	 * on disk:
@@ -704,9 +728,13 @@ static void btree_update_nodes_written(struct btree_update *as)
 	bch2_fs_fatal_err_on(ret && !bch2_journal_error(&c->journal), c,
 			     "%s", bch2_err_str(ret));
 err:
-	if (as->b) {
-
-		b = as->b;
+	/*
+	 * We have to be careful because another thread might be getting ready
+	 * to free as->b and calling btree_update_reparent() on us - we'll
+	 * recheck under btree_update_lock below:
+	 */
+	b = READ_ONCE(as->b);
+	if (b) {
 		btree_path_idx_t path_idx = get_unlocked_mut_path(trans,
 						as->btree_id, b->c.level, b->key.k.p);
 		struct btree_path *path = trans->paths + path_idx;
@@ -850,15 +878,17 @@ static void btree_update_updated_node(struct btree_update *as, struct btree *b)
 {
 	struct bch_fs *c = as->c;
 
-	mutex_lock(&c->btree_interior_update_lock);
-	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
-
 	BUG_ON(as->mode != BTREE_UPDATE_none);
+	BUG_ON(as->update_level_end < b->c.level);
 	BUG_ON(!btree_node_dirty(b));
 	BUG_ON(!b->c.level);
 
+	mutex_lock(&c->btree_interior_update_lock);
+	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
+
 	as->mode	= BTREE_UPDATE_node;
 	as->b		= b;
+	as->update_level_end = b->c.level;
 
 	set_btree_node_write_blocked(b);
 	list_add(&as->write_blocked_list, &b->write_blocked);
@@ -1100,7 +1130,7 @@ static void bch2_btree_update_done(struct btree_update *as, struct btree_trans *
 
 static struct btree_update *
 bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
-			unsigned level, bool split, unsigned flags)
+			unsigned level_start, bool split, unsigned flags)
 {
 	struct bch_fs *c = trans->c;
 	struct btree_update *as;
@@ -1108,7 +1138,7 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	int disk_res_flags = (flags & BCH_TRANS_COMMIT_no_enospc)
 		? BCH_DISK_RESERVATION_NOFAIL : 0;
 	unsigned nr_nodes[2] = { 0, 0 };
-	unsigned update_level = level;
+	unsigned level_end = level_start;
 	enum bch_watermark watermark = flags & BCH_WATERMARK_MASK;
 	int ret = 0;
 	u32 restart_count = trans->restart_count;
@@ -1123,34 +1153,30 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	flags &= ~BCH_WATERMARK_MASK;
 	flags |= watermark;
 
-	if (watermark < c->journal.watermark) {
-		struct journal_res res = { 0 };
-		unsigned journal_flags = watermark|JOURNAL_RES_GET_CHECK;
+	if (watermark < BCH_WATERMARK_reclaim &&
+	    test_bit(JOURNAL_SPACE_LOW, &c->journal.flags)) {
+		if (flags & BCH_TRANS_COMMIT_journal_reclaim)
+			return ERR_PTR(-BCH_ERR_journal_reclaim_would_deadlock);
 
-		if ((flags & BCH_TRANS_COMMIT_journal_reclaim) &&
-		    watermark < BCH_WATERMARK_reclaim)
-			journal_flags |= JOURNAL_RES_GET_NONBLOCK;
-
-		ret = drop_locks_do(trans,
-			bch2_journal_res_get(&c->journal, &res, 1, journal_flags));
-		if (bch2_err_matches(ret, BCH_ERR_operation_blocked))
-			ret = -BCH_ERR_journal_reclaim_would_deadlock;
+		bch2_trans_unlock(trans);
+		wait_event(c->journal.wait, !test_bit(JOURNAL_SPACE_LOW, &c->journal.flags));
+		ret = bch2_trans_relock(trans);
 		if (ret)
 			return ERR_PTR(ret);
 	}
 
 	while (1) {
-		nr_nodes[!!update_level] += 1 + split;
-		update_level++;
+		nr_nodes[!!level_end] += 1 + split;
+		level_end++;
 
-		ret = bch2_btree_path_upgrade(trans, path, update_level + 1);
+		ret = bch2_btree_path_upgrade(trans, path, level_end + 1);
 		if (ret)
 			return ERR_PTR(ret);
 
-		if (!btree_path_node(path, update_level)) {
+		if (!btree_path_node(path, level_end)) {
 			/* Allocating new root? */
 			nr_nodes[1] += split;
-			update_level = BTREE_MAX_DEPTH;
+			level_end = BTREE_MAX_DEPTH;
 			break;
 		}
 
@@ -1158,11 +1184,11 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 		 * Always check for space for two keys, even if we won't have to
 		 * split at prior level - it might have been a merge instead:
 		 */
-		if (bch2_btree_node_insert_fits(path->l[update_level].b,
+		if (bch2_btree_node_insert_fits(path->l[level_end].b,
 						BKEY_BTREE_PTR_U64s_MAX * 2))
 			break;
 
-		split = path->l[update_level].b->nr.live_u64s > BTREE_SPLIT_THRESHOLD(c);
+		split = path->l[level_end].b->nr.live_u64s > BTREE_SPLIT_THRESHOLD(c);
 	}
 
 	if (!down_read_trylock(&c->gc_lock)) {
@@ -1176,14 +1202,15 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path,
 	as = mempool_alloc(&c->btree_interior_update_pool, GFP_NOFS);
 	memset(as, 0, sizeof(*as));
 	closure_init(&as->cl, NULL);
-	as->c		= c;
-	as->start_time	= start_time;
-	as->ip_started	= _RET_IP_;
-	as->mode	= BTREE_UPDATE_none;
-	as->watermark	= watermark;
-	as->took_gc_lock = true;
-	as->btree_id	= path->btree_id;
-	as->update_level = update_level;
+	as->c			= c;
+	as->start_time		= start_time;
+	as->ip_started		= _RET_IP_;
+	as->mode		= BTREE_UPDATE_none;
+	as->watermark		= watermark;
+	as->took_gc_lock	= true;
+	as->btree_id		= path->btree_id;
+	as->update_level_start	= level_start;
+	as->update_level_end	= level_end;
 	INIT_LIST_HEAD(&as->list);
 	INIT_LIST_HEAD(&as->unwritten_list);
 	INIT_LIST_HEAD(&as->write_blocked_list);
@@ -1277,23 +1304,29 @@ static void bch2_btree_set_root_inmem(struct bch_fs *c, struct btree *b)
 	bch2_recalc_btree_reserve(c);
 }
 
-static void bch2_btree_set_root(struct btree_update *as,
-				struct btree_trans *trans,
-				struct btree_path *path,
-				struct btree *b)
+static int bch2_btree_set_root(struct btree_update *as,
+			       struct btree_trans *trans,
+			       struct btree_path *path,
+			       struct btree *b,
+			       bool nofail)
 {
 	struct bch_fs *c = as->c;
-	struct btree *old;
 
 	trace_and_count(c, btree_node_set_root, trans, b);
 
-	old = btree_node_root(c, b);
+	struct btree *old = btree_node_root(c, b);
 
 	/*
 	 * Ensure no one is using the old root while we switch to the
 	 * new root:
 	 */
-	bch2_btree_node_lock_write_nofail(trans, path, &old->c);
+	if (nofail) {
+		bch2_btree_node_lock_write_nofail(trans, path, &old->c);
+	} else {
+		int ret = bch2_btree_node_lock_write(trans, path, &old->c);
+		if (ret)
+			return ret;
+	}
 
 	bch2_btree_set_root_inmem(c, b);
 
@@ -1307,6 +1340,7 @@ static void bch2_btree_set_root(struct btree_update *as,
 	 * depend on the new root would have to update the new root.
 	 */
 	bch2_btree_node_unlock_write(trans, path, old);
+	return 0;
 }
 
 /* Interior node updates: */
@@ -1373,12 +1407,12 @@ static void bch2_insert_fixup_btree_ptr(struct btree_update *as,
 }
 
 static void
-__bch2_btree_insert_keys_interior(struct btree_update *as,
-				  struct btree_trans *trans,
-				  struct btree_path *path,
-				  struct btree *b,
-				  struct btree_node_iter node_iter,
-				  struct keylist *keys)
+bch2_btree_insert_keys_interior(struct btree_update *as,
+				struct btree_trans *trans,
+				struct btree_path *path,
+				struct btree *b,
+				struct btree_node_iter node_iter,
+				struct keylist *keys)
 {
 	struct bkey_i *insert = bch2_keylist_front(keys);
 	struct bkey_packed *k;
@@ -1534,7 +1568,7 @@ static void btree_split_insert_keys(struct btree_update *as,
 
 		bch2_btree_node_iter_init(&node_iter, b, &bch2_keylist_front(keys)->k.p);
 
-		__bch2_btree_insert_keys_interior(as, trans, path, b, node_iter, keys);
+		bch2_btree_insert_keys_interior(as, trans, path, b, node_iter, keys);
 
 		BUG_ON(bch2_btree_node_check_topology(trans, b));
 	}
@@ -1649,15 +1683,16 @@ static int btree_split(struct btree_update *as, struct btree_trans *trans,
 	if (parent) {
 		/* Split a non root node */
 		ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys);
-		if (ret)
-			goto err;
 	} else if (n3) {
-		bch2_btree_set_root(as, trans, trans->paths + path, n3);
+		ret = bch2_btree_set_root(as, trans, trans->paths + path, n3, false);
 	} else {
 		/* Root filled up but didn't need to be split */
-		bch2_btree_set_root(as, trans, trans->paths + path, n1);
+		ret = bch2_btree_set_root(as, trans, trans->paths + path, n1, false);
 	}
 
+	if (ret)
+		goto err;
+
 	if (n3) {
 		bch2_btree_update_get_open_buckets(as, n3);
 		bch2_btree_node_write(c, n3, SIX_LOCK_intent, 0);
@@ -1714,27 +1749,6 @@ err:
 	goto out;
 }
 
-static void
-bch2_btree_insert_keys_interior(struct btree_update *as,
-				struct btree_trans *trans,
-				struct btree_path *path,
-				struct btree *b,
-				struct keylist *keys)
-{
-	struct btree_path *linked;
-	unsigned i;
-
-	__bch2_btree_insert_keys_interior(as, trans, path, b,
-					  path->l[b->c.level].iter, keys);
-
-	btree_update_updated_node(as, b);
-
-	trans_for_each_path_with_node(trans, b, linked, i)
-		bch2_btree_node_iter_peek(&linked->l[b->c.level].iter, b);
-
-	bch2_trans_verify_paths(trans);
-}
-
 /**
  * bch2_btree_insert_node - insert bkeys into a given btree node
  *
@@ -1755,7 +1769,8 @@ static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *t
 				  struct keylist *keys)
 {
 	struct bch_fs *c = as->c;
-	struct btree_path *path = trans->paths + path_idx;
+	struct btree_path *path = trans->paths + path_idx, *linked;
+	unsigned i;
 	int old_u64s = le16_to_cpu(btree_bset_last(b)->u64s);
 	int old_live_u64s = b->nr.live_u64s;
 	int live_u64s_added, u64s_added;
@@ -1784,7 +1799,13 @@ static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *t
 		return ret;
 	}
 
-	bch2_btree_insert_keys_interior(as, trans, path, b, keys);
+	bch2_btree_insert_keys_interior(as, trans, path, b,
+					path->l[b->c.level].iter, keys);
+
+	trans_for_each_path_with_node(trans, b, linked, i)
+		bch2_btree_node_iter_peek(&linked->l[b->c.level].iter, b);
+
+	bch2_trans_verify_paths(trans);
 
 	live_u64s_added = (int) b->nr.live_u64s - old_live_u64s;
 	u64s_added = (int) le16_to_cpu(btree_bset_last(b)->u64s) - old_u64s;
@@ -1798,6 +1819,7 @@ static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *t
 	    bch2_maybe_compact_whiteouts(c, b))
 		bch2_trans_node_reinit_iter(trans, b);
 
+	btree_update_updated_node(as, b);
 	bch2_btree_node_unlock_write(trans, path, b);
 
 	BUG_ON(bch2_btree_node_check_topology(trans, b));
@@ -1807,7 +1829,7 @@ split:
 	 * We could attempt to avoid the transaction restart, by calling
 	 * bch2_btree_path_upgrade() and allocating more nodes:
 	 */
-	if (b->c.level >= as->update_level) {
+	if (b->c.level >= as->update_level_end) {
 		trace_and_count(c, trans_restart_split_race, trans, _THIS_IP_, b);
 		return btree_trans_restart(trans, BCH_ERR_transaction_restart_split_race);
 	}
@@ -1873,7 +1895,9 @@ static void __btree_increase_depth(struct btree_update *as, struct btree_trans *
 	bch2_keylist_add(&as->parent_keys, &b->key);
 	btree_split_insert_keys(as, trans, path_idx, n, &as->parent_keys);
 
-	bch2_btree_set_root(as, trans, path, n);
+	int ret = bch2_btree_set_root(as, trans, path, n, true);
+	BUG_ON(ret);
+
 	bch2_btree_update_get_open_buckets(as, n);
 	bch2_btree_node_write(c, n, SIX_LOCK_intent, 0);
 	bch2_trans_node_add(trans, path, n);
@@ -1926,6 +1950,22 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans,
 	BUG_ON(!trans->paths[path].should_be_locked);
 	BUG_ON(!btree_node_locked(&trans->paths[path], level));
 
+	/*
+	 * Work around a deadlock caused by the btree write buffer not doing
+	 * merges and leaving tons of merges for us to do - we really don't need
+	 * to be doing merges at all from the interior update path, and if the
+	 * interior update path is generating too many new interior updates we
+	 * deadlock:
+	 */
+	if ((flags & BCH_WATERMARK_MASK) == BCH_WATERMARK_interior_updates)
+		return 0;
+
+	if ((flags & BCH_WATERMARK_MASK) <= BCH_WATERMARK_reclaim) {
+		flags &= ~BCH_WATERMARK_MASK;
+		flags |= BCH_WATERMARK_btree;
+		flags |= BCH_TRANS_COMMIT_journal_reclaim;
+	}
+
 	b = trans->paths[path].l[level].b;
 
 	if ((sib == btree_prev_sib && bpos_eq(b->data->min_key, POS_MIN)) ||
@@ -2071,6 +2111,10 @@ err:
 		bch2_path_put(trans, new_path, true);
 	bch2_path_put(trans, sib_path, true);
 	bch2_trans_verify_locks(trans);
+	if (ret == -BCH_ERR_journal_reclaim_would_deadlock)
+		ret = 0;
+	if (!ret)
+		ret = bch2_trans_relock(trans);
 	return ret;
 err_free_update:
 	bch2_btree_node_free_never_used(as, trans, n);
@@ -2116,12 +2160,13 @@ int bch2_btree_node_rewrite(struct btree_trans *trans,
 	if (parent) {
 		bch2_keylist_add(&as->parent_keys, &n->key);
 		ret = bch2_btree_insert_node(as, trans, iter->path, parent, &as->parent_keys);
-		if (ret)
-			goto err;
 	} else {
-		bch2_btree_set_root(as, trans, btree_iter_path(trans, iter), n);
+		ret = bch2_btree_set_root(as, trans, btree_iter_path(trans, iter), n, false);
 	}
 
+	if (ret)
+		goto err;
+
 	bch2_btree_update_get_open_buckets(as, n);
 	bch2_btree_node_write(c, n, SIX_LOCK_intent, 0);
 
@@ -2519,9 +2564,11 @@ void bch2_btree_root_alloc_fake(struct bch_fs *c, enum btree_id id, unsigned lev
 
 static void bch2_btree_update_to_text(struct printbuf *out, struct btree_update *as)
 {
-	prt_printf(out, "%ps: btree=%s watermark=%s mode=%s nodes_written=%u cl.remaining=%u journal_seq=%llu\n",
+	prt_printf(out, "%ps: btree=%s l=%u-%u watermark=%s mode=%s nodes_written=%u cl.remaining=%u journal_seq=%llu\n",
 		   (void *) as->ip_started,
 		   bch2_btree_id_str(as->btree_id),
+		   as->update_level_start,
+		   as->update_level_end,
 		   bch2_watermarks[as->watermark],
 		   bch2_btree_update_modes[as->mode],
 		   as->nodes_written,
diff --git a/fs/bcachefs/btree_update_interior.h b/fs/bcachefs/btree_update_interior.h
index 88dcf5a22a3b..c1a479ebaad1 100644
--- a/fs/bcachefs/btree_update_interior.h
+++ b/fs/bcachefs/btree_update_interior.h
@@ -57,7 +57,8 @@ struct btree_update {
 	unsigned			took_gc_lock:1;
 
 	enum btree_id			btree_id;
-	unsigned			update_level;
+	unsigned			update_level_start;
+	unsigned			update_level_end;
 
 	struct disk_reservation		disk_res;
 
diff --git a/fs/bcachefs/btree_write_buffer.c b/fs/bcachefs/btree_write_buffer.c
index baf63e2fddb6..36a6f42aba5e 100644
--- a/fs/bcachefs/btree_write_buffer.c
+++ b/fs/bcachefs/btree_write_buffer.c
@@ -316,6 +316,16 @@ static int bch2_btree_write_buffer_flush_locked(struct btree_trans *trans)
 			    bpos_gt(k->k.k.p, path->l[0].b->key.k.p)) {
 				bch2_btree_node_unlock_write(trans, path, path->l[0].b);
 				write_locked = false;
+
+				ret = lockrestart_do(trans,
+					bch2_btree_iter_traverse(&iter) ?:
+					bch2_foreground_maybe_merge(trans, iter.path, 0,
+							BCH_WATERMARK_reclaim|
+							BCH_TRANS_COMMIT_journal_reclaim|
+							BCH_TRANS_COMMIT_no_check_rw|
+							BCH_TRANS_COMMIT_no_enospc));
+				if (ret)
+					goto err;
 			}
 		}
 
@@ -382,10 +392,10 @@ static int bch2_btree_write_buffer_flush_locked(struct btree_trans *trans)
 
 			ret = commit_do(trans, NULL, NULL,
 					BCH_WATERMARK_reclaim|
+					BCH_TRANS_COMMIT_journal_reclaim|
 					BCH_TRANS_COMMIT_no_check_rw|
 					BCH_TRANS_COMMIT_no_enospc|
-					BCH_TRANS_COMMIT_no_journal_res|
-					BCH_TRANS_COMMIT_journal_reclaim,
+					BCH_TRANS_COMMIT_no_journal_res ,
 					btree_write_buffered_insert(trans, i));
 			if (ret)
 				goto err;
diff --git a/fs/bcachefs/buckets.c b/fs/bcachefs/buckets.c
index 941401a210f5..82f179258867 100644
--- a/fs/bcachefs/buckets.c
+++ b/fs/bcachefs/buckets.c
@@ -525,7 +525,6 @@ int bch2_mark_metadata_bucket(struct bch_fs *c, struct bch_dev *ca,
 			"different types of data in same bucket: %s, %s",
 			bch2_data_type_str(g->data_type),
 			bch2_data_type_str(data_type))) {
-		BUG();
 		ret = -EIO;
 		goto err;
 	}
@@ -629,7 +628,6 @@ int bch2_check_bucket_ref(struct btree_trans *trans,
 			bch2_data_type_str(ptr_data_type),
 			(printbuf_reset(&buf),
 			 bch2_bkey_val_to_text(&buf, c, k), buf.buf));
-		BUG();
 		ret = -EIO;
 		goto err;
 	}
diff --git a/fs/bcachefs/buckets.h b/fs/bcachefs/buckets.h
index 00aaf4bb5139..f9af5adabe83 100644
--- a/fs/bcachefs/buckets.h
+++ b/fs/bcachefs/buckets.h
@@ -395,14 +395,6 @@ static inline const char *bch2_data_type_str(enum bch_data_type type)
 		: "(invalid data type)";
 }
 
-static inline void bch2_prt_data_type(struct printbuf *out, enum bch_data_type type)
-{
-	if (type < BCH_DATA_NR)
-		prt_str(out, __bch2_data_types[type]);
-	else
-		prt_printf(out, "(invalid data type %u)", type);
-}
-
 /* disk reservations: */
 
 static inline void bch2_disk_reservation_put(struct bch_fs *c,
diff --git a/fs/bcachefs/chardev.c b/fs/bcachefs/chardev.c
index cbfa6459bdbc..4d14f19f5185 100644
--- a/fs/bcachefs/chardev.c
+++ b/fs/bcachefs/chardev.c
@@ -134,42 +134,38 @@ static long bch2_ioctl_incremental(struct bch_ioctl_incremental __user *user_arg
 struct fsck_thread {
 	struct thread_with_stdio thr;
 	struct bch_fs		*c;
-	char			**devs;
-	size_t			nr_devs;
 	struct bch_opts		opts;
 };
 
 static void bch2_fsck_thread_exit(struct thread_with_stdio *_thr)
 {
 	struct fsck_thread *thr = container_of(_thr, struct fsck_thread, thr);
-	if (thr->devs)
-		for (size_t i = 0; i < thr->nr_devs; i++)
-			kfree(thr->devs[i]);
-	kfree(thr->devs);
 	kfree(thr);
 }
 
 static int bch2_fsck_offline_thread_fn(struct thread_with_stdio *stdio)
 {
 	struct fsck_thread *thr = container_of(stdio, struct fsck_thread, thr);
-	struct bch_fs *c = bch2_fs_open(thr->devs, thr->nr_devs, thr->opts);
-
-	if (IS_ERR(c))
-		return PTR_ERR(c);
+	struct bch_fs *c = thr->c;
 
-	int ret = 0;
-	if (test_bit(BCH_FS_errors_fixed, &c->flags))
-		ret |= 1;
-	if (test_bit(BCH_FS_error, &c->flags))
-		ret |= 4;
+	int ret = PTR_ERR_OR_ZERO(c);
+	if (ret)
+		return ret;
 
-	bch2_fs_stop(c);
+	ret = bch2_fs_start(thr->c);
+	if (ret)
+		goto err;
 
-	if (ret & 1)
+	if (test_bit(BCH_FS_errors_fixed, &c->flags)) {
 		bch2_stdio_redirect_printf(&stdio->stdio, false, "%s: errors fixed\n", c->name);
-	if (ret & 4)
+		ret |= 1;
+	}
+	if (test_bit(BCH_FS_error, &c->flags)) {
 		bch2_stdio_redirect_printf(&stdio->stdio, false, "%s: still has errors\n", c->name);
-
+		ret |= 4;
+	}
+err:
+	bch2_fs_stop(c);
 	return ret;
 }
 
@@ -182,7 +178,7 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a
 {
 	struct bch_ioctl_fsck_offline arg;
 	struct fsck_thread *thr = NULL;
-	u64 *devs = NULL;
+	darray_str(devs) = {};
 	long ret = 0;
 
 	if (copy_from_user(&arg, user_arg, sizeof(arg)))
@@ -194,29 +190,32 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (!(devs = kcalloc(arg.nr_devs, sizeof(*devs), GFP_KERNEL)) ||
-	    !(thr = kzalloc(sizeof(*thr), GFP_KERNEL)) ||
-	    !(thr->devs = kcalloc(arg.nr_devs, sizeof(*thr->devs), GFP_KERNEL))) {
-		ret = -ENOMEM;
-		goto err;
-	}
+	for (size_t i = 0; i < arg.nr_devs; i++) {
+		u64 dev_u64;
+		ret = copy_from_user_errcode(&dev_u64, &user_arg->devs[i], sizeof(u64));
+		if (ret)
+			goto err;
 
-	thr->opts = bch2_opts_empty();
-	thr->nr_devs = arg.nr_devs;
+		char *dev_str = strndup_user((char __user *)(unsigned long) dev_u64, PATH_MAX);
+		ret = PTR_ERR_OR_ZERO(dev_str);
+		if (ret)
+			goto err;
 
-	if (copy_from_user(devs, &user_arg->devs[0],
-			   array_size(sizeof(user_arg->devs[0]), arg.nr_devs))) {
-		ret = -EINVAL;
-		goto err;
+		ret = darray_push(&devs, dev_str);
+		if (ret) {
+			kfree(dev_str);
+			goto err;
+		}
 	}
 
-	for (size_t i = 0; i < arg.nr_devs; i++) {
-		thr->devs[i] = strndup_user((char __user *)(unsigned long) devs[i], PATH_MAX);
-		ret = PTR_ERR_OR_ZERO(thr->devs[i]);
-		if (ret)
-			goto err;
+	thr = kzalloc(sizeof(*thr), GFP_KERNEL);
+	if (!thr) {
+		ret = -ENOMEM;
+		goto err;
 	}
 
+	thr->opts = bch2_opts_empty();
+
 	if (arg.opts) {
 		char *optstr = strndup_user((char __user *)(unsigned long) arg.opts, 1 << 16);
 
@@ -230,15 +229,28 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a
 
 	opt_set(thr->opts, stdio, (u64)(unsigned long)&thr->thr.stdio);
 
-	ret = bch2_run_thread_with_stdio(&thr->thr, &bch2_offline_fsck_ops);
-err:
-	if (ret < 0) {
-		if (thr)
-			bch2_fsck_thread_exit(&thr->thr);
-		pr_err("ret %s", bch2_err_str(ret));
-	}
-	kfree(devs);
+	/* We need request_key() to be called before we punt to kthread: */
+	opt_set(thr->opts, nostart, true);
+
+	bch2_thread_with_stdio_init(&thr->thr, &bch2_offline_fsck_ops);
+
+	thr->c = bch2_fs_open(devs.data, arg.nr_devs, thr->opts);
+
+	if (!IS_ERR(thr->c) &&
+	    thr->c->opts.errors == BCH_ON_ERROR_panic)
+		thr->c->opts.errors = BCH_ON_ERROR_ro;
+
+	ret = __bch2_run_thread_with_stdio(&thr->thr);
+out:
+	darray_for_each(devs, i)
+		kfree(*i);
+	darray_exit(&devs);
 	return ret;
+err:
+	if (thr)
+		bch2_fsck_thread_exit(&thr->thr);
+	pr_err("ret %s", bch2_err_str(ret));
+	goto out;
 }
 
 static long bch2_global_ioctl(unsigned cmd, void __user *arg)
diff --git a/fs/bcachefs/checksum.c b/fs/bcachefs/checksum.c
index 4701457f6381..7ed779b411f6 100644
--- a/fs/bcachefs/checksum.c
+++ b/fs/bcachefs/checksum.c
@@ -429,15 +429,20 @@ int bch2_rechecksum_bio(struct bch_fs *c, struct bio *bio,
 				extent_nonce(version, crc_old), bio);
 
 	if (bch2_crc_cmp(merged, crc_old.csum) && !c->opts.no_data_io) {
-		bch_err(c, "checksum error in %s() (memory corruption or bug?)\n"
-			"expected %0llx:%0llx got %0llx:%0llx (old type %s new type %s)",
-			__func__,
-			crc_old.csum.hi,
-			crc_old.csum.lo,
-			merged.hi,
-			merged.lo,
-			bch2_csum_types[crc_old.csum_type],
-			bch2_csum_types[new_csum_type]);
+		struct printbuf buf = PRINTBUF;
+		prt_printf(&buf, "checksum error in %s() (memory corruption or bug?)\n"
+			   "expected %0llx:%0llx got %0llx:%0llx (old type ",
+			   __func__,
+			   crc_old.csum.hi,
+			   crc_old.csum.lo,
+			   merged.hi,
+			   merged.lo);
+		bch2_prt_csum_type(&buf, crc_old.csum_type);
+		prt_str(&buf, " new type ");
+		bch2_prt_csum_type(&buf, new_csum_type);
+		prt_str(&buf, ")");
+		bch_err(c, "%s", buf.buf);
+		printbuf_exit(&buf);
 		return -EIO;
 	}
 
diff --git a/fs/bcachefs/checksum.h b/fs/bcachefs/checksum.h
index 1b8c2c1016dc..e40499fde9a4 100644
--- a/fs/bcachefs/checksum.h
+++ b/fs/bcachefs/checksum.h
@@ -61,11 +61,12 @@ static inline void bch2_csum_err_msg(struct printbuf *out,
 				     struct bch_csum expected,
 				     struct bch_csum got)
 {
-	prt_printf(out, "checksum error: got ");
+	prt_str(out, "checksum error, type ");
+	bch2_prt_csum_type(out, type);
+	prt_str(out, ": got ");
 	bch2_csum_to_text(out, type, got);
 	prt_str(out, " should be ");
 	bch2_csum_to_text(out, type, expected);
-	prt_printf(out, " type %s", bch2_csum_types[type]);
 }
 
 int bch2_chacha_encrypt_key(struct bch_key *, struct nonce, void *, size_t);
diff --git a/fs/bcachefs/compress.h b/fs/bcachefs/compress.h
index 58c2eb45570f..607fd5e232c9 100644
--- a/fs/bcachefs/compress.h
+++ b/fs/bcachefs/compress.h
@@ -47,14 +47,6 @@ static inline enum bch_compression_type bch2_compression_opt_to_type(unsigned v)
 	return __bch2_compression_opt_to_type[bch2_compression_decode(v).type];
 }
 
-static inline void bch2_prt_compression_type(struct printbuf *out, enum bch_compression_type type)
-{
-	if (type < BCH_COMPRESSION_TYPE_NR)
-		prt_str(out, __bch2_compression_types[type]);
-	else
-		prt_printf(out, "(invalid compression type %u)", type);
-}
-
 int bch2_bio_uncompress_inplace(struct bch_fs *, struct bio *,
 				struct bch_extent_crc_unpacked *);
 int bch2_bio_uncompress(struct bch_fs *, struct bio *, struct bio *,
diff --git a/fs/bcachefs/data_update.c b/fs/bcachefs/data_update.c
index 34731ee0217f..0022b51ce3c0 100644
--- a/fs/bcachefs/data_update.c
+++ b/fs/bcachefs/data_update.c
@@ -598,6 +598,8 @@ int bch2_data_update_init(struct btree_trans *trans,
 		i++;
 	}
 
+	unsigned durability_required = max(0, (int) (io_opts.data_replicas - durability_have));
+
 	/*
 	 * If current extent durability is less than io_opts.data_replicas,
 	 * we're not trying to rereplicate the extent up to data_replicas here -
@@ -607,7 +609,7 @@ int bch2_data_update_init(struct btree_trans *trans,
 	 * rereplicate, currently, so that users don't get an unexpected -ENOSPC
 	 */
 	if (!(m->data_opts.write_flags & BCH_WRITE_CACHED) &&
-	    durability_have >= io_opts.data_replicas) {
+	    !durability_required) {
 		m->data_opts.kill_ptrs |= m->data_opts.rewrite_ptrs;
 		m->data_opts.rewrite_ptrs = 0;
 		/* if iter == NULL, it's just a promote */
@@ -616,11 +618,18 @@ int bch2_data_update_init(struct btree_trans *trans,
 		goto done;
 	}
 
-	m->op.nr_replicas = min(durability_removing, io_opts.data_replicas - durability_have) +
+	m->op.nr_replicas = min(durability_removing, durability_required) +
 		m->data_opts.extra_replicas;
-	m->op.nr_replicas_required = m->op.nr_replicas;
 
-	BUG_ON(!m->op.nr_replicas);
+	/*
+	 * If device(s) were set to durability=0 after data was written to them
+	 * we can end up with a duribilty=0 extent, and the normal algorithm
+	 * that tries not to increase durability doesn't work:
+	 */
+	if (!(durability_have + durability_removing))
+		m->op.nr_replicas = max((unsigned) m->op.nr_replicas, 1);
+
+	m->op.nr_replicas_required = m->op.nr_replicas;
 
 	if (reserve_sectors) {
 		ret = bch2_disk_reservation_add(c, &m->op.res, reserve_sectors,
diff --git a/fs/bcachefs/debug.c b/fs/bcachefs/debug.c
index 208ce6f0fc43..cd99b7399414 100644
--- a/fs/bcachefs/debug.c
+++ b/fs/bcachefs/debug.c
@@ -13,6 +13,7 @@
 #include "btree_iter.h"
 #include "btree_locking.h"
 #include "btree_update.h"
+#include "btree_update_interior.h"
 #include "buckets.h"
 #include "debug.h"
 #include "error.h"
@@ -668,7 +669,7 @@ static ssize_t bch2_journal_pins_read(struct file *file, char __user *buf,
 	i->size	= size;
 	i->ret	= 0;
 
-	do {
+	while (1) {
 		err = flush_buf(i);
 		if (err)
 			return err;
@@ -676,9 +677,12 @@ static ssize_t bch2_journal_pins_read(struct file *file, char __user *buf,
 		if (!i->size)
 			break;
 
+		if (done)
+			break;
+
 		done = bch2_journal_seq_pins_to_text(&i->buf, &c->journal, &i->iter);
 		i->iter++;
-	} while (!done);
+	}
 
 	if (i->buf.allocation_failure)
 		return -ENOMEM;
@@ -693,13 +697,45 @@ static const struct file_operations journal_pins_ops = {
 	.read		= bch2_journal_pins_read,
 };
 
+static ssize_t bch2_btree_updates_read(struct file *file, char __user *buf,
+				       size_t size, loff_t *ppos)
+{
+	struct dump_iter *i = file->private_data;
+	struct bch_fs *c = i->c;
+	int err;
+
+	i->ubuf = buf;
+	i->size	= size;
+	i->ret	= 0;
+
+	if (!i->iter) {
+		bch2_btree_updates_to_text(&i->buf, c);
+		i->iter++;
+	}
+
+	err = flush_buf(i);
+	if (err)
+		return err;
+
+	if (i->buf.allocation_failure)
+		return -ENOMEM;
+
+	return i->ret;
+}
+
+static const struct file_operations btree_updates_ops = {
+	.owner		= THIS_MODULE,
+	.open		= bch2_dump_open,
+	.release	= bch2_dump_release,
+	.read		= bch2_btree_updates_read,
+};
+
 static int btree_transaction_stats_open(struct inode *inode, struct file *file)
 {
 	struct bch_fs *c = inode->i_private;
 	struct dump_iter *i;
 
 	i = kzalloc(sizeof(struct dump_iter), GFP_KERNEL);
-
 	if (!i)
 		return -ENOMEM;
 
@@ -866,6 +902,20 @@ void bch2_fs_debug_exit(struct bch_fs *c)
 		debugfs_remove_recursive(c->fs_debug_dir);
 }
 
+static void bch2_fs_debug_btree_init(struct bch_fs *c, struct btree_debug *bd)
+{
+	struct dentry *d;
+
+	d = debugfs_create_dir(bch2_btree_id_str(bd->id), c->btree_debug_dir);
+
+	debugfs_create_file("keys", 0400, d, bd, &btree_debug_ops);
+
+	debugfs_create_file("formats", 0400, d, bd, &btree_format_debug_ops);
+
+	debugfs_create_file("bfloat-failed", 0400, d, bd,
+			    &bfloat_failed_debug_ops);
+}
+
 void bch2_fs_debug_init(struct bch_fs *c)
 {
 	struct btree_debug *bd;
@@ -888,6 +938,9 @@ void bch2_fs_debug_init(struct bch_fs *c)
 	debugfs_create_file("journal_pins", 0400, c->fs_debug_dir,
 			    c->btree_debug, &journal_pins_ops);
 
+	debugfs_create_file("btree_updates", 0400, c->fs_debug_dir,
+			    c->btree_debug, &btree_updates_ops);
+
 	debugfs_create_file("btree_transaction_stats", 0400, c->fs_debug_dir,
 			    c, &btree_transaction_stats_op);
 
@@ -902,21 +955,7 @@ void bch2_fs_debug_init(struct bch_fs *c)
 	     bd < c->btree_debug + ARRAY_SIZE(c->btree_debug);
 	     bd++) {
 		bd->id = bd - c->btree_debug;
-		debugfs_create_file(bch2_btree_id_str(bd->id),
-				    0400, c->btree_debug_dir, bd,
-				    &btree_debug_ops);
-
-		snprintf(name, sizeof(name), "%s-formats",
-			 bch2_btree_id_str(bd->id));
-
-		debugfs_create_file(name, 0400, c->btree_debug_dir, bd,
-				    &btree_format_debug_ops);
-
-		snprintf(name, sizeof(name), "%s-bfloat-failed",
-			 bch2_btree_id_str(bd->id));
-
-		debugfs_create_file(name, 0400, c->btree_debug_dir, bd,
-				    &bfloat_failed_debug_ops);
+		bch2_fs_debug_btree_init(c, bd);
 	}
 }
 
diff --git a/fs/bcachefs/ec.c b/fs/bcachefs/ec.c
index 082075244e16..556a217108d3 100644
--- a/fs/bcachefs/ec.c
+++ b/fs/bcachefs/ec.c
@@ -131,29 +131,33 @@ fsck_err:
 void bch2_stripe_to_text(struct printbuf *out, struct bch_fs *c,
 			 struct bkey_s_c k)
 {
-	const struct bch_stripe *s = bkey_s_c_to_stripe(k).v;
-	unsigned i, nr_data = s->nr_blocks - s->nr_redundant;
+	const struct bch_stripe *sp = bkey_s_c_to_stripe(k).v;
+	struct bch_stripe s = {};
+
+	memcpy(&s, sp, min(sizeof(s), bkey_val_bytes(k.k)));
+
+	unsigned nr_data = s.nr_blocks - s.nr_redundant;
+
+	prt_printf(out, "algo %u sectors %u blocks %u:%u csum ",
+		   s.algorithm,
+		   le16_to_cpu(s.sectors),
+		   nr_data,
+		   s.nr_redundant);
+	bch2_prt_csum_type(out, s.csum_type);
+	prt_printf(out, " gran %u", 1U << s.csum_granularity_bits);
+
+	for (unsigned i = 0; i < s.nr_blocks; i++) {
+		const struct bch_extent_ptr *ptr = sp->ptrs + i;
+
+		if ((void *) ptr >= bkey_val_end(k))
+			break;
+
+		bch2_extent_ptr_to_text(out, c, ptr);
 
-	prt_printf(out, "algo %u sectors %u blocks %u:%u csum %u gran %u",
-	       s->algorithm,
-	       le16_to_cpu(s->sectors),
-	       nr_data,
-	       s->nr_redundant,
-	       s->csum_type,
-	       1U << s->csum_granularity_bits);
-
-	for (i = 0; i < s->nr_blocks; i++) {
-		const struct bch_extent_ptr *ptr = s->ptrs + i;
-		struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev);
-		u32 offset;
-		u64 b = sector_to_bucket_and_offset(ca, ptr->offset, &offset);
-
-		prt_printf(out, " %u:%llu:%u", ptr->dev, b, offset);
-		if (i < nr_data)
-			prt_printf(out, "#%u", stripe_blockcount_get(s, i));
-		prt_printf(out, " gen %u", ptr->gen);
-		if (ptr_stale(ca, ptr))
-			prt_printf(out, " stale");
+		if (s.csum_type < BCH_CSUM_NR &&
+		    i < nr_data &&
+		    stripe_blockcount_offset(&s, i) < bkey_val_bytes(k.k))
+			prt_printf(out,  "#%u", stripe_blockcount_get(sp, i));
 	}
 }
 
@@ -607,10 +611,8 @@ static void ec_validate_checksums(struct bch_fs *c, struct ec_stripe_buf *buf)
 				struct printbuf err = PRINTBUF;
 				struct bch_dev *ca = bch_dev_bkey_exists(c, v->ptrs[i].dev);
 
-				prt_printf(&err, "stripe checksum error: expected %0llx:%0llx got %0llx:%0llx (type %s)\n",
-					   want.hi, want.lo,
-					   got.hi, got.lo,
-					   bch2_csum_types[v->csum_type]);
+				prt_str(&err, "stripe ");
+				bch2_csum_err_msg(&err, v->csum_type, want, got);
 				prt_printf(&err, "  for %ps at %u of\n  ", (void *) _RET_IP_, i);
 				bch2_bkey_val_to_text(&err, c, bkey_i_to_s_c(&buf->key));
 				bch_err_ratelimited(ca, "%s", err.buf);
diff --git a/fs/bcachefs/ec.h b/fs/bcachefs/ec.h
index f4369b02e805..f042616888b0 100644
--- a/fs/bcachefs/ec.h
+++ b/fs/bcachefs/ec.h
@@ -32,6 +32,8 @@ static inline unsigned stripe_csums_per_device(const struct bch_stripe *s)
 static inline unsigned stripe_csum_offset(const struct bch_stripe *s,
 					  unsigned dev, unsigned csum_idx)
 {
+	EBUG_ON(s->csum_type >= BCH_CSUM_NR);
+
 	unsigned csum_bytes = bch_crc_bytes[s->csum_type];
 
 	return sizeof(struct bch_stripe) +
diff --git a/fs/bcachefs/extents.c b/fs/bcachefs/extents.c
index 0e3ca99fbd2d..1a331e539204 100644
--- a/fs/bcachefs/extents.c
+++ b/fs/bcachefs/extents.c
@@ -998,7 +998,9 @@ void bch2_extent_ptr_to_text(struct printbuf *out, struct bch_fs *c, const struc
 			prt_str(out, " cached");
 		if (ptr->unwritten)
 			prt_str(out, " unwritten");
-		if (ca && ptr_stale(ca, ptr))
+		if (b >= ca->mi.first_bucket &&
+		    b <  ca->mi.nbuckets &&
+		    ptr_stale(ca, ptr))
 			prt_printf(out, " stale");
 	}
 }
@@ -1028,11 +1030,12 @@ void bch2_bkey_ptrs_to_text(struct printbuf *out, struct bch_fs *c,
 			struct bch_extent_crc_unpacked crc =
 				bch2_extent_crc_unpack(k.k, entry_to_crc(entry));
 
-			prt_printf(out, "crc: c_size %u size %u offset %u nonce %u csum %s compress ",
+			prt_printf(out, "crc: c_size %u size %u offset %u nonce %u csum ",
 			       crc.compressed_size,
 			       crc.uncompressed_size,
-			       crc.offset, crc.nonce,
-			       bch2_csum_types[crc.csum_type]);
+			       crc.offset, crc.nonce);
+			bch2_prt_csum_type(out, crc.csum_type);
+			prt_str(out, " compress ");
 			bch2_prt_compression_type(out, crc.compression_type);
 			break;
 		}
diff --git a/fs/bcachefs/eytzinger.c b/fs/bcachefs/eytzinger.c
index 4ce5e957a6e9..0f955c3c76a7 100644
--- a/fs/bcachefs/eytzinger.c
+++ b/fs/bcachefs/eytzinger.c
@@ -115,7 +115,7 @@ static void swap_bytes(void *a, void *b, size_t n)
 
 struct wrapper {
 	cmp_func_t cmp;
-	swap_func_t swap;
+	swap_func_t swap_func;
 };
 
 /*
@@ -125,7 +125,7 @@ struct wrapper {
 static void do_swap(void *a, void *b, size_t size, swap_r_func_t swap_func, const void *priv)
 {
 	if (swap_func == SWAP_WRAPPER) {
-		((const struct wrapper *)priv)->swap(a, b, (int)size);
+		((const struct wrapper *)priv)->swap_func(a, b, (int)size);
 		return;
 	}
 
@@ -174,7 +174,7 @@ void eytzinger0_sort_r(void *base, size_t n, size_t size,
 	int i, c, r;
 
 	/* called from 'sort' without swap function, let's pick the default */
-	if (swap_func == SWAP_WRAPPER && !((struct wrapper *)priv)->swap)
+	if (swap_func == SWAP_WRAPPER && !((struct wrapper *)priv)->swap_func)
 		swap_func = NULL;
 
 	if (!swap_func) {
@@ -227,7 +227,7 @@ void eytzinger0_sort(void *base, size_t n, size_t size,
 {
 	struct wrapper w = {
 		.cmp  = cmp_func,
-		.swap = swap_func,
+		.swap_func = swap_func,
 	};
 
 	return eytzinger0_sort_r(base, n, size, _CMP_WRAPPER, SWAP_WRAPPER, &w);
diff --git a/fs/bcachefs/eytzinger.h b/fs/bcachefs/eytzinger.h
index ee0e2df33322..24840aee335c 100644
--- a/fs/bcachefs/eytzinger.h
+++ b/fs/bcachefs/eytzinger.h
@@ -242,8 +242,8 @@ static inline unsigned inorder_to_eytzinger0(unsigned i, unsigned size)
 	     (_i) = eytzinger0_next((_i), (_size)))
 
 /* return greatest node <= @search, or -1 if not found */
-static inline ssize_t eytzinger0_find_le(void *base, size_t nr, size_t size,
-					 cmp_func_t cmp, const void *search)
+static inline int eytzinger0_find_le(void *base, size_t nr, size_t size,
+				     cmp_func_t cmp, const void *search)
 {
 	unsigned i, n = 0;
 
@@ -256,18 +256,32 @@ static inline ssize_t eytzinger0_find_le(void *base, size_t nr, size_t size,
 	} while (n < nr);
 
 	if (n & 1) {
-		/* @i was greater than @search, return previous node: */
+		/*
+		 * @i was greater than @search, return previous node:
+		 *
+		 * if @i was leftmost/smallest element,
+		 * eytzinger0_prev(eytzinger0_first())) returns -1, as expected
+		 */
 		return eytzinger0_prev(i, nr);
 	} else {
 		return i;
 	}
 }
 
-static inline ssize_t eytzinger0_find_gt(void *base, size_t nr, size_t size,
-					 cmp_func_t cmp, const void *search)
+static inline int eytzinger0_find_gt(void *base, size_t nr, size_t size,
+				     cmp_func_t cmp, const void *search)
 {
 	ssize_t idx = eytzinger0_find_le(base, nr, size, cmp, search);
-	return eytzinger0_next(idx, size);
+
+	/*
+	 * if eytitzinger0_find_le() returned -1 - no element was <= search - we
+	 * want to return the first element; next/prev identities mean this work
+	 * as expected
+	 *
+	 * similarly if find_le() returns last element, we should return -1;
+	 * identities mean this all works out:
+	 */
+	return eytzinger0_next(idx, nr);
 }
 
 #define eytzinger0_find(base, nr, size, _cmp, search)			\
diff --git a/fs/bcachefs/fs-io-direct.c b/fs/bcachefs/fs-io-direct.c
index f49e6c0f0f68..b889370a5088 100644
--- a/fs/bcachefs/fs-io-direct.c
+++ b/fs/bcachefs/fs-io-direct.c
@@ -387,6 +387,8 @@ static __always_inline long bch2_dio_write_done(struct dio_write *dio)
 	ret = dio->op.error ?: ((long) dio->written << 9);
 	bio_put(&dio->op.wbio.bio);
 
+	bch2_write_ref_put(dio->op.c, BCH_WRITE_REF_dio_write);
+
 	/* inode->i_dio_count is our ref on inode and thus bch_fs */
 	inode_dio_end(&inode->v);
 
@@ -590,22 +592,25 @@ ssize_t bch2_direct_write(struct kiocb *req, struct iov_iter *iter)
 	prefetch(&inode->ei_inode);
 	prefetch((void *) &inode->ei_inode + 64);
 
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_dio_write))
+		return -EROFS;
+
 	inode_lock(&inode->v);
 
 	ret = generic_write_checks(req, iter);
 	if (unlikely(ret <= 0))
-		goto err;
+		goto err_put_write_ref;
 
 	ret = file_remove_privs(file);
 	if (unlikely(ret))
-		goto err;
+		goto err_put_write_ref;
 
 	ret = file_update_time(file);
 	if (unlikely(ret))
-		goto err;
+		goto err_put_write_ref;
 
 	if (unlikely((req->ki_pos|iter->count) & (block_bytes(c) - 1)))
-		goto err;
+		goto err_put_write_ref;
 
 	inode_dio_begin(&inode->v);
 	bch2_pagecache_block_get(inode);
@@ -645,7 +650,7 @@ ssize_t bch2_direct_write(struct kiocb *req, struct iov_iter *iter)
 	}
 
 	ret = bch2_dio_write_loop(dio);
-err:
+out:
 	if (locked)
 		inode_unlock(&inode->v);
 	return ret;
@@ -653,7 +658,9 @@ err_put_bio:
 	bch2_pagecache_block_put(inode);
 	bio_put(bio);
 	inode_dio_end(&inode->v);
-	goto err;
+err_put_write_ref:
+	bch2_write_ref_put(c, BCH_WRITE_REF_dio_write);
+	goto out;
 }
 
 void bch2_fs_fs_io_direct_exit(struct bch_fs *c)
diff --git a/fs/bcachefs/fs-io.c b/fs/bcachefs/fs-io.c
index 8c70123b6a0c..20b40477425f 100644
--- a/fs/bcachefs/fs-io.c
+++ b/fs/bcachefs/fs-io.c
@@ -174,18 +174,18 @@ void __bch2_i_sectors_acct(struct bch_fs *c, struct bch_inode_info *inode,
 static int bch2_flush_inode(struct bch_fs *c,
 			    struct bch_inode_info *inode)
 {
-	struct bch_inode_unpacked u;
-	int ret;
-
 	if (c->opts.journal_flush_disabled)
 		return 0;
 
-	ret = bch2_inode_find_by_inum(c, inode_inum(inode), &u);
-	if (ret)
-		return ret;
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_fsync))
+		return -EROFS;
 
-	return bch2_journal_flush_seq(&c->journal, u.bi_journal_seq) ?:
-		bch2_inode_flush_nocow_writes(c, inode);
+	struct bch_inode_unpacked u;
+	int ret = bch2_inode_find_by_inum(c, inode_inum(inode), &u) ?:
+		  bch2_journal_flush_seq(&c->journal, u.bi_journal_seq) ?:
+		  bch2_inode_flush_nocow_writes(c, inode);
+	bch2_write_ref_put(c, BCH_WRITE_REF_fsync);
+	return ret;
 }
 
 int bch2_fsync(struct file *file, loff_t start, loff_t end, int datasync)
diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c
index b5ea9fa1259d..fce690007edf 100644
--- a/fs/bcachefs/fs.c
+++ b/fs/bcachefs/fs.c
@@ -188,7 +188,8 @@ static struct bch_inode_info *bch2_inode_insert(struct bch_fs *c, struct bch_ino
 	BUG_ON(!old);
 
 	if (unlikely(old != inode)) {
-		discard_new_inode(&inode->v);
+		__destroy_inode(&inode->v);
+		kmem_cache_free(bch2_inode_cache, inode);
 		inode = old;
 	} else {
 		mutex_lock(&c->vfs_inodes_lock);
@@ -225,8 +226,10 @@ static struct bch_inode_info *bch2_new_inode(struct btree_trans *trans)
 
 	if (unlikely(!inode)) {
 		int ret = drop_locks_do(trans, (inode = to_bch_ei(new_inode(c->vfs_sb))) ? 0 : -ENOMEM);
-		if (ret && inode)
-			discard_new_inode(&inode->v);
+		if (ret && inode) {
+			__destroy_inode(&inode->v);
+			kmem_cache_free(bch2_inode_cache, inode);
+		}
 		if (ret)
 			return ERR_PTR(ret);
 	}
diff --git a/fs/bcachefs/inode.c b/fs/bcachefs/inode.c
index ca4a066e9a54..0f95d7fb5ec0 100644
--- a/fs/bcachefs/inode.c
+++ b/fs/bcachefs/inode.c
@@ -606,7 +606,7 @@ int bch2_trigger_inode(struct btree_trans *trans,
 		       struct bkey_s new,
 		       unsigned flags)
 {
-	s64 nr = bkey_is_inode(new.k) - bkey_is_inode(old.k);
+	s64 nr = (s64) bkey_is_inode(new.k) - (s64) bkey_is_inode(old.k);
 
 	if (flags & BTREE_TRIGGER_TRANSACTIONAL) {
 		if (nr) {
diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c
index 725fcf46f631..eb1f9d6f5a19 100644
--- a/fs/bcachefs/journal_io.c
+++ b/fs/bcachefs/journal_io.c
@@ -247,7 +247,7 @@ static void journal_entry_err_msg(struct printbuf *out,
 
 	if (entry) {
 		prt_str(out, " type=");
-		prt_str(out, bch2_jset_entry_types[entry->type]);
+		bch2_prt_jset_entry_type(out, entry->type);
 	}
 
 	if (!jset) {
@@ -403,7 +403,8 @@ static void journal_entry_btree_keys_to_text(struct printbuf *out, struct bch_fs
 	jset_entry_for_each_key(entry, k) {
 		if (!first) {
 			prt_newline(out);
-			prt_printf(out, "%s: ", bch2_jset_entry_types[entry->type]);
+			bch2_prt_jset_entry_type(out, entry->type);
+			prt_str(out, ": ");
 		}
 		prt_printf(out, "btree=%s l=%u ", bch2_btree_id_str(entry->btree_id), entry->level);
 		bch2_bkey_val_to_text(out, c, bkey_i_to_s_c(k));
@@ -563,9 +564,9 @@ static void journal_entry_usage_to_text(struct printbuf *out, struct bch_fs *c,
 	struct jset_entry_usage *u =
 		container_of(entry, struct jset_entry_usage, entry);
 
-	prt_printf(out, "type=%s v=%llu",
-	       bch2_fs_usage_types[u->entry.btree_id],
-	       le64_to_cpu(u->v));
+	prt_str(out, "type=");
+	bch2_prt_fs_usage_type(out, u->entry.btree_id);
+	prt_printf(out, " v=%llu", le64_to_cpu(u->v));
 }
 
 static int journal_entry_data_usage_validate(struct bch_fs *c,
@@ -827,11 +828,11 @@ int bch2_journal_entry_validate(struct bch_fs *c,
 void bch2_journal_entry_to_text(struct printbuf *out, struct bch_fs *c,
 				struct jset_entry *entry)
 {
+	bch2_prt_jset_entry_type(out, entry->type);
+
 	if (entry->type < BCH_JSET_ENTRY_NR) {
-		prt_printf(out, "%s: ", bch2_jset_entry_types[entry->type]);
+		prt_str(out, ": ");
 		bch2_jset_entry_ops[entry->type].to_text(out, c, entry);
-	} else {
-		prt_printf(out, "(unknown type %u)", entry->type);
 	}
 }
 
@@ -1722,7 +1723,7 @@ static void journal_write_endio(struct bio *bio)
 	percpu_ref_put(&ca->io_ref);
 }
 
-static CLOSURE_CALLBACK(do_journal_write)
+static CLOSURE_CALLBACK(journal_write_submit)
 {
 	closure_type(w, struct journal_buf, io);
 	struct journal *j = container_of(w, struct journal, buf[w->idx]);
@@ -1767,6 +1768,44 @@ static CLOSURE_CALLBACK(do_journal_write)
 	continue_at(cl, journal_write_done, j->wq);
 }
 
+static CLOSURE_CALLBACK(journal_write_preflush)
+{
+	closure_type(w, struct journal_buf, io);
+	struct journal *j = container_of(w, struct journal, buf[w->idx]);
+	struct bch_fs *c = container_of(j, struct bch_fs, journal);
+
+	if (j->seq_ondisk + 1 != le64_to_cpu(w->data->seq)) {
+		spin_lock(&j->lock);
+		closure_wait(&j->async_wait, cl);
+		spin_unlock(&j->lock);
+
+		continue_at(cl, journal_write_preflush, j->wq);
+		return;
+	}
+
+	if (w->separate_flush) {
+		for_each_rw_member(c, ca) {
+			percpu_ref_get(&ca->io_ref);
+
+			struct journal_device *ja = &ca->journal;
+			struct bio *bio = &ja->bio[w->idx]->bio;
+			bio_reset(bio, ca->disk_sb.bdev,
+				  REQ_OP_WRITE|REQ_SYNC|REQ_META|REQ_PREFLUSH);
+			bio->bi_end_io		= journal_write_endio;
+			bio->bi_private		= ca;
+			closure_bio_submit(bio, cl);
+		}
+
+		continue_at(cl, journal_write_submit, j->wq);
+	} else {
+		/*
+		 * no need to punt to another work item if we're not waiting on
+		 * preflushes
+		 */
+		journal_write_submit(&cl->work);
+	}
+}
+
 static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w)
 {
 	struct bch_fs *c = container_of(j, struct bch_fs, journal);
@@ -2032,23 +2071,9 @@ CLOSURE_CALLBACK(bch2_journal_write)
 		goto err;
 
 	if (!JSET_NO_FLUSH(w->data))
-		closure_wait_event(&j->async_wait, j->seq_ondisk + 1 == le64_to_cpu(w->data->seq));
-
-	if (!JSET_NO_FLUSH(w->data) && w->separate_flush) {
-		for_each_rw_member(c, ca) {
-			percpu_ref_get(&ca->io_ref);
-
-			struct journal_device *ja = &ca->journal;
-			struct bio *bio = &ja->bio[w->idx]->bio;
-			bio_reset(bio, ca->disk_sb.bdev,
-				  REQ_OP_WRITE|REQ_SYNC|REQ_META|REQ_PREFLUSH);
-			bio->bi_end_io		= journal_write_endio;
-			bio->bi_private		= ca;
-			closure_bio_submit(bio, cl);
-		}
-	}
-
-	continue_at(cl, do_journal_write, j->wq);
+		continue_at(cl, journal_write_preflush, j->wq);
+	else
+		continue_at(cl, journal_write_submit, j->wq);
 	return;
 no_io:
 	continue_at(cl, journal_write_done, j->wq);
diff --git a/fs/bcachefs/journal_reclaim.c b/fs/bcachefs/journal_reclaim.c
index ab811c0dad26..04a577848b01 100644
--- a/fs/bcachefs/journal_reclaim.c
+++ b/fs/bcachefs/journal_reclaim.c
@@ -67,6 +67,8 @@ void bch2_journal_set_watermark(struct journal *j)
 	    track_event_change(&c->times[BCH_TIME_blocked_write_buffer_full], low_on_wb))
 		trace_and_count(c, journal_full, c);
 
+	mod_bit(JOURNAL_SPACE_LOW, &j->flags, low_on_space || low_on_pin);
+
 	swap(watermark, j->watermark);
 	if (watermark > j->watermark)
 		journal_wake(j);
diff --git a/fs/bcachefs/journal_types.h b/fs/bcachefs/journal_types.h
index 8c053cb64ca5..b5161b5d76a0 100644
--- a/fs/bcachefs/journal_types.h
+++ b/fs/bcachefs/journal_types.h
@@ -134,6 +134,7 @@ enum journal_flags {
 	JOURNAL_STARTED,
 	JOURNAL_MAY_SKIP_FLUSH,
 	JOURNAL_NEED_FLUSH_WRITE,
+	JOURNAL_SPACE_LOW,
 };
 
 /* Reasons we may fail to get a journal reservation: */
diff --git a/fs/bcachefs/opts.c b/fs/bcachefs/opts.c
index e1800c4119b5..bb068fd72465 100644
--- a/fs/bcachefs/opts.c
+++ b/fs/bcachefs/opts.c
@@ -43,7 +43,7 @@ const char * const __bch2_btree_ids[] = {
 	NULL
 };
 
-const char * const bch2_csum_types[] = {
+static const char * const __bch2_csum_types[] = {
 	BCH_CSUM_TYPES()
 	NULL
 };
@@ -53,7 +53,7 @@ const char * const bch2_csum_opts[] = {
 	NULL
 };
 
-const char * const __bch2_compression_types[] = {
+static const char * const __bch2_compression_types[] = {
 	BCH_COMPRESSION_TYPES()
 	NULL
 };
@@ -83,18 +83,39 @@ const char * const bch2_member_states[] = {
 	NULL
 };
 
-const char * const bch2_jset_entry_types[] = {
+static const char * const __bch2_jset_entry_types[] = {
 	BCH_JSET_ENTRY_TYPES()
 	NULL
 };
 
-const char * const bch2_fs_usage_types[] = {
+static const char * const __bch2_fs_usage_types[] = {
 	BCH_FS_USAGE_TYPES()
 	NULL
 };
 
 #undef x
 
+static void prt_str_opt_boundscheck(struct printbuf *out, const char * const opts[],
+				    unsigned nr, const char *type, unsigned idx)
+{
+	if (idx < nr)
+		prt_str(out, opts[idx]);
+	else
+		prt_printf(out, "(unknown %s %u)", type, idx);
+}
+
+#define PRT_STR_OPT_BOUNDSCHECKED(name, type)					\
+void bch2_prt_##name(struct printbuf *out, type t)				\
+{										\
+	prt_str_opt_boundscheck(out, __bch2_##name##s, ARRAY_SIZE(__bch2_##name##s) - 1, #name, t);\
+}
+
+PRT_STR_OPT_BOUNDSCHECKED(jset_entry_type,	enum bch_jset_entry_type);
+PRT_STR_OPT_BOUNDSCHECKED(fs_usage_type,	enum bch_fs_usage_type);
+PRT_STR_OPT_BOUNDSCHECKED(data_type,		enum bch_data_type);
+PRT_STR_OPT_BOUNDSCHECKED(csum_type,		enum bch_csum_type);
+PRT_STR_OPT_BOUNDSCHECKED(compression_type,	enum bch_compression_type);
+
 static int bch2_opt_fix_errors_parse(struct bch_fs *c, const char *val, u64 *res,
 				     struct printbuf *err)
 {
diff --git a/fs/bcachefs/opts.h b/fs/bcachefs/opts.h
index 1ac4135cca1c..84e452835a17 100644
--- a/fs/bcachefs/opts.h
+++ b/fs/bcachefs/opts.h
@@ -16,18 +16,20 @@ extern const char * const bch2_version_upgrade_opts[];
 extern const char * const bch2_sb_features[];
 extern const char * const bch2_sb_compat[];
 extern const char * const __bch2_btree_ids[];
-extern const char * const bch2_csum_types[];
 extern const char * const bch2_csum_opts[];
-extern const char * const __bch2_compression_types[];
 extern const char * const bch2_compression_opts[];
 extern const char * const bch2_str_hash_types[];
 extern const char * const bch2_str_hash_opts[];
 extern const char * const __bch2_data_types[];
 extern const char * const bch2_member_states[];
-extern const char * const bch2_jset_entry_types[];
-extern const char * const bch2_fs_usage_types[];
 extern const char * const bch2_d_types[];
 
+void bch2_prt_jset_entry_type(struct printbuf *,	enum bch_jset_entry_type);
+void bch2_prt_fs_usage_type(struct printbuf *,		enum bch_fs_usage_type);
+void bch2_prt_data_type(struct printbuf *,		enum bch_data_type);
+void bch2_prt_csum_type(struct printbuf *,		enum bch_csum_type);
+void bch2_prt_compression_type(struct printbuf *,	enum bch_compression_type);
+
 static inline const char *bch2_d_type_str(unsigned d_type)
 {
 	return (d_type < BCH_DT_MAX ? bch2_d_types[d_type] : NULL) ?: "(bad d_type)";
diff --git a/fs/bcachefs/recovery.c b/fs/bcachefs/recovery.c
index b76c16152579..be5b47619327 100644
--- a/fs/bcachefs/recovery.c
+++ b/fs/bcachefs/recovery.c
@@ -47,20 +47,6 @@ void bch2_btree_lost_data(struct bch_fs *c, enum btree_id btree)
 	}
 }
 
-static bool btree_id_is_alloc(enum btree_id id)
-{
-	switch (id) {
-	case BTREE_ID_alloc:
-	case BTREE_ID_backpointers:
-	case BTREE_ID_need_discard:
-	case BTREE_ID_freespace:
-	case BTREE_ID_bucket_gens:
-		return true;
-	default:
-		return false;
-	}
-}
-
 /* for -o reconstruct_alloc: */
 static void bch2_reconstruct_alloc(struct bch_fs *c)
 {
@@ -263,7 +249,10 @@ int bch2_journal_replay(struct bch_fs *c)
 
 		struct journal_key *k = *kp;
 
-		replay_now_at(j, k->journal_seq);
+		if (k->journal_seq)
+			replay_now_at(j, k->journal_seq);
+		else
+			replay_now_at(j, j->replay_journal_seq_end);
 
 		ret = commit_do(trans, NULL, NULL,
 				BCH_TRANS_COMMIT_no_enospc|
diff --git a/fs/bcachefs/recovery_passes.c b/fs/bcachefs/recovery_passes.c
index cb501460d615..0cec0f7d9703 100644
--- a/fs/bcachefs/recovery_passes.c
+++ b/fs/bcachefs/recovery_passes.c
@@ -44,7 +44,7 @@ static int bch2_set_may_go_rw(struct bch_fs *c)
 
 	set_bit(BCH_FS_may_go_rw, &c->flags);
 
-	if (keys->nr || c->opts.fsck || !c->sb.clean)
+	if (keys->nr || c->opts.fsck || !c->sb.clean || c->recovery_passes_explicit)
 		return bch2_fs_read_write_early(c);
 	return 0;
 }
diff --git a/fs/bcachefs/sb-clean.c b/fs/bcachefs/sb-clean.c
index 5980ba2563fe..35ca3f138de6 100644
--- a/fs/bcachefs/sb-clean.c
+++ b/fs/bcachefs/sb-clean.c
@@ -29,6 +29,14 @@ int bch2_sb_clean_validate_late(struct bch_fs *c, struct bch_sb_field_clean *cle
 	for (entry = clean->start;
 	     entry < (struct jset_entry *) vstruct_end(&clean->field);
 	     entry = vstruct_next(entry)) {
+		if (vstruct_end(entry) > vstruct_end(&clean->field)) {
+			bch_err(c, "journal entry (u64s %u) overran end of superblock clean section (u64s %u) by %zu",
+				le16_to_cpu(entry->u64s), le32_to_cpu(clean->field.u64s),
+				(u64 *) vstruct_end(entry) - (u64 *) vstruct_end(&clean->field));
+			bch2_sb_error_count(c, BCH_FSCK_ERR_sb_clean_entry_overrun);
+			return -BCH_ERR_fsck_repair_unimplemented;
+		}
+
 		ret = bch2_journal_entry_validate(c, NULL, entry,
 						  le16_to_cpu(c->disk_sb.sb->version),
 						  BCH_SB_BIG_ENDIAN(c->disk_sb.sb),
diff --git a/fs/bcachefs/sb-downgrade.c b/fs/bcachefs/sb-downgrade.c
index d6f81179c3a2..a98ef940b7a3 100644
--- a/fs/bcachefs/sb-downgrade.c
+++ b/fs/bcachefs/sb-downgrade.c
@@ -51,7 +51,10 @@
 	  BCH_FSCK_ERR_subvol_fs_path_parent_wrong)		\
 	x(btree_subvolume_children,				\
 	  BIT_ULL(BCH_RECOVERY_PASS_check_subvols),		\
-	  BCH_FSCK_ERR_subvol_children_not_set)
+	  BCH_FSCK_ERR_subvol_children_not_set)			\
+	x(mi_btree_bitmap,					\
+	  BIT_ULL(BCH_RECOVERY_PASS_check_allocations),		\
+	  BCH_FSCK_ERR_btree_bitmap_not_marked)
 
 #define DOWNGRADE_TABLE()
 
diff --git a/fs/bcachefs/sb-errors_types.h b/fs/bcachefs/sb-errors_types.h
index d7d609131030..06c7a644f4a4 100644
--- a/fs/bcachefs/sb-errors_types.h
+++ b/fs/bcachefs/sb-errors_types.h
@@ -130,7 +130,7 @@
 	x(bucket_gens_nonzero_for_invalid_buckets,		122)	\
 	x(need_discard_freespace_key_to_invalid_dev_bucket,	123)	\
 	x(need_discard_freespace_key_bad,			124)	\
-	x(backpointer_pos_wrong,				125)	\
+	x(backpointer_bucket_offset_wrong,			125)	\
 	x(backpointer_to_missing_device,			126)	\
 	x(backpointer_to_missing_alloc,				127)	\
 	x(backpointer_to_missing_ptr,				128)	\
@@ -270,7 +270,9 @@
 	x(btree_ptr_v2_min_key_bad,				262)	\
 	x(btree_root_unreadable_and_scan_found_nothing,		263)	\
 	x(snapshot_node_missing,				264)	\
-	x(dup_backpointer_to_bad_csum_extent,			265)
+	x(dup_backpointer_to_bad_csum_extent,			265)	\
+	x(btree_bitmap_not_marked,				266)	\
+	x(sb_clean_entry_overrun,				267)
 
 enum bch_sb_error_id {
 #define x(t, n) BCH_FSCK_ERR_##t = n,
diff --git a/fs/bcachefs/sb-members.c b/fs/bcachefs/sb-members.c
index eff5ce18c69c..5b8e621ac5eb 100644
--- a/fs/bcachefs/sb-members.c
+++ b/fs/bcachefs/sb-members.c
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 
 #include "bcachefs.h"
+#include "btree_cache.h"
 #include "disk_groups.h"
 #include "opts.h"
 #include "replicas.h"
@@ -426,3 +427,55 @@ void bch2_dev_errors_reset(struct bch_dev *ca)
 	bch2_write_super(c);
 	mutex_unlock(&c->sb_lock);
 }
+
+/*
+ * Per member "range has btree nodes" bitmap:
+ *
+ * This is so that if we ever have to run the btree node scan to repair we don't
+ * have to scan full devices:
+ */
+
+bool bch2_dev_btree_bitmap_marked(struct bch_fs *c, struct bkey_s_c k)
+{
+	bkey_for_each_ptr(bch2_bkey_ptrs_c(k), ptr)
+		if (!bch2_dev_btree_bitmap_marked_sectors(bch_dev_bkey_exists(c, ptr->dev),
+							  ptr->offset, btree_sectors(c)))
+			return false;
+	return true;
+}
+
+static void __bch2_dev_btree_bitmap_mark(struct bch_sb_field_members_v2 *mi, unsigned dev,
+				u64 start, unsigned sectors)
+{
+	struct bch_member *m = __bch2_members_v2_get_mut(mi, dev);
+	u64 bitmap = le64_to_cpu(m->btree_allocated_bitmap);
+
+	u64 end = start + sectors;
+
+	int resize = ilog2(roundup_pow_of_two(end)) - (m->btree_bitmap_shift + 6);
+	if (resize > 0) {
+		u64 new_bitmap = 0;
+
+		for (unsigned i = 0; i < 64; i++)
+			if (bitmap & BIT_ULL(i))
+				new_bitmap |= BIT_ULL(i >> resize);
+		bitmap = new_bitmap;
+		m->btree_bitmap_shift += resize;
+	}
+
+	for (unsigned bit = start >> m->btree_bitmap_shift;
+	     (u64) bit << m->btree_bitmap_shift < end;
+	     bit++)
+		bitmap |= BIT_ULL(bit);
+
+	m->btree_allocated_bitmap = cpu_to_le64(bitmap);
+}
+
+void bch2_dev_btree_bitmap_mark(struct bch_fs *c, struct bkey_s_c k)
+{
+	lockdep_assert_held(&c->sb_lock);
+
+	struct bch_sb_field_members_v2 *mi = bch2_sb_field_get(c->disk_sb.sb, members_v2);
+	bkey_for_each_ptr(bch2_bkey_ptrs_c(k), ptr)
+		__bch2_dev_btree_bitmap_mark(mi, ptr->dev, ptr->offset, btree_sectors(c));
+}
diff --git a/fs/bcachefs/sb-members.h b/fs/bcachefs/sb-members.h
index be0a94183271..5efa64eca5f8 100644
--- a/fs/bcachefs/sb-members.h
+++ b/fs/bcachefs/sb-members.h
@@ -3,6 +3,7 @@
 #define _BCACHEFS_SB_MEMBERS_H
 
 #include "darray.h"
+#include "bkey_types.h"
 
 extern char * const bch2_member_error_strs[];
 
@@ -220,6 +221,8 @@ static inline struct bch_member_cpu bch2_mi_to_cpu(struct bch_member *mi)
 			: 1,
 		.freespace_initialized = BCH_MEMBER_FREESPACE_INITIALIZED(mi),
 		.valid		= bch2_member_exists(mi),
+		.btree_bitmap_shift	= mi->btree_bitmap_shift,
+		.btree_allocated_bitmap = le64_to_cpu(mi->btree_allocated_bitmap),
 	};
 }
 
@@ -228,4 +231,22 @@ void bch2_sb_members_from_cpu(struct bch_fs *);
 void bch2_dev_io_errors_to_text(struct printbuf *, struct bch_dev *);
 void bch2_dev_errors_reset(struct bch_dev *);
 
+static inline bool bch2_dev_btree_bitmap_marked_sectors(struct bch_dev *ca, u64 start, unsigned sectors)
+{
+	u64 end = start + sectors;
+
+	if (end > 64ULL << ca->mi.btree_bitmap_shift)
+		return false;
+
+	for (unsigned bit = start >> ca->mi.btree_bitmap_shift;
+	     (u64) bit << ca->mi.btree_bitmap_shift < end;
+	     bit++)
+		if (!(ca->mi.btree_allocated_bitmap & BIT_ULL(bit)))
+			return false;
+	return true;
+}
+
+bool bch2_dev_btree_bitmap_marked(struct bch_fs *, struct bkey_s_c);
+void bch2_dev_btree_bitmap_mark(struct bch_fs *, struct bkey_s_c);
+
 #endif /* _BCACHEFS_SB_MEMBERS_H */
diff --git a/fs/bcachefs/snapshot.c b/fs/bcachefs/snapshot.c
index 0e806f04f3d7..544322d5c251 100644
--- a/fs/bcachefs/snapshot.c
+++ b/fs/bcachefs/snapshot.c
@@ -125,6 +125,15 @@ static inline u32 get_ancestor_below(struct snapshot_table *t, u32 id, u32 ances
 	return s->parent;
 }
 
+static bool test_ancestor_bitmap(struct snapshot_table *t, u32 id, u32 ancestor)
+{
+	const struct snapshot_t *s = __snapshot_t(t, id);
+	if (!s)
+		return false;
+
+	return test_bit(ancestor - id - 1, s->is_ancestor);
+}
+
 bool __bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ancestor)
 {
 	bool ret;
@@ -140,13 +149,11 @@ bool __bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ancestor)
 	while (id && id < ancestor - IS_ANCESTOR_BITMAP)
 		id = get_ancestor_below(t, id, ancestor);
 
-	if (id && id < ancestor) {
-		ret = test_bit(ancestor - id - 1, __snapshot_t(t, id)->is_ancestor);
+	ret = id && id < ancestor
+		? test_ancestor_bitmap(t, id, ancestor)
+		: id == ancestor;
 
-		EBUG_ON(ret != __bch2_snapshot_is_ancestor_early(t, id, ancestor));
-	} else {
-		ret = id == ancestor;
-	}
+	EBUG_ON(ret != __bch2_snapshot_is_ancestor_early(t, id, ancestor));
 out:
 	rcu_read_unlock();
 
diff --git a/fs/bcachefs/super-io.c b/fs/bcachefs/super-io.c
index 5eee055ee272..08ea3dbbbe97 100644
--- a/fs/bcachefs/super-io.c
+++ b/fs/bcachefs/super-io.c
@@ -700,8 +700,11 @@ retry:
 		return -ENOMEM;
 
 	sb->sb_name = kstrdup(path, GFP_KERNEL);
-	if (!sb->sb_name)
-		return -ENOMEM;
+	if (!sb->sb_name) {
+		ret = -ENOMEM;
+		prt_printf(&err, "error allocating memory for sb_name");
+		goto err;
+	}
 
 #ifndef __KERNEL__
 	if (opt_get(*opts, direct_io) == false)
diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c
index ed63018f21be..88e214c609bb 100644
--- a/fs/bcachefs/super.c
+++ b/fs/bcachefs/super.c
@@ -288,8 +288,13 @@ static void __bch2_fs_read_only(struct bch_fs *c)
 	if (test_bit(JOURNAL_REPLAY_DONE, &c->journal.flags) &&
 	    !test_bit(BCH_FS_emergency_ro, &c->flags))
 		set_bit(BCH_FS_clean_shutdown, &c->flags);
+
 	bch2_fs_journal_stop(&c->journal);
 
+	bch_info(c, "%sshutdown complete, journal seq %llu",
+		 test_bit(BCH_FS_clean_shutdown, &c->flags) ? "" : "un",
+		 c->journal.seq_ondisk);
+
 	/*
 	 * After stopping journal:
 	 */
@@ -539,6 +544,7 @@ static void __bch2_fs_free(struct bch_fs *c)
 
 	bch2_find_btree_nodes_exit(&c->found_btree_nodes);
 	bch2_free_pending_node_rewrites(c);
+	bch2_fs_allocator_background_exit(c);
 	bch2_fs_sb_errors_exit(c);
 	bch2_fs_counters_exit(c);
 	bch2_fs_snapshots_exit(c);
diff --git a/fs/bcachefs/super_types.h b/fs/bcachefs/super_types.h
index ec784d975f66..11bcef170c2c 100644
--- a/fs/bcachefs/super_types.h
+++ b/fs/bcachefs/super_types.h
@@ -37,6 +37,8 @@ struct bch_member_cpu {
 	u8			durability;
 	u8			freespace_initialized;
 	u8			valid;
+	u8			btree_bitmap_shift;
+	u64			btree_allocated_bitmap;
 };
 
 #endif /* _BCACHEFS_SUPER_TYPES_H */
diff --git a/fs/bcachefs/sysfs.c b/fs/bcachefs/sysfs.c
index c86a93a8d8fc..5be92fe3f4ea 100644
--- a/fs/bcachefs/sysfs.c
+++ b/fs/bcachefs/sysfs.c
@@ -17,7 +17,6 @@
 #include "btree_iter.h"
 #include "btree_key_cache.h"
 #include "btree_update.h"
-#include "btree_update_interior.h"
 #include "btree_gc.h"
 #include "buckets.h"
 #include "clock.h"
@@ -26,6 +25,7 @@
 #include "ec.h"
 #include "inode.h"
 #include "journal.h"
+#include "journal_reclaim.h"
 #include "keylist.h"
 #include "move.h"
 #include "movinggc.h"
@@ -139,6 +139,7 @@ do {									\
 write_attribute(trigger_gc);
 write_attribute(trigger_discards);
 write_attribute(trigger_invalidates);
+write_attribute(trigger_journal_flush);
 write_attribute(prune_cache);
 write_attribute(btree_wakeup);
 rw_attribute(btree_gc_periodic);
@@ -166,7 +167,6 @@ read_attribute(btree_write_stats);
 read_attribute(btree_cache_size);
 read_attribute(compression_stats);
 read_attribute(journal_debug);
-read_attribute(btree_updates);
 read_attribute(btree_cache);
 read_attribute(btree_key_cache);
 read_attribute(stripes_heap);
@@ -415,9 +415,6 @@ SHOW(bch2_fs)
 	if (attr == &sysfs_journal_debug)
 		bch2_journal_debug_to_text(out, &c->journal);
 
-	if (attr == &sysfs_btree_updates)
-		bch2_btree_updates_to_text(out, c);
-
 	if (attr == &sysfs_btree_cache)
 		bch2_btree_cache_to_text(out, c);
 
@@ -505,7 +502,7 @@ STORE(bch2_fs)
 
 	/* Debugging: */
 
-	if (!test_bit(BCH_FS_rw, &c->flags))
+	if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_sysfs))
 		return -EROFS;
 
 	if (attr == &sysfs_prune_cache) {
@@ -538,6 +535,11 @@ STORE(bch2_fs)
 	if (attr == &sysfs_trigger_invalidates)
 		bch2_do_invalidates(c);
 
+	if (attr == &sysfs_trigger_journal_flush) {
+		bch2_journal_flush_all_pins(&c->journal);
+		bch2_journal_meta(&c->journal);
+	}
+
 #ifdef CONFIG_BCACHEFS_TESTS
 	if (attr == &sysfs_perf_test) {
 		char *tmp = kstrdup(buf, GFP_KERNEL), *p = tmp;
@@ -558,6 +560,7 @@ STORE(bch2_fs)
 			size = ret;
 	}
 #endif
+	bch2_write_ref_put(c, BCH_WRITE_REF_sysfs);
 	return size;
 }
 SYSFS_OPS(bch2_fs);
@@ -639,7 +642,6 @@ SYSFS_OPS(bch2_fs_internal);
 struct attribute *bch2_fs_internal_files[] = {
 	&sysfs_flags,
 	&sysfs_journal_debug,
-	&sysfs_btree_updates,
 	&sysfs_btree_cache,
 	&sysfs_btree_key_cache,
 	&sysfs_new_stripes,
@@ -657,6 +659,7 @@ struct attribute *bch2_fs_internal_files[] = {
 	&sysfs_trigger_gc,
 	&sysfs_trigger_discards,
 	&sysfs_trigger_invalidates,
+	&sysfs_trigger_journal_flush,
 	&sysfs_prune_cache,
 	&sysfs_btree_wakeup,
 
diff --git a/fs/bcachefs/tests.c b/fs/bcachefs/tests.c
index b3fe9fc57747..bfec656f94c0 100644
--- a/fs/bcachefs/tests.c
+++ b/fs/bcachefs/tests.c
@@ -672,7 +672,7 @@ static int __do_delete(struct btree_trans *trans, struct bpos pos)
 
 	bch2_trans_iter_init(trans, &iter, BTREE_ID_xattrs, pos,
 			     BTREE_ITER_INTENT);
-	k = bch2_btree_iter_peek(&iter);
+	k = bch2_btree_iter_peek_upto(&iter, POS(0, U64_MAX));
 	ret = bkey_err(k);
 	if (ret)
 		goto err;
diff --git a/fs/bcachefs/thread_with_file.c b/fs/bcachefs/thread_with_file.c
index 940db15d6a93..b1af7ac430f6 100644
--- a/fs/bcachefs/thread_with_file.c
+++ b/fs/bcachefs/thread_with_file.c
@@ -294,16 +294,27 @@ static int thread_with_stdio_fn(void *arg)
 	return 0;
 }
 
-int bch2_run_thread_with_stdio(struct thread_with_stdio *thr,
-			       const struct thread_with_stdio_ops *ops)
+void bch2_thread_with_stdio_init(struct thread_with_stdio *thr,
+				 const struct thread_with_stdio_ops *ops)
 {
 	stdio_buf_init(&thr->stdio.input);
 	stdio_buf_init(&thr->stdio.output);
 	thr->ops = ops;
+}
 
+int __bch2_run_thread_with_stdio(struct thread_with_stdio *thr)
+{
 	return bch2_run_thread_with_file(&thr->thr, &thread_with_stdio_fops, thread_with_stdio_fn);
 }
 
+int bch2_run_thread_with_stdio(struct thread_with_stdio *thr,
+			       const struct thread_with_stdio_ops *ops)
+{
+	bch2_thread_with_stdio_init(thr, ops);
+
+	return __bch2_run_thread_with_stdio(thr);
+}
+
 int bch2_run_thread_with_stdout(struct thread_with_stdio *thr,
 				const struct thread_with_stdio_ops *ops)
 {
diff --git a/fs/bcachefs/thread_with_file.h b/fs/bcachefs/thread_with_file.h
index af54ea8f5b0f..1d63d14d7dca 100644
--- a/fs/bcachefs/thread_with_file.h
+++ b/fs/bcachefs/thread_with_file.h
@@ -63,6 +63,9 @@ struct thread_with_stdio {
 	const struct thread_with_stdio_ops	*ops;
 };
 
+void bch2_thread_with_stdio_init(struct thread_with_stdio *,
+				 const struct thread_with_stdio_ops *);
+int __bch2_run_thread_with_stdio(struct thread_with_stdio *);
 int bch2_run_thread_with_stdio(struct thread_with_stdio *,
 			       const struct thread_with_stdio_ops *);
 int bch2_run_thread_with_stdout(struct thread_with_stdio *,
diff --git a/fs/bcachefs/util.h b/fs/bcachefs/util.h
index b7e7c29278fc..5cf885b09986 100644
--- a/fs/bcachefs/util.h
+++ b/fs/bcachefs/util.h
@@ -788,6 +788,14 @@ static inline int copy_from_user_errcode(void *to, const void __user *from, unsi
 
 #endif
 
+static inline void mod_bit(long nr, volatile unsigned long *addr, bool v)
+{
+	if (v)
+		set_bit(nr, addr);
+	else
+		clear_bit(nr, addr);
+}
+
 static inline void __set_bit_le64(size_t bit, __le64 *addr)
 {
 	addr[bit / 64] |= cpu_to_le64(BIT_ULL(bit % 64));
@@ -795,7 +803,7 @@ static inline void __set_bit_le64(size_t bit, __le64 *addr)
 
 static inline void __clear_bit_le64(size_t bit, __le64 *addr)
 {
-	addr[bit / 64] &= !cpu_to_le64(BIT_ULL(bit % 64));
+	addr[bit / 64] &= ~cpu_to_le64(BIT_ULL(bit % 64));
 }
 
 static inline bool test_bit_le64(size_t bit, __le64 *addr)
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index c1e6a5bbeeaf..58110c968667 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -2776,20 +2776,14 @@ struct btrfs_data_container *init_data_container(u32 total_bytes)
 	size_t alloc_bytes;
 
 	alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
-	data = kvmalloc(alloc_bytes, GFP_KERNEL);
+	data = kvzalloc(alloc_bytes, GFP_KERNEL);
 	if (!data)
 		return ERR_PTR(-ENOMEM);
 
-	if (total_bytes >= sizeof(*data)) {
+	if (total_bytes >= sizeof(*data))
 		data->bytes_left = total_bytes - sizeof(*data);
-		data->bytes_missing = 0;
-	} else {
+	else
 		data->bytes_missing = sizeof(*data) - total_bytes;
-		data->bytes_left = 0;
-	}
-
-	data->elem_cnt = 0;
-	data->elem_missed = 0;
 
 	return data;
 }
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index dd6f566a383f..121ab890bd05 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -1133,6 +1133,9 @@ __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
 	if (ret)
 		return ret;
 
+	ret = btrfs_record_root_in_trans(trans, node->root);
+	if (ret)
+		return ret;
 	ret = btrfs_update_delayed_inode(trans, node->root, path, node);
 	return ret;
 }
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index beedd6ed64d3..257d044bca91 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3464,6 +3464,14 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
 	if (root_id != BTRFS_TREE_LOG_OBJECTID) {
 		struct btrfs_ref generic_ref = { 0 };
 
+		/*
+		 * Assert that the extent buffer is not cleared due to
+		 * EXTENT_BUFFER_ZONED_ZEROOUT. Please refer
+		 * btrfs_clear_buffer_dirty() and btree_csum_one_bio() for
+		 * detail.
+		 */
+		ASSERT(btrfs_header_bytenr(buf) != 0);
+
 		btrfs_init_generic_ref(&generic_ref, BTRFS_DROP_DELAYED_REF,
 				       buf->start, buf->len, parent,
 				       btrfs_header_owner(buf));
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 61594eaf1f89..2776112dbdf8 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -681,31 +681,21 @@ static void end_bbio_data_read(struct btrfs_bio *bbio)
 int btrfs_alloc_page_array(unsigned int nr_pages, struct page **page_array,
 			   gfp_t extra_gfp)
 {
+	const gfp_t gfp = GFP_NOFS | extra_gfp;
 	unsigned int allocated;
 
 	for (allocated = 0; allocated < nr_pages;) {
 		unsigned int last = allocated;
 
-		allocated = alloc_pages_bulk_array(GFP_NOFS | extra_gfp,
-						   nr_pages, page_array);
-
-		if (allocated == nr_pages)
-			return 0;
-
-		/*
-		 * During this iteration, no page could be allocated, even
-		 * though alloc_pages_bulk_array() falls back to alloc_page()
-		 * if  it could not bulk-allocate. So we must be out of memory.
-		 */
-		if (allocated == last) {
+		allocated = alloc_pages_bulk_array(gfp, nr_pages, page_array);
+		if (unlikely(allocated == last)) {
+			/* No progress, fail and do cleanup. */
 			for (int i = 0; i < allocated; i++) {
 				__free_page(page_array[i]);
 				page_array[i] = NULL;
 			}
 			return -ENOMEM;
 		}
-
-		memalloc_retry_wait(GFP_NOFS);
 	}
 	return 0;
 }
@@ -4154,7 +4144,7 @@ void btrfs_clear_buffer_dirty(struct btrfs_trans_handle *trans,
 	 * The actual zeroout of the buffer will happen later in
 	 * btree_csum_one_bio.
 	 */
-	if (btrfs_is_zoned(fs_info)) {
+	if (btrfs_is_zoned(fs_info) && test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
 		set_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags);
 		return;
 	}
@@ -4193,6 +4183,7 @@ void set_extent_buffer_dirty(struct extent_buffer *eb)
 	num_folios = num_extent_folios(eb);
 	WARN_ON(atomic_read(&eb->refs) == 0);
 	WARN_ON(!test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags));
+	WARN_ON(test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags));
 
 	if (!was_dirty) {
 		bool subpage = eb->fs_info->nodesize < PAGE_SIZE;
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index 445f7716f1e2..24a048210b15 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -817,7 +817,7 @@ void btrfs_drop_extent_map_range(struct btrfs_inode *inode, u64 start, u64 end,
 					split->block_len = em->block_len;
 					split->orig_start = em->orig_start;
 				} else {
-					const u64 diff = start + len - em->start;
+					const u64 diff = end - em->start;
 
 					split->block_len = split->len;
 					split->block_start += diff;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 37701531eeb1..7fed887e700c 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -1145,13 +1145,13 @@ static void submit_one_async_extent(struct async_chunk *async_chunk,
 				   0, *alloc_hint, &ins, 1, 1);
 	if (ret) {
 		/*
-		 * Here we used to try again by going back to non-compressed
-		 * path for ENOSPC.  But we can't reserve space even for
-		 * compressed size, how could it work for uncompressed size
-		 * which requires larger size?  So here we directly go error
-		 * path.
+		 * We can't reserve contiguous space for the compressed size.
+		 * Unlikely, but it's possible that we could have enough
+		 * non-contiguous space for the uncompressed size instead.  So
+		 * fall back to uncompressed.
 		 */
-		goto out_free;
+		submit_uncompressed_range(inode, async_extent, locked_page);
+		goto done;
 	}
 
 	/* Here we're doing allocation and writeback of the compressed pages */
@@ -1203,7 +1203,6 @@ done:
 out_free_reserve:
 	btrfs_dec_block_group_reservations(fs_info, ins.objectid);
 	btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 1);
-out_free:
 	mapping_set_error(inode->vfs_inode.i_mapping, -EIO);
 	extent_clear_unlock_delalloc(inode, start, end,
 				     NULL, EXTENT_LOCKED | EXTENT_DELALLOC |
@@ -2533,7 +2532,7 @@ void btrfs_clear_delalloc_extent(struct btrfs_inode *inode,
 		 */
 		if (bits & EXTENT_CLEAR_META_RESV &&
 		    root != fs_info->tree_root)
-			btrfs_delalloc_release_metadata(inode, len, false);
+			btrfs_delalloc_release_metadata(inode, len, true);
 
 		/* For sanity tests. */
 		if (btrfs_is_testing(fs_info))
@@ -4503,6 +4502,7 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 	struct btrfs_trans_handle *trans;
 	struct btrfs_block_rsv block_rsv;
 	u64 root_flags;
+	u64 qgroup_reserved = 0;
 	int ret;
 
 	down_write(&fs_info->subvol_sem);
@@ -4547,12 +4547,20 @@ int btrfs_delete_subvolume(struct btrfs_inode *dir, struct dentry *dentry)
 	ret = btrfs_subvolume_reserve_metadata(root, &block_rsv, 5, true);
 	if (ret)
 		goto out_undead;
+	qgroup_reserved = block_rsv.qgroup_rsv_reserved;
 
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
 		goto out_release;
 	}
+	ret = btrfs_record_root_in_trans(trans, root);
+	if (ret) {
+		btrfs_abort_transaction(trans, ret);
+		goto out_end_trans;
+	}
+	btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved);
+	qgroup_reserved = 0;
 	trans->block_rsv = &block_rsv;
 	trans->bytes_reserved = block_rsv.size;
 
@@ -4611,7 +4619,9 @@ out_end_trans:
 	ret = btrfs_end_transaction(trans);
 	inode->i_flags |= S_DEAD;
 out_release:
-	btrfs_subvolume_release_metadata(root, &block_rsv);
+	btrfs_block_rsv_release(fs_info, &block_rsv, (u64)-1, NULL);
+	if (qgroup_reserved)
+		btrfs_qgroup_free_meta_prealloc(root, qgroup_reserved);
 out_undead:
 	if (ret) {
 		spin_lock(&dest->root_item_lock);
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 294e31edec9d..0493272a7668 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -613,6 +613,7 @@ static noinline int create_subvol(struct mnt_idmap *idmap,
 	int ret;
 	dev_t anon_dev;
 	u64 objectid;
+	u64 qgroup_reserved = 0;
 
 	root_item = kzalloc(sizeof(*root_item), GFP_KERNEL);
 	if (!root_item)
@@ -650,13 +651,18 @@ static noinline int create_subvol(struct mnt_idmap *idmap,
 					       trans_num_items, false);
 	if (ret)
 		goto out_new_inode_args;
+	qgroup_reserved = block_rsv.qgroup_rsv_reserved;
 
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
-		btrfs_subvolume_release_metadata(root, &block_rsv);
-		goto out_new_inode_args;
+		goto out_release_rsv;
 	}
+	ret = btrfs_record_root_in_trans(trans, BTRFS_I(dir)->root);
+	if (ret)
+		goto out;
+	btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved);
+	qgroup_reserved = 0;
 	trans->block_rsv = &block_rsv;
 	trans->bytes_reserved = block_rsv.size;
 	/* Tree log can't currently deal with an inode which is a new root. */
@@ -767,9 +773,11 @@ static noinline int create_subvol(struct mnt_idmap *idmap,
 out:
 	trans->block_rsv = NULL;
 	trans->bytes_reserved = 0;
-	btrfs_subvolume_release_metadata(root, &block_rsv);
-
 	btrfs_end_transaction(trans);
+out_release_rsv:
+	btrfs_block_rsv_release(fs_info, &block_rsv, (u64)-1, NULL);
+	if (qgroup_reserved)
+		btrfs_qgroup_free_meta_prealloc(root, qgroup_reserved);
 out_new_inode_args:
 	btrfs_new_inode_args_destroy(&new_inode_args);
 out_inode:
@@ -791,6 +799,8 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 	struct btrfs_pending_snapshot *pending_snapshot;
 	unsigned int trans_num_items;
 	struct btrfs_trans_handle *trans;
+	struct btrfs_block_rsv *block_rsv;
+	u64 qgroup_reserved = 0;
 	int ret;
 
 	/* We do not support snapshotting right now. */
@@ -827,19 +837,19 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 		goto free_pending;
 	}
 
-	btrfs_init_block_rsv(&pending_snapshot->block_rsv,
-			     BTRFS_BLOCK_RSV_TEMP);
+	block_rsv = &pending_snapshot->block_rsv;
+	btrfs_init_block_rsv(block_rsv, BTRFS_BLOCK_RSV_TEMP);
 	/*
 	 * 1 to add dir item
 	 * 1 to add dir index
 	 * 1 to update parent inode item
 	 */
 	trans_num_items = create_subvol_num_items(inherit) + 3;
-	ret = btrfs_subvolume_reserve_metadata(BTRFS_I(dir)->root,
-					       &pending_snapshot->block_rsv,
+	ret = btrfs_subvolume_reserve_metadata(BTRFS_I(dir)->root, block_rsv,
 					       trans_num_items, false);
 	if (ret)
 		goto free_pending;
+	qgroup_reserved = block_rsv->qgroup_rsv_reserved;
 
 	pending_snapshot->dentry = dentry;
 	pending_snapshot->root = root;
@@ -852,6 +862,13 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 		ret = PTR_ERR(trans);
 		goto fail;
 	}
+	ret = btrfs_record_root_in_trans(trans, BTRFS_I(dir)->root);
+	if (ret) {
+		btrfs_end_transaction(trans);
+		goto fail;
+	}
+	btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved);
+	qgroup_reserved = 0;
 
 	trans->pending_snapshot = pending_snapshot;
 
@@ -881,7 +898,9 @@ fail:
 	if (ret && pending_snapshot->snap)
 		pending_snapshot->snap->anon_dev = 0;
 	btrfs_put_root(pending_snapshot->snap);
-	btrfs_subvolume_release_metadata(root, &pending_snapshot->block_rsv);
+	btrfs_block_rsv_release(fs_info, block_rsv, (u64)-1, NULL);
+	if (qgroup_reserved)
+		btrfs_qgroup_free_meta_prealloc(root, qgroup_reserved);
 free_pending:
 	if (pending_snapshot->anon_dev)
 		free_anon_bdev(pending_snapshot->anon_dev);
@@ -3739,15 +3758,43 @@ static long btrfs_ioctl_quota_ctl(struct file *file, void __user *arg)
 		goto drop_write;
 	}
 
-	down_write(&fs_info->subvol_sem);
-
 	switch (sa->cmd) {
 	case BTRFS_QUOTA_CTL_ENABLE:
 	case BTRFS_QUOTA_CTL_ENABLE_SIMPLE_QUOTA:
+		down_write(&fs_info->subvol_sem);
 		ret = btrfs_quota_enable(fs_info, sa);
+		up_write(&fs_info->subvol_sem);
 		break;
 	case BTRFS_QUOTA_CTL_DISABLE:
+		/*
+		 * Lock the cleaner mutex to prevent races with concurrent
+		 * relocation, because relocation may be building backrefs for
+		 * blocks of the quota root while we are deleting the root. This
+		 * is like dropping fs roots of deleted snapshots/subvolumes, we
+		 * need the same protection.
+		 *
+		 * This also prevents races between concurrent tasks trying to
+		 * disable quotas, because we will unlock and relock
+		 * qgroup_ioctl_lock across BTRFS_FS_QUOTA_ENABLED changes.
+		 *
+		 * We take this here because we have the dependency of
+		 *
+		 * inode_lock -> subvol_sem
+		 *
+		 * because of rename.  With relocation we can prealloc extents,
+		 * so that makes the dependency chain
+		 *
+		 * cleaner_mutex -> inode_lock -> subvol_sem
+		 *
+		 * so we must take the cleaner_mutex here before we take the
+		 * subvol_sem.  The deadlock can't actually happen, but this
+		 * quiets lockdep.
+		 */
+		mutex_lock(&fs_info->cleaner_mutex);
+		down_write(&fs_info->subvol_sem);
 		ret = btrfs_quota_disable(fs_info);
+		up_write(&fs_info->subvol_sem);
+		mutex_unlock(&fs_info->cleaner_mutex);
 		break;
 	default:
 		ret = -EINVAL;
@@ -3755,7 +3802,6 @@ static long btrfs_ioctl_quota_ctl(struct file *file, void __user *arg)
 	}
 
 	kfree(sa);
-	up_write(&fs_info->subvol_sem);
 drop_write:
 	mnt_drop_write_file(file);
 	return ret;
diff --git a/fs/btrfs/messages.c b/fs/btrfs/messages.c
index c96dd66fd0f7..210d9c82e2ae 100644
--- a/fs/btrfs/messages.c
+++ b/fs/btrfs/messages.c
@@ -7,7 +7,7 @@
 
 #ifdef CONFIG_PRINTK
 
-#define STATE_STRING_PREFACE	": state "
+#define STATE_STRING_PREFACE	" state "
 #define STATE_STRING_BUF_LEN	(sizeof(STATE_STRING_PREFACE) + BTRFS_FS_STATE_COUNT + 1)
 
 /*
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index b749ba45da2b..c2a42bcde98e 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -1188,6 +1188,7 @@ struct btrfs_ordered_extent *btrfs_split_ordered_extent(
 	ordered->disk_bytenr += len;
 	ordered->num_bytes -= len;
 	ordered->disk_num_bytes -= len;
+	ordered->ram_bytes -= len;
 
 	if (test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags)) {
 		ASSERT(ordered->bytes_left == 0);
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index 5f90f0605b12..364acc9bbe73 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -1342,16 +1342,10 @@ int btrfs_quota_disable(struct btrfs_fs_info *fs_info)
 	lockdep_assert_held_write(&fs_info->subvol_sem);
 
 	/*
-	 * Lock the cleaner mutex to prevent races with concurrent relocation,
-	 * because relocation may be building backrefs for blocks of the quota
-	 * root while we are deleting the root. This is like dropping fs roots
-	 * of deleted snapshots/subvolumes, we need the same protection.
-	 *
-	 * This also prevents races between concurrent tasks trying to disable
-	 * quotas, because we will unlock and relock qgroup_ioctl_lock across
-	 * BTRFS_FS_QUOTA_ENABLED changes.
+	 * Relocation will mess with backrefs, so make sure we have the
+	 * cleaner_mutex held to protect us from relocate.
 	 */
-	mutex_lock(&fs_info->cleaner_mutex);
+	lockdep_assert_held(&fs_info->cleaner_mutex);
 
 	mutex_lock(&fs_info->qgroup_ioctl_lock);
 	if (!fs_info->quota_root)
@@ -1373,9 +1367,13 @@ int btrfs_quota_disable(struct btrfs_fs_info *fs_info)
 	clear_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags);
 	btrfs_qgroup_wait_for_completion(fs_info, false);
 
+	/*
+	 * We have nothing held here and no trans handle, just return the error
+	 * if there is one.
+	 */
 	ret = flush_reservations(fs_info);
 	if (ret)
-		goto out_unlock_cleaner;
+		return ret;
 
 	/*
 	 * 1 For the root item
@@ -1439,9 +1437,6 @@ out:
 		btrfs_end_transaction(trans);
 	else if (trans)
 		ret = btrfs_commit_transaction(trans);
-out_unlock_cleaner:
-	mutex_unlock(&fs_info->cleaner_mutex);
-
 	return ret;
 }
 
@@ -4495,6 +4490,8 @@ void btrfs_qgroup_convert_reserved_meta(struct btrfs_root *root, int num_bytes)
 				      BTRFS_QGROUP_RSV_META_PREALLOC);
 	trace_qgroup_meta_convert(root, num_bytes);
 	qgroup_convert_meta(fs_info, root->root_key.objectid, num_bytes);
+	if (!sb_rdonly(fs_info->sb))
+		add_root_meta_rsv(root, num_bytes, BTRFS_QGROUP_RSV_META_PERTRANS);
 }
 
 /*
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 4bb538a372ce..7007f9e0c972 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -548,13 +548,3 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 	}
 	return ret;
 }
-
-void btrfs_subvolume_release_metadata(struct btrfs_root *root,
-				      struct btrfs_block_rsv *rsv)
-{
-	struct btrfs_fs_info *fs_info = root->fs_info;
-	u64 qgroup_to_release;
-
-	btrfs_block_rsv_release(fs_info, rsv, (u64)-1, &qgroup_to_release);
-	btrfs_qgroup_convert_reserved_meta(root, qgroup_to_release);
-}
diff --git a/fs/btrfs/root-tree.h b/fs/btrfs/root-tree.h
index 6f929cf3bd49..8f5739e732b9 100644
--- a/fs/btrfs/root-tree.h
+++ b/fs/btrfs/root-tree.h
@@ -18,8 +18,6 @@ struct btrfs_trans_handle;
 int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 				     struct btrfs_block_rsv *rsv,
 				     int nitems, bool use_global_rsv);
-void btrfs_subvolume_release_metadata(struct btrfs_root *root,
-				      struct btrfs_block_rsv *rsv);
 int btrfs_add_root_ref(struct btrfs_trans_handle *trans, u64 root_id,
 		       u64 ref_id, u64 dirid, u64 sequence,
 		       const struct fscrypt_str *name);
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index fa25004ab04e..4b22cfe9a98c 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -1012,6 +1012,7 @@ static void scrub_stripe_read_repair_worker(struct work_struct *work)
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
 	int num_copies = btrfs_num_copies(fs_info, stripe->bg->start,
 					  stripe->bg->length);
+	unsigned long repaired;
 	int mirror;
 	int i;
 
@@ -1078,16 +1079,15 @@ out:
 	 * Submit the repaired sectors.  For zoned case, we cannot do repair
 	 * in-place, but queue the bg to be relocated.
 	 */
-	if (btrfs_is_zoned(fs_info)) {
-		if (!bitmap_empty(&stripe->error_bitmap, stripe->nr_sectors))
+	bitmap_andnot(&repaired, &stripe->init_error_bitmap, &stripe->error_bitmap,
+		      stripe->nr_sectors);
+	if (!sctx->readonly && !bitmap_empty(&repaired, stripe->nr_sectors)) {
+		if (btrfs_is_zoned(fs_info)) {
 			btrfs_repair_one_zone(fs_info, sctx->stripes[0].bg->start);
-	} else if (!sctx->readonly) {
-		unsigned long repaired;
-
-		bitmap_andnot(&repaired, &stripe->init_error_bitmap,
-			      &stripe->error_bitmap, stripe->nr_sectors);
-		scrub_write_sectors(sctx, stripe, repaired, false);
-		wait_scrub_stripe_io(stripe);
+		} else {
+			scrub_write_sectors(sctx, stripe, repaired, false);
+			wait_scrub_stripe_io(stripe);
+		}
 	}
 
 	scrub_stripe_report_errors(sctx, stripe);
diff --git a/fs/btrfs/tests/extent-map-tests.c b/fs/btrfs/tests/extent-map-tests.c
index 253cce7ffecf..47b5d301038e 100644
--- a/fs/btrfs/tests/extent-map-tests.c
+++ b/fs/btrfs/tests/extent-map-tests.c
@@ -847,6 +847,11 @@ static int test_case_7(struct btrfs_fs_info *fs_info)
 		goto out;
 	}
 
+	if (em->block_start != SZ_32K + SZ_4K) {
+		test_err("em->block_start is %llu, expected 36K", em->block_start);
+		goto out;
+	}
+
 	free_extent_map(em);
 
 	read_lock(&em_tree->lock);
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 46e8426adf4f..85f359e0e0a7 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -745,14 +745,6 @@ again:
 		h->reloc_reserved = reloc_reserved;
 	}
 
-	/*
-	 * Now that we have found a transaction to be a part of, convert the
-	 * qgroup reservation from prealloc to pertrans. A different transaction
-	 * can't race in and free our pertrans out from under us.
-	 */
-	if (qgroup_reserved)
-		btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved);
-
 got_it:
 	if (!current->journal_info)
 		current->journal_info = h;
@@ -786,8 +778,15 @@ got_it:
 		 * not just freed.
 		 */
 		btrfs_end_transaction(h);
-		return ERR_PTR(ret);
+		goto reserve_fail;
 	}
+	/*
+	 * Now that we have found a transaction to be a part of, convert the
+	 * qgroup reservation from prealloc to pertrans. A different transaction
+	 * can't race in and free our pertrans out from under us.
+	 */
+	if (qgroup_reserved)
+		btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved);
 
 	return h;
 
@@ -1495,6 +1494,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans)
 			radix_tree_tag_clear(&fs_info->fs_roots_radix,
 					(unsigned long)root->root_key.objectid,
 					BTRFS_ROOT_TRANS_TAG);
+			btrfs_qgroup_free_meta_all_pertrans(root);
 			spin_unlock(&fs_info->fs_roots_radix_lock);
 
 			btrfs_free_log(trans, root);
@@ -1519,7 +1519,6 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans)
 			if (ret2)
 				return ret2;
 			spin_lock(&fs_info->fs_roots_radix_lock);
-			btrfs_qgroup_free_meta_all_pertrans(root);
 		}
 	}
 	spin_unlock(&fs_info->fs_roots_radix_lock);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index f15591f3e54f..ef6bd2f4251b 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -3455,6 +3455,7 @@ again:
 			 * alignment and size).
 			 */
 			ret = -EUCLEAN;
+			mutex_unlock(&fs_info->reclaim_bgs_lock);
 			goto error;
 		}
 
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 1340d77124ae..ee9caf7916fb 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -795,8 +795,10 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc)
 	ihold(inode);
 
 	if (wbc->sync_mode == WB_SYNC_NONE &&
-	    ceph_inode_to_fs_client(inode)->write_congested)
+	    ceph_inode_to_fs_client(inode)->write_congested) {
+		redirty_page_for_writepage(wbc, page);
 		return AOP_WRITEPAGE_ACTIVATE;
+	}
 
 	wait_on_page_fscache(page);
 
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 55051ad09c19..c4941ba245ac 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -4783,13 +4783,13 @@ int ceph_drop_caps_for_unlink(struct inode *inode)
 
 			doutc(mdsc->fsc->client, "%p %llx.%llx\n", inode,
 			      ceph_vinop(inode));
-			spin_lock(&mdsc->cap_unlink_delay_lock);
+			spin_lock(&mdsc->cap_delay_lock);
 			ci->i_ceph_flags |= CEPH_I_FLUSH;
 			if (!list_empty(&ci->i_cap_delay_list))
 				list_del_init(&ci->i_cap_delay_list);
 			list_add_tail(&ci->i_cap_delay_list,
 				      &mdsc->cap_unlink_delay_list);
-			spin_unlock(&mdsc->cap_unlink_delay_lock);
+			spin_unlock(&mdsc->cap_delay_lock);
 
 			/*
 			 * Fire the work immediately, because the MDS maybe
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 3ab9c268a8bb..360b686c3c67 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -2504,7 +2504,7 @@ static void ceph_cap_unlink_work(struct work_struct *work)
 	struct ceph_client *cl = mdsc->fsc->client;
 
 	doutc(cl, "begin\n");
-	spin_lock(&mdsc->cap_unlink_delay_lock);
+	spin_lock(&mdsc->cap_delay_lock);
 	while (!list_empty(&mdsc->cap_unlink_delay_list)) {
 		struct ceph_inode_info *ci;
 		struct inode *inode;
@@ -2516,15 +2516,15 @@ static void ceph_cap_unlink_work(struct work_struct *work)
 
 		inode = igrab(&ci->netfs.inode);
 		if (inode) {
-			spin_unlock(&mdsc->cap_unlink_delay_lock);
+			spin_unlock(&mdsc->cap_delay_lock);
 			doutc(cl, "on %p %llx.%llx\n", inode,
 			      ceph_vinop(inode));
 			ceph_check_caps(ci, CHECK_CAPS_FLUSH);
 			iput(inode);
-			spin_lock(&mdsc->cap_unlink_delay_lock);
+			spin_lock(&mdsc->cap_delay_lock);
 		}
 	}
-	spin_unlock(&mdsc->cap_unlink_delay_lock);
+	spin_unlock(&mdsc->cap_delay_lock);
 	doutc(cl, "done\n");
 }
 
@@ -5404,7 +5404,6 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc)
 	INIT_LIST_HEAD(&mdsc->cap_wait_list);
 	spin_lock_init(&mdsc->cap_delay_lock);
 	INIT_LIST_HEAD(&mdsc->cap_unlink_delay_list);
-	spin_lock_init(&mdsc->cap_unlink_delay_lock);
 	INIT_LIST_HEAD(&mdsc->snap_flush_list);
 	spin_lock_init(&mdsc->snap_flush_lock);
 	mdsc->last_cap_flush_tid = 1;
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index 03f8ff00874f..b88e80415224 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -461,9 +461,8 @@ struct ceph_mds_client {
 	struct delayed_work    delayed_work;  /* delayed work */
 	unsigned long    last_renew_caps;  /* last time we renewed our caps */
 	struct list_head cap_delay_list;   /* caps with delayed release */
-	spinlock_t       cap_delay_lock;   /* protects cap_delay_list */
 	struct list_head cap_unlink_delay_list;  /* caps with delayed release for unlink */
-	spinlock_t       cap_unlink_delay_lock;  /* protects cap_unlink_delay_list */
+	spinlock_t       cap_delay_lock;   /* protects cap_delay_list and cap_unlink_delay_list */
 	struct list_head snap_flush_list;  /* cap_snaps ready to flush */
 	spinlock_t       snap_flush_lock;
 
diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
index 8aff1a724805..62da538d91cb 100644
--- a/fs/erofs/fscache.c
+++ b/fs/erofs/fscache.c
@@ -151,7 +151,7 @@ static int erofs_fscache_read_io_async(struct fscache_cookie *cookie,
 		if (WARN_ON(len == 0))
 			source = NETFS_INVALID_READ;
 		if (source != NETFS_READ_FROM_CACHE) {
-			erofs_err(NULL, "prepare_read failed (source %d)", source);
+			erofs_err(NULL, "prepare_ondemand_read failed (source %d)", source);
 			return -EIO;
 		}
 
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 39c67119f43b..d28ccfc0352b 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -84,13 +84,6 @@ struct erofs_dev_context {
 	bool flatdev;
 };
 
-struct erofs_fs_context {
-	struct erofs_mount_opts opt;
-	struct erofs_dev_context *devs;
-	char *fsid;
-	char *domain_id;
-};
-
 /* all filesystem-wide lz4 configurations */
 struct erofs_sb_lz4_info {
 	/* # of pages needed for EROFS lz4 rolling decompression */
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index c0eb139adb07..30b49b2eee53 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -370,18 +370,18 @@ out:
 	return ret;
 }
 
-static void erofs_default_options(struct erofs_fs_context *ctx)
+static void erofs_default_options(struct erofs_sb_info *sbi)
 {
 #ifdef CONFIG_EROFS_FS_ZIP
-	ctx->opt.cache_strategy = EROFS_ZIP_CACHE_READAROUND;
-	ctx->opt.max_sync_decompress_pages = 3;
-	ctx->opt.sync_decompress = EROFS_SYNC_DECOMPRESS_AUTO;
+	sbi->opt.cache_strategy = EROFS_ZIP_CACHE_READAROUND;
+	sbi->opt.max_sync_decompress_pages = 3;
+	sbi->opt.sync_decompress = EROFS_SYNC_DECOMPRESS_AUTO;
 #endif
 #ifdef CONFIG_EROFS_FS_XATTR
-	set_opt(&ctx->opt, XATTR_USER);
+	set_opt(&sbi->opt, XATTR_USER);
 #endif
 #ifdef CONFIG_EROFS_FS_POSIX_ACL
-	set_opt(&ctx->opt, POSIX_ACL);
+	set_opt(&sbi->opt, POSIX_ACL);
 #endif
 }
 
@@ -426,16 +426,16 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {
 static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
 {
 #ifdef CONFIG_FS_DAX
-	struct erofs_fs_context *ctx = fc->fs_private;
+	struct erofs_sb_info *sbi = fc->s_fs_info;
 
 	switch (mode) {
 	case EROFS_MOUNT_DAX_ALWAYS:
-		set_opt(&ctx->opt, DAX_ALWAYS);
-		clear_opt(&ctx->opt, DAX_NEVER);
+		set_opt(&sbi->opt, DAX_ALWAYS);
+		clear_opt(&sbi->opt, DAX_NEVER);
 		return true;
 	case EROFS_MOUNT_DAX_NEVER:
-		set_opt(&ctx->opt, DAX_NEVER);
-		clear_opt(&ctx->opt, DAX_ALWAYS);
+		set_opt(&sbi->opt, DAX_NEVER);
+		clear_opt(&sbi->opt, DAX_ALWAYS);
 		return true;
 	default:
 		DBG_BUGON(1);
@@ -450,7 +450,7 @@ static bool erofs_fc_set_dax_mode(struct fs_context *fc, unsigned int mode)
 static int erofs_fc_parse_param(struct fs_context *fc,
 				struct fs_parameter *param)
 {
-	struct erofs_fs_context *ctx = fc->fs_private;
+	struct erofs_sb_info *sbi = fc->s_fs_info;
 	struct fs_parse_result result;
 	struct erofs_device_info *dif;
 	int opt, ret;
@@ -463,9 +463,9 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 	case Opt_user_xattr:
 #ifdef CONFIG_EROFS_FS_XATTR
 		if (result.boolean)
-			set_opt(&ctx->opt, XATTR_USER);
+			set_opt(&sbi->opt, XATTR_USER);
 		else
-			clear_opt(&ctx->opt, XATTR_USER);
+			clear_opt(&sbi->opt, XATTR_USER);
 #else
 		errorfc(fc, "{,no}user_xattr options not supported");
 #endif
@@ -473,16 +473,16 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 	case Opt_acl:
 #ifdef CONFIG_EROFS_FS_POSIX_ACL
 		if (result.boolean)
-			set_opt(&ctx->opt, POSIX_ACL);
+			set_opt(&sbi->opt, POSIX_ACL);
 		else
-			clear_opt(&ctx->opt, POSIX_ACL);
+			clear_opt(&sbi->opt, POSIX_ACL);
 #else
 		errorfc(fc, "{,no}acl options not supported");
 #endif
 		break;
 	case Opt_cache_strategy:
 #ifdef CONFIG_EROFS_FS_ZIP
-		ctx->opt.cache_strategy = result.uint_32;
+		sbi->opt.cache_strategy = result.uint_32;
 #else
 		errorfc(fc, "compression not supported, cache_strategy ignored");
 #endif
@@ -504,27 +504,27 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 			kfree(dif);
 			return -ENOMEM;
 		}
-		down_write(&ctx->devs->rwsem);
-		ret = idr_alloc(&ctx->devs->tree, dif, 0, 0, GFP_KERNEL);
-		up_write(&ctx->devs->rwsem);
+		down_write(&sbi->devs->rwsem);
+		ret = idr_alloc(&sbi->devs->tree, dif, 0, 0, GFP_KERNEL);
+		up_write(&sbi->devs->rwsem);
 		if (ret < 0) {
 			kfree(dif->path);
 			kfree(dif);
 			return ret;
 		}
-		++ctx->devs->extra_devices;
+		++sbi->devs->extra_devices;
 		break;
 #ifdef CONFIG_EROFS_FS_ONDEMAND
 	case Opt_fsid:
-		kfree(ctx->fsid);
-		ctx->fsid = kstrdup(param->string, GFP_KERNEL);
-		if (!ctx->fsid)
+		kfree(sbi->fsid);
+		sbi->fsid = kstrdup(param->string, GFP_KERNEL);
+		if (!sbi->fsid)
 			return -ENOMEM;
 		break;
 	case Opt_domain_id:
-		kfree(ctx->domain_id);
-		ctx->domain_id = kstrdup(param->string, GFP_KERNEL);
-		if (!ctx->domain_id)
+		kfree(sbi->domain_id);
+		sbi->domain_id = kstrdup(param->string, GFP_KERNEL);
+		if (!sbi->domain_id)
 			return -ENOMEM;
 		break;
 #else
@@ -581,8 +581,7 @@ static const struct export_operations erofs_export_ops = {
 static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 {
 	struct inode *inode;
-	struct erofs_sb_info *sbi;
-	struct erofs_fs_context *ctx = fc->fs_private;
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
 	int err;
 
 	sb->s_magic = EROFS_SUPER_MAGIC;
@@ -590,19 +589,6 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 	sb->s_maxbytes = MAX_LFS_FILESIZE;
 	sb->s_op = &erofs_sops;
 
-	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
-	if (!sbi)
-		return -ENOMEM;
-
-	sb->s_fs_info = sbi;
-	sbi->opt = ctx->opt;
-	sbi->devs = ctx->devs;
-	ctx->devs = NULL;
-	sbi->fsid = ctx->fsid;
-	ctx->fsid = NULL;
-	sbi->domain_id = ctx->domain_id;
-	ctx->domain_id = NULL;
-
 	sbi->blkszbits = PAGE_SHIFT;
 	if (erofs_is_fscache_mode(sb)) {
 		sb->s_blocksize = PAGE_SIZE;
@@ -706,9 +692,9 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)
 
 static int erofs_fc_get_tree(struct fs_context *fc)
 {
-	struct erofs_fs_context *ctx = fc->fs_private;
+	struct erofs_sb_info *sbi = fc->s_fs_info;
 
-	if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && ctx->fsid)
+	if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && sbi->fsid)
 		return get_tree_nodev(fc, erofs_fc_fill_super);
 
 	return get_tree_bdev(fc, erofs_fc_fill_super);
@@ -718,19 +704,19 @@ static int erofs_fc_reconfigure(struct fs_context *fc)
 {
 	struct super_block *sb = fc->root->d_sb;
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
-	struct erofs_fs_context *ctx = fc->fs_private;
+	struct erofs_sb_info *new_sbi = fc->s_fs_info;
 
 	DBG_BUGON(!sb_rdonly(sb));
 
-	if (ctx->fsid || ctx->domain_id)
+	if (new_sbi->fsid || new_sbi->domain_id)
 		erofs_info(sb, "ignoring reconfiguration for fsid|domain_id.");
 
-	if (test_opt(&ctx->opt, POSIX_ACL))
+	if (test_opt(&new_sbi->opt, POSIX_ACL))
 		fc->sb_flags |= SB_POSIXACL;
 	else
 		fc->sb_flags &= ~SB_POSIXACL;
 
-	sbi->opt = ctx->opt;
+	sbi->opt = new_sbi->opt;
 
 	fc->sb_flags |= SB_RDONLY;
 	return 0;
@@ -761,12 +747,15 @@ static void erofs_free_dev_context(struct erofs_dev_context *devs)
 
 static void erofs_fc_free(struct fs_context *fc)
 {
-	struct erofs_fs_context *ctx = fc->fs_private;
+	struct erofs_sb_info *sbi = fc->s_fs_info;
 
-	erofs_free_dev_context(ctx->devs);
-	kfree(ctx->fsid);
-	kfree(ctx->domain_id);
-	kfree(ctx);
+	if (!sbi)
+		return;
+
+	erofs_free_dev_context(sbi->devs);
+	kfree(sbi->fsid);
+	kfree(sbi->domain_id);
+	kfree(sbi);
 }
 
 static const struct fs_context_operations erofs_context_ops = {
@@ -778,38 +767,35 @@ static const struct fs_context_operations erofs_context_ops = {
 
 static int erofs_init_fs_context(struct fs_context *fc)
 {
-	struct erofs_fs_context *ctx;
+	struct erofs_sb_info *sbi;
 
-	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
-	if (!ctx)
+	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
+	if (!sbi)
 		return -ENOMEM;
-	ctx->devs = kzalloc(sizeof(struct erofs_dev_context), GFP_KERNEL);
-	if (!ctx->devs) {
-		kfree(ctx);
+
+	sbi->devs = kzalloc(sizeof(struct erofs_dev_context), GFP_KERNEL);
+	if (!sbi->devs) {
+		kfree(sbi);
 		return -ENOMEM;
 	}
-	fc->fs_private = ctx;
+	fc->s_fs_info = sbi;
 
-	idr_init(&ctx->devs->tree);
-	init_rwsem(&ctx->devs->rwsem);
-	erofs_default_options(ctx);
+	idr_init(&sbi->devs->tree);
+	init_rwsem(&sbi->devs->rwsem);
+	erofs_default_options(sbi);
 	fc->ops = &erofs_context_ops;
 	return 0;
 }
 
 static void erofs_kill_sb(struct super_block *sb)
 {
-	struct erofs_sb_info *sbi;
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
 
-	if (erofs_is_fscache_mode(sb))
+	if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && sbi->fsid)
 		kill_anon_super(sb);
 	else
 		kill_block_super(sb);
 
-	sbi = EROFS_SB(sb);
-	if (!sbi)
-		return;
-
 	erofs_free_dev_context(sbi->devs);
 	fs_put_dax(sbi->dax_dev, NULL);
 	erofs_fscache_unregister_fs(sb);
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 882b89edc52a..f53ca4f7fced 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -980,6 +980,34 @@ static __poll_t __ep_eventpoll_poll(struct file *file, poll_table *wait, int dep
 }
 
 /*
+ * The ffd.file pointer may be in the process of being torn down due to
+ * being closed, but we may not have finished eventpoll_release() yet.
+ *
+ * Normally, even with the atomic_long_inc_not_zero, the file may have
+ * been free'd and then gotten re-allocated to something else (since
+ * files are not RCU-delayed, they are SLAB_TYPESAFE_BY_RCU).
+ *
+ * But for epoll, users hold the ep->mtx mutex, and as such any file in
+ * the process of being free'd will block in eventpoll_release_file()
+ * and thus the underlying file allocation will not be free'd, and the
+ * file re-use cannot happen.
+ *
+ * For the same reason we can avoid a rcu_read_lock() around the
+ * operation - 'ffd.file' cannot go away even if the refcount has
+ * reached zero (but we must still not call out to ->poll() functions
+ * etc).
+ */
+static struct file *epi_fget(const struct epitem *epi)
+{
+	struct file *file;
+
+	file = epi->ffd.file;
+	if (!atomic_long_inc_not_zero(&file->f_count))
+		file = NULL;
+	return file;
+}
+
+/*
  * Differs from ep_eventpoll_poll() in that internal callers already have
  * the ep->mtx so we need to start from depth=1, such that mutex_lock_nested()
  * is correctly annotated.
@@ -987,14 +1015,22 @@ static __poll_t __ep_eventpoll_poll(struct file *file, poll_table *wait, int dep
 static __poll_t ep_item_poll(const struct epitem *epi, poll_table *pt,
 				 int depth)
 {
-	struct file *file = epi->ffd.file;
+	struct file *file = epi_fget(epi);
 	__poll_t res;
 
+	/*
+	 * We could return EPOLLERR | EPOLLHUP or something, but let's
+	 * treat this more as "file doesn't exist, poll didn't happen".
+	 */
+	if (!file)
+		return 0;
+
 	pt->_key = epi->event.events;
 	if (!is_file_epoll(file))
 		res = vfs_poll(file, pt);
 	else
 		res = __ep_eventpoll_poll(file, pt, depth);
+	fput(file);
 	return res & epi->event.events;
 }
 
diff --git a/fs/fuse/cuse.c b/fs/fuse/cuse.c
index b6cad106c37e..0b2da7b7e2ad 100644
--- a/fs/fuse/cuse.c
+++ b/fs/fuse/cuse.c
@@ -310,6 +310,10 @@ struct cuse_init_args {
 /**
  * cuse_process_init_reply - finish initializing CUSE channel
  *
+ * @fm: The fuse mount information containing the CUSE connection.
+ * @args: The arguments passed to the init reply.
+ * @error: The error code signifying if any error occurred during the process.
+ *
  * This function creates the character device and sets up all the
  * required data structures for it.  Please read the comment at the
  * top of this file for high level overview.
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 4a6df591add6..2b0d4781f394 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -1321,6 +1321,7 @@ retry:
 			err = fuse_do_statx(inode, file, stat);
 			if (err == -ENOSYS) {
 				fc->no_statx = 1;
+				err = 0;
 				goto retry;
 			}
 		} else {
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index a56e7bffd000..b57ce4157640 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -1362,7 +1362,7 @@ static void fuse_dio_lock(struct kiocb *iocb, struct iov_iter *from,
 			  bool *exclusive)
 {
 	struct inode *inode = file_inode(iocb->ki_filp);
-	struct fuse_file *ff = iocb->ki_filp->private_data;
+	struct fuse_inode *fi = get_fuse_inode(inode);
 
 	*exclusive = fuse_dio_wr_exclusive_lock(iocb, from);
 	if (*exclusive) {
@@ -1377,7 +1377,7 @@ static void fuse_dio_lock(struct kiocb *iocb, struct iov_iter *from,
 		 * have raced, so check it again.
 		 */
 		if (fuse_io_past_eof(iocb, from) ||
-		    fuse_file_uncached_io_start(inode, ff, NULL) != 0) {
+		    fuse_inode_uncached_io_start(fi, NULL) != 0) {
 			inode_unlock_shared(inode);
 			inode_lock(inode);
 			*exclusive = true;
@@ -1388,13 +1388,13 @@ static void fuse_dio_lock(struct kiocb *iocb, struct iov_iter *from,
 static void fuse_dio_unlock(struct kiocb *iocb, bool exclusive)
 {
 	struct inode *inode = file_inode(iocb->ki_filp);
-	struct fuse_file *ff = iocb->ki_filp->private_data;
+	struct fuse_inode *fi = get_fuse_inode(inode);
 
 	if (exclusive) {
 		inode_unlock(inode);
 	} else {
 		/* Allow opens in caching mode after last parallel dio end */
-		fuse_file_uncached_io_end(inode, ff);
+		fuse_inode_uncached_io_end(fi);
 		inode_unlock_shared(inode);
 	}
 }
@@ -2574,8 +2574,10 @@ static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma)
 		 * First mmap of direct_io file enters caching inode io mode.
 		 * Also waits for parallel dio writers to go into serial mode
 		 * (exclusive instead of shared lock).
+		 * After first mmap, the inode stays in caching io mode until
+		 * the direct_io file release.
 		 */
-		rc = fuse_file_cached_io_start(inode, ff);
+		rc = fuse_file_cached_io_open(inode, ff);
 		if (rc)
 			return rc;
 	}
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index b24084b60864..f23919610313 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -1394,9 +1394,10 @@ int fuse_fileattr_set(struct mnt_idmap *idmap,
 		      struct dentry *dentry, struct fileattr *fa);
 
 /* iomode.c */
-int fuse_file_cached_io_start(struct inode *inode, struct fuse_file *ff);
-int fuse_file_uncached_io_start(struct inode *inode, struct fuse_file *ff, struct fuse_backing *fb);
-void fuse_file_uncached_io_end(struct inode *inode, struct fuse_file *ff);
+int fuse_file_cached_io_open(struct inode *inode, struct fuse_file *ff);
+int fuse_inode_uncached_io_start(struct fuse_inode *fi,
+				 struct fuse_backing *fb);
+void fuse_inode_uncached_io_end(struct fuse_inode *fi);
 
 int fuse_file_io_open(struct file *file, struct inode *inode);
 void fuse_file_io_release(struct fuse_file *ff, struct inode *inode);
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 3a5d88878335..99e44ea7d875 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -175,6 +175,7 @@ static void fuse_evict_inode(struct inode *inode)
 		}
 	}
 	if (S_ISREG(inode->i_mode) && !fuse_is_bad(inode)) {
+		WARN_ON(fi->iocachectr != 0);
 		WARN_ON(!list_empty(&fi->write_files));
 		WARN_ON(!list_empty(&fi->queued_writes));
 	}
diff --git a/fs/fuse/iomode.c b/fs/fuse/iomode.c
index c653ddcf0578..c99e285f3183 100644
--- a/fs/fuse/iomode.c
+++ b/fs/fuse/iomode.c
@@ -21,12 +21,13 @@ static inline bool fuse_is_io_cache_wait(struct fuse_inode *fi)
 }
 
 /*
- * Start cached io mode.
+ * Called on cached file open() and on first mmap() of direct_io file.
+ * Takes cached_io inode mode reference to be dropped on file release.
  *
  * Blocks new parallel dio writes and waits for the in-progress parallel dio
  * writes to complete.
  */
-int fuse_file_cached_io_start(struct inode *inode, struct fuse_file *ff)
+int fuse_file_cached_io_open(struct inode *inode, struct fuse_file *ff)
 {
 	struct fuse_inode *fi = get_fuse_inode(inode);
 
@@ -67,10 +68,9 @@ int fuse_file_cached_io_start(struct inode *inode, struct fuse_file *ff)
 	return 0;
 }
 
-static void fuse_file_cached_io_end(struct inode *inode, struct fuse_file *ff)
+static void fuse_file_cached_io_release(struct fuse_file *ff,
+					struct fuse_inode *fi)
 {
-	struct fuse_inode *fi = get_fuse_inode(inode);
-
 	spin_lock(&fi->lock);
 	WARN_ON(fi->iocachectr <= 0);
 	WARN_ON(ff->iomode != IOM_CACHED);
@@ -82,16 +82,15 @@ static void fuse_file_cached_io_end(struct inode *inode, struct fuse_file *ff)
 }
 
 /* Start strictly uncached io mode where cache access is not allowed */
-int fuse_file_uncached_io_start(struct inode *inode, struct fuse_file *ff, struct fuse_backing *fb)
+int fuse_inode_uncached_io_start(struct fuse_inode *fi, struct fuse_backing *fb)
 {
-	struct fuse_inode *fi = get_fuse_inode(inode);
 	struct fuse_backing *oldfb;
 	int err = 0;
 
 	spin_lock(&fi->lock);
 	/* deny conflicting backing files on same fuse inode */
 	oldfb = fuse_inode_backing(fi);
-	if (oldfb && oldfb != fb) {
+	if (fb && oldfb && oldfb != fb) {
 		err = -EBUSY;
 		goto unlock;
 	}
@@ -99,12 +98,10 @@ int fuse_file_uncached_io_start(struct inode *inode, struct fuse_file *ff, struc
 		err = -ETXTBSY;
 		goto unlock;
 	}
-	WARN_ON(ff->iomode != IOM_NONE);
 	fi->iocachectr--;
-	ff->iomode = IOM_UNCACHED;
 
 	/* fuse inode holds a single refcount of backing file */
-	if (!oldfb) {
+	if (fb && !oldfb) {
 		oldfb = fuse_inode_backing_set(fi, fb);
 		WARN_ON_ONCE(oldfb != NULL);
 	} else {
@@ -115,15 +112,29 @@ unlock:
 	return err;
 }
 
-void fuse_file_uncached_io_end(struct inode *inode, struct fuse_file *ff)
+/* Takes uncached_io inode mode reference to be dropped on file release */
+static int fuse_file_uncached_io_open(struct inode *inode,
+				      struct fuse_file *ff,
+				      struct fuse_backing *fb)
 {
 	struct fuse_inode *fi = get_fuse_inode(inode);
+	int err;
+
+	err = fuse_inode_uncached_io_start(fi, fb);
+	if (err)
+		return err;
+
+	WARN_ON(ff->iomode != IOM_NONE);
+	ff->iomode = IOM_UNCACHED;
+	return 0;
+}
+
+void fuse_inode_uncached_io_end(struct fuse_inode *fi)
+{
 	struct fuse_backing *oldfb = NULL;
 
 	spin_lock(&fi->lock);
 	WARN_ON(fi->iocachectr >= 0);
-	WARN_ON(ff->iomode != IOM_UNCACHED);
-	ff->iomode = IOM_NONE;
 	fi->iocachectr++;
 	if (!fi->iocachectr) {
 		wake_up(&fi->direct_io_waitq);
@@ -134,6 +145,15 @@ void fuse_file_uncached_io_end(struct inode *inode, struct fuse_file *ff)
 		fuse_backing_put(oldfb);
 }
 
+/* Drop uncached_io reference from passthrough open */
+static void fuse_file_uncached_io_release(struct fuse_file *ff,
+					  struct fuse_inode *fi)
+{
+	WARN_ON(ff->iomode != IOM_UNCACHED);
+	ff->iomode = IOM_NONE;
+	fuse_inode_uncached_io_end(fi);
+}
+
 /*
  * Open flags that are allowed in combination with FOPEN_PASSTHROUGH.
  * A combination of FOPEN_PASSTHROUGH and FOPEN_DIRECT_IO means that read/write
@@ -163,7 +183,7 @@ static int fuse_file_passthrough_open(struct inode *inode, struct file *file)
 		return PTR_ERR(fb);
 
 	/* First passthrough file open denies caching inode io mode */
-	err = fuse_file_uncached_io_start(inode, ff, fb);
+	err = fuse_file_uncached_io_open(inode, ff, fb);
 	if (!err)
 		return 0;
 
@@ -216,7 +236,7 @@ int fuse_file_io_open(struct file *file, struct inode *inode)
 	if (ff->open_flags & FOPEN_PASSTHROUGH)
 		err = fuse_file_passthrough_open(inode, file);
 	else
-		err = fuse_file_cached_io_start(inode, ff);
+		err = fuse_file_cached_io_open(inode, ff);
 	if (err)
 		goto fail;
 
@@ -236,8 +256,10 @@ fail:
 /* No more pending io and no new io possible to inode via open/mmapped file */
 void fuse_file_io_release(struct fuse_file *ff, struct inode *inode)
 {
+	struct fuse_inode *fi = get_fuse_inode(inode);
+
 	/*
-	 * Last parallel dio close allows caching inode io mode.
+	 * Last passthrough file close allows caching inode io mode.
 	 * Last caching file close exits caching inode io mode.
 	 */
 	switch (ff->iomode) {
@@ -245,10 +267,10 @@ void fuse_file_io_release(struct fuse_file *ff, struct inode *inode)
 		/* Nothing to do */
 		break;
 	case IOM_UNCACHED:
-		fuse_file_uncached_io_end(inode, ff);
+		fuse_file_uncached_io_release(ff, fi);
 		break;
 	case IOM_CACHED:
-		fuse_file_cached_io_end(inode, ff);
+		fuse_file_cached_io_release(ff, fi);
 		break;
 	}
 }
diff --git a/fs/ioctl.c b/fs/ioctl.c
index 1d5abfdf0f22..fb0628e680c4 100644
--- a/fs/ioctl.c
+++ b/fs/ioctl.c
@@ -769,7 +769,7 @@ static int ioctl_getfsuuid(struct file *file, void __user *argp)
 	struct fsuuid2 u = { .len = sb->s_uuid_len, };
 
 	if (!sb->s_uuid_len)
-		return -ENOIOCTLCMD;
+		return -ENOTTY;
 
 	memcpy(&u.uuid[0], &sb->s_uuid, sb->s_uuid_len);
 
@@ -781,7 +781,7 @@ static int ioctl_get_fs_sysfs_path(struct file *file, void __user *argp)
 	struct super_block *sb = file_inode(file)->i_sb;
 
 	if (!strlen(sb->s_sysfs_name))
-		return -ENOIOCTLCMD;
+		return -ENOTTY;
 
 	struct fs_sysfs_path u = {};
 
diff --git a/fs/kernfs/file.c b/fs/kernfs/file.c
index e9df2f87072c..8502ef68459b 100644
--- a/fs/kernfs/file.c
+++ b/fs/kernfs/file.c
@@ -636,11 +636,18 @@ static int kernfs_fop_open(struct inode *inode, struct file *file)
 	 * each file a separate locking class.  Let's differentiate on
 	 * whether the file has mmap or not for now.
 	 *
-	 * Both paths of the branch look the same.  They're supposed to
+	 * For similar reasons, writable and readonly files are given different
+	 * lockdep key, because the writable file /sys/power/resume may call vfs
+	 * lookup helpers for arbitrary paths and readonly files can be read by
+	 * overlayfs from vfs helpers when sysfs is a lower layer of overalyfs.
+	 *
+	 * All three cases look the same.  They're supposed to
 	 * look that way and give @of->mutex different static lockdep keys.
 	 */
 	if (has_mmap)
 		mutex_init(&of->mutex);
+	else if (file->f_mode & FMODE_WRITE)
+		mutex_init(&of->mutex);
 	else
 		mutex_init(&of->mutex);
 
diff --git a/fs/netfs/buffered_write.c b/fs/netfs/buffered_write.c
index 9a0d32e4b422..267b622d923b 100644
--- a/fs/netfs/buffered_write.c
+++ b/fs/netfs/buffered_write.c
@@ -164,7 +164,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 	enum netfs_how_to_modify howto;
 	enum netfs_folio_trace trace;
 	unsigned int bdp_flags = (iocb->ki_flags & IOCB_SYNC) ? 0: BDP_ASYNC;
-	ssize_t written = 0, ret;
+	ssize_t written = 0, ret, ret2;
 	loff_t i_size, pos = iocb->ki_pos, from, to;
 	size_t max_chunk = PAGE_SIZE << MAX_PAGECACHE_ORDER;
 	bool maybe_trouble = false;
@@ -172,15 +172,14 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 	if (unlikely(test_bit(NETFS_ICTX_WRITETHROUGH, &ctx->flags) ||
 		     iocb->ki_flags & (IOCB_DSYNC | IOCB_SYNC))
 	    ) {
-		if (pos < i_size_read(inode)) {
-			ret = filemap_write_and_wait_range(mapping, pos, pos + iter->count);
-			if (ret < 0) {
-				goto out;
-			}
-		}
-
 		wbc_attach_fdatawrite_inode(&wbc, mapping->host);
 
+		ret = filemap_write_and_wait_range(mapping, pos, pos + iter->count);
+		if (ret < 0) {
+			wbc_detach_inode(&wbc);
+			goto out;
+		}
+
 		wreq = netfs_begin_writethrough(iocb, iter->count);
 		if (IS_ERR(wreq)) {
 			wbc_detach_inode(&wbc);
@@ -395,10 +394,12 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 
 out:
 	if (unlikely(wreq)) {
-		ret = netfs_end_writethrough(wreq, iocb);
+		ret2 = netfs_end_writethrough(wreq, iocb);
 		wbc_detach_inode(&wbc);
-		if (ret == -EIOCBQUEUED)
-			return ret;
+		if (ret2 == -EIOCBQUEUED)
+			return ret2;
+		if (ret == 0)
+			ret = ret2;
 	}
 
 	iocb->ki_pos += written;
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index c709c296ea9a..acef52ecb1bb 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -2429,7 +2429,12 @@ static int nfs_net_init(struct net *net)
 	struct nfs_net *nn = net_generic(net, nfs_net_id);
 
 	nfs_clients_init(net);
-	rpc_proc_register(net, &nn->rpcstats);
+
+	if (!rpc_proc_register(net, &nn->rpcstats)) {
+		nfs_clients_exit(net);
+		return -ENOMEM;
+	}
+
 	return nfs_fs_proc_net_init(net);
 }
 
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 87c9547989f6..e88aca0c6e8e 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -983,15 +983,7 @@ static struct workqueue_struct *callback_wq;
 static bool nfsd4_queue_cb(struct nfsd4_callback *cb)
 {
 	trace_nfsd_cb_queue(cb->cb_clp, cb);
-	return queue_delayed_work(callback_wq, &cb->cb_work, 0);
-}
-
-static void nfsd4_queue_cb_delayed(struct nfsd4_callback *cb,
-				   unsigned long msecs)
-{
-	trace_nfsd_cb_queue(cb->cb_clp, cb);
-	queue_delayed_work(callback_wq, &cb->cb_work,
-			   msecs_to_jiffies(msecs));
+	return queue_work(callback_wq, &cb->cb_work);
 }
 
 static void nfsd41_cb_inflight_begin(struct nfs4_client *clp)
@@ -1490,7 +1482,7 @@ static void
 nfsd4_run_cb_work(struct work_struct *work)
 {
 	struct nfsd4_callback *cb =
-		container_of(work, struct nfsd4_callback, cb_work.work);
+		container_of(work, struct nfsd4_callback, cb_work);
 	struct nfs4_client *clp = cb->cb_clp;
 	struct rpc_clnt *clnt;
 	int flags;
@@ -1502,16 +1494,8 @@ nfsd4_run_cb_work(struct work_struct *work)
 
 	clnt = clp->cl_cb_client;
 	if (!clnt) {
-		if (test_bit(NFSD4_CLIENT_CB_KILL, &clp->cl_flags))
-			nfsd41_destroy_cb(cb);
-		else {
-			/*
-			 * XXX: Ideally, we could wait for the client to
-			 *	reconnect, but I haven't figured out how
-			 *	to do that yet.
-			 */
-			nfsd4_queue_cb_delayed(cb, 25);
-		}
+		/* Callback channel broken, or client killed; give up: */
+		nfsd41_destroy_cb(cb);
 		return;
 	}
 
@@ -1544,7 +1528,7 @@ void nfsd4_init_cb(struct nfsd4_callback *cb, struct nfs4_client *clp,
 	cb->cb_msg.rpc_argp = cb;
 	cb->cb_msg.rpc_resp = cb;
 	cb->cb_ops = ops;
-	INIT_DELAYED_WORK(&cb->cb_work, nfsd4_run_cb_work);
+	INIT_WORK(&cb->cb_work, nfsd4_run_cb_work);
 	cb->cb_status = 0;
 	cb->cb_need_restart = false;
 	cb->cb_holds_slot = false;
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index fac938f563ad..a644460f3a5e 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -3490,11 +3490,13 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 		    struct dentry *dentry, const u32 *bmval,
 		    int ignore_crossmnt)
 {
+	DECLARE_BITMAP(attr_bitmap, ARRAY_SIZE(nfsd4_enc_fattr4_encode_ops));
 	struct nfsd4_fattr_args args;
 	struct svc_fh *tempfh = NULL;
 	int starting_len = xdr->buf->len;
 	__be32 *attrlen_p, status;
 	int attrlen_offset;
+	u32 attrmask[3];
 	int err;
 	struct nfsd4_compoundres *resp = rqstp->rq_resp;
 	u32 minorversion = resp->cstate.minorversion;
@@ -3502,10 +3504,6 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 		.mnt	= exp->ex_path.mnt,
 		.dentry	= dentry,
 	};
-	union {
-		u32		attrmask[3];
-		unsigned long	mask[2];
-	} u;
 	unsigned long bit;
 	bool file_modified = false;
 	u64 size = 0;
@@ -3517,24 +3515,24 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 	args.exp = exp;
 	args.dentry = dentry;
 	args.ignore_crossmnt = (ignore_crossmnt != 0);
+	args.acl = NULL;
 
 	/*
 	 * Make a local copy of the attribute bitmap that can be modified.
 	 */
-	memset(&u, 0, sizeof(u));
-	u.attrmask[0] = bmval[0];
-	u.attrmask[1] = bmval[1];
-	u.attrmask[2] = bmval[2];
+	attrmask[0] = bmval[0];
+	attrmask[1] = bmval[1];
+	attrmask[2] = bmval[2];
 
 	args.rdattr_err = 0;
 	if (exp->ex_fslocs.migrated) {
-		status = fattr_handle_absent_fs(&u.attrmask[0], &u.attrmask[1],
-						&u.attrmask[2], &args.rdattr_err);
+		status = fattr_handle_absent_fs(&attrmask[0], &attrmask[1],
+						&attrmask[2], &args.rdattr_err);
 		if (status)
 			goto out;
 	}
 	args.size = 0;
-	if (u.attrmask[0] & (FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE)) {
+	if (attrmask[0] & (FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE)) {
 		status = nfsd4_deleg_getattr_conflict(rqstp, d_inode(dentry),
 					&file_modified, &size);
 		if (status)
@@ -3553,16 +3551,16 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 
 	if (!(args.stat.result_mask & STATX_BTIME))
 		/* underlying FS does not offer btime so we can't share it */
-		u.attrmask[1] &= ~FATTR4_WORD1_TIME_CREATE;
-	if ((u.attrmask[0] & (FATTR4_WORD0_FILES_AVAIL | FATTR4_WORD0_FILES_FREE |
+		attrmask[1] &= ~FATTR4_WORD1_TIME_CREATE;
+	if ((attrmask[0] & (FATTR4_WORD0_FILES_AVAIL | FATTR4_WORD0_FILES_FREE |
 			FATTR4_WORD0_FILES_TOTAL | FATTR4_WORD0_MAXNAME)) ||
-	    (u.attrmask[1] & (FATTR4_WORD1_SPACE_AVAIL | FATTR4_WORD1_SPACE_FREE |
+	    (attrmask[1] & (FATTR4_WORD1_SPACE_AVAIL | FATTR4_WORD1_SPACE_FREE |
 		       FATTR4_WORD1_SPACE_TOTAL))) {
 		err = vfs_statfs(&path, &args.statfs);
 		if (err)
 			goto out_nfserr;
 	}
-	if ((u.attrmask[0] & (FATTR4_WORD0_FILEHANDLE | FATTR4_WORD0_FSID)) &&
+	if ((attrmask[0] & (FATTR4_WORD0_FILEHANDLE | FATTR4_WORD0_FSID)) &&
 	    !fhp) {
 		tempfh = kmalloc(sizeof(struct svc_fh), GFP_KERNEL);
 		status = nfserr_jukebox;
@@ -3576,11 +3574,10 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 	} else
 		args.fhp = fhp;
 
-	args.acl = NULL;
-	if (u.attrmask[0] & FATTR4_WORD0_ACL) {
+	if (attrmask[0] & FATTR4_WORD0_ACL) {
 		err = nfsd4_get_nfs4_acl(rqstp, dentry, &args.acl);
 		if (err == -EOPNOTSUPP)
-			u.attrmask[0] &= ~FATTR4_WORD0_ACL;
+			attrmask[0] &= ~FATTR4_WORD0_ACL;
 		else if (err == -EINVAL) {
 			status = nfserr_attrnotsupp;
 			goto out;
@@ -3592,17 +3589,17 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 
 #ifdef CONFIG_NFSD_V4_SECURITY_LABEL
 	args.context = NULL;
-	if ((u.attrmask[2] & FATTR4_WORD2_SECURITY_LABEL) ||
-	     u.attrmask[0] & FATTR4_WORD0_SUPPORTED_ATTRS) {
+	if ((attrmask[2] & FATTR4_WORD2_SECURITY_LABEL) ||
+	     attrmask[0] & FATTR4_WORD0_SUPPORTED_ATTRS) {
 		if (exp->ex_flags & NFSEXP_SECURITY_LABEL)
 			err = security_inode_getsecctx(d_inode(dentry),
 						&args.context, &args.contextlen);
 		else
 			err = -EOPNOTSUPP;
 		args.contextsupport = (err == 0);
-		if (u.attrmask[2] & FATTR4_WORD2_SECURITY_LABEL) {
+		if (attrmask[2] & FATTR4_WORD2_SECURITY_LABEL) {
 			if (err == -EOPNOTSUPP)
-				u.attrmask[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
+				attrmask[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
 			else if (err)
 				goto out_nfserr;
 		}
@@ -3610,8 +3607,8 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 #endif /* CONFIG_NFSD_V4_SECURITY_LABEL */
 
 	/* attrmask */
-	status = nfsd4_encode_bitmap4(xdr, u.attrmask[0],
-				      u.attrmask[1], u.attrmask[2]);
+	status = nfsd4_encode_bitmap4(xdr, attrmask[0], attrmask[1],
+				      attrmask[2]);
 	if (status)
 		goto out;
 
@@ -3620,7 +3617,9 @@ nfsd4_encode_fattr4(struct svc_rqst *rqstp, struct xdr_stream *xdr,
 	attrlen_p = xdr_reserve_space(xdr, XDR_UNIT);
 	if (!attrlen_p)
 		goto out_resource;
-	for_each_set_bit(bit, (const unsigned long *)&u.mask,
+	bitmap_from_arr32(attr_bitmap, attrmask,
+			  ARRAY_SIZE(nfsd4_enc_fattr4_encode_ops));
+	for_each_set_bit(bit, attr_bitmap,
 			 ARRAY_SIZE(nfsd4_enc_fattr4_encode_ops)) {
 		status = nfsd4_enc_fattr4_encode_ops[bit](xdr, &args);
 		if (status != nfs_ok)
diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
index 01c6f3445646..2ed0fcf879fd 100644
--- a/fs/nfsd/state.h
+++ b/fs/nfsd/state.h
@@ -68,7 +68,7 @@ struct nfsd4_callback {
 	struct nfs4_client *cb_clp;
 	struct rpc_message cb_msg;
 	const struct nfsd4_callback_ops *cb_ops;
-	struct delayed_work cb_work;
+	struct work_struct cb_work;
 	int cb_seq_status;
 	int cb_status;
 	bool cb_need_restart;
diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c
index bc846b904b68..aee40db7a036 100644
--- a/fs/nilfs2/dir.c
+++ b/fs/nilfs2/dir.c
@@ -240,7 +240,7 @@ nilfs_filetype_table[NILFS_FT_MAX] = {
 
 #define S_SHIFT 12
 static unsigned char
-nilfs_type_by_mode[S_IFMT >> S_SHIFT] = {
+nilfs_type_by_mode[(S_IFMT >> S_SHIFT) + 1] = {
 	[S_IFREG >> S_SHIFT]	= NILFS_FT_REG_FILE,
 	[S_IFDIR >> S_SHIFT]	= NILFS_FT_DIR,
 	[S_IFCHR >> S_SHIFT]	= NILFS_FT_CHRDEV,
diff --git a/fs/ntfs3/Kconfig b/fs/ntfs3/Kconfig
index cdfdf51e55d7..7bc31d69f680 100644
--- a/fs/ntfs3/Kconfig
+++ b/fs/ntfs3/Kconfig
@@ -46,3 +46,12 @@ config NTFS3_FS_POSIX_ACL
 	  NOTE: this is linux only feature. Windows will ignore these ACLs.
 
 	  If you don't know what Access Control Lists are, say N.
+
+config NTFS_FS
+	tristate "NTFS file system support"
+	select NTFS3_FS
+	select BUFFER_HEAD
+	select NLS
+	help
+	  This config option is here only for backward compatibility. NTFS
+	  filesystem is now handled by the NTFS3 driver.
diff --git a/fs/ntfs3/dir.c b/fs/ntfs3/dir.c
index 5cf3d9decf64..263635199b60 100644
--- a/fs/ntfs3/dir.c
+++ b/fs/ntfs3/dir.c
@@ -616,4 +616,11 @@ const struct file_operations ntfs_dir_operations = {
 	.compat_ioctl   = ntfs_compat_ioctl,
 #endif
 };
+
+const struct file_operations ntfs_legacy_dir_operations = {
+	.llseek		= generic_file_llseek,
+	.read		= generic_read_dir,
+	.iterate_shared	= ntfs_readdir,
+	.open		= ntfs_file_open,
+};
 // clang-format on
diff --git a/fs/ntfs3/file.c b/fs/ntfs3/file.c
index 5418662c80d8..b73969e05052 100644
--- a/fs/ntfs3/file.c
+++ b/fs/ntfs3/file.c
@@ -1236,4 +1236,12 @@ const struct file_operations ntfs_file_operations = {
 	.fallocate	= ntfs_fallocate,
 	.release	= ntfs_file_release,
 };
+
+const struct file_operations ntfs_legacy_file_operations = {
+	.llseek		= generic_file_llseek,
+	.read_iter	= ntfs_file_read_iter,
+	.splice_read	= ntfs_file_splice_read,
+	.open		= ntfs_file_open,
+	.release	= ntfs_file_release,
+};
 // clang-format on
diff --git a/fs/ntfs3/inode.c b/fs/ntfs3/inode.c
index eb7a8c9fba01..d273eda1cf45 100644
--- a/fs/ntfs3/inode.c
+++ b/fs/ntfs3/inode.c
@@ -440,7 +440,10 @@ end_enum:
 		 * Usually a hard links to directories are disabled.
 		 */
 		inode->i_op = &ntfs_dir_inode_operations;
-		inode->i_fop = &ntfs_dir_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_dir_operations;
+		else
+			inode->i_fop = &ntfs_dir_operations;
 		ni->i_valid = 0;
 	} else if (S_ISLNK(mode)) {
 		ni->std_fa &= ~FILE_ATTRIBUTE_DIRECTORY;
@@ -450,7 +453,10 @@ end_enum:
 	} else if (S_ISREG(mode)) {
 		ni->std_fa &= ~FILE_ATTRIBUTE_DIRECTORY;
 		inode->i_op = &ntfs_file_inode_operations;
-		inode->i_fop = &ntfs_file_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_file_operations;
+		else
+			inode->i_fop = &ntfs_file_operations;
 		inode->i_mapping->a_ops = is_compressed(ni) ? &ntfs_aops_cmpr :
 							      &ntfs_aops;
 		if (ino != MFT_REC_MFT)
@@ -1614,7 +1620,10 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 
 	if (S_ISDIR(mode)) {
 		inode->i_op = &ntfs_dir_inode_operations;
-		inode->i_fop = &ntfs_dir_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_dir_operations;
+		else
+			inode->i_fop = &ntfs_dir_operations;
 	} else if (S_ISLNK(mode)) {
 		inode->i_op = &ntfs_link_inode_operations;
 		inode->i_fop = NULL;
@@ -1623,7 +1632,10 @@ struct inode *ntfs_create_inode(struct mnt_idmap *idmap, struct inode *dir,
 		inode_nohighmem(inode);
 	} else if (S_ISREG(mode)) {
 		inode->i_op = &ntfs_file_inode_operations;
-		inode->i_fop = &ntfs_file_operations;
+		if (is_legacy_ntfs(inode->i_sb))
+			inode->i_fop = &ntfs_legacy_file_operations;
+		else
+			inode->i_fop = &ntfs_file_operations;
 		inode->i_mapping->a_ops = is_compressed(ni) ? &ntfs_aops_cmpr :
 							      &ntfs_aops;
 		init_rwsem(&ni->file.run_lock);
diff --git a/fs/ntfs3/ntfs_fs.h b/fs/ntfs3/ntfs_fs.h
index 79356fd29a14..5f4d288c6adf 100644
--- a/fs/ntfs3/ntfs_fs.h
+++ b/fs/ntfs3/ntfs_fs.h
@@ -493,6 +493,7 @@ struct inode *dir_search_u(struct inode *dir, const struct cpu_str *uni,
 			   struct ntfs_fnd *fnd);
 bool dir_is_empty(struct inode *dir);
 extern const struct file_operations ntfs_dir_operations;
+extern const struct file_operations ntfs_legacy_dir_operations;
 
 /* Globals from file.c */
 int ntfs_getattr(struct mnt_idmap *idmap, const struct path *path,
@@ -507,6 +508,7 @@ long ntfs_compat_ioctl(struct file *filp, u32 cmd, unsigned long arg);
 extern const struct inode_operations ntfs_special_inode_operations;
 extern const struct inode_operations ntfs_file_inode_operations;
 extern const struct file_operations ntfs_file_operations;
+extern const struct file_operations ntfs_legacy_file_operations;
 
 /* Globals from frecord.c */
 void ni_remove_mi(struct ntfs_inode *ni, struct mft_inode *mi);
@@ -1154,4 +1156,6 @@ static inline void le64_sub_cpu(__le64 *var, u64 val)
 	*var = cpu_to_le64(le64_to_cpu(*var) - val);
 }
 
+bool is_legacy_ntfs(struct super_block *sb);
+
 #endif /* _LINUX_NTFS3_NTFS_FS_H */
diff --git a/fs/ntfs3/super.c b/fs/ntfs3/super.c
index 9df7c20d066f..b26d95a8d327 100644
--- a/fs/ntfs3/super.c
+++ b/fs/ntfs3/super.c
@@ -408,6 +408,12 @@ static int ntfs_fs_reconfigure(struct fs_context *fc)
 	struct ntfs_mount_options *new_opts = fc->fs_private;
 	int ro_rw;
 
+	/* If ntfs3 is used as legacy ntfs enforce read-only mode. */
+	if (is_legacy_ntfs(sb)) {
+		fc->sb_flags |= SB_RDONLY;
+		goto out;
+	}
+
 	ro_rw = sb_rdonly(sb) && !(fc->sb_flags & SB_RDONLY);
 	if (ro_rw && (sbi->flags & NTFS_FLAGS_NEED_REPLAY)) {
 		errorf(fc,
@@ -427,8 +433,6 @@ static int ntfs_fs_reconfigure(struct fs_context *fc)
 			fc,
 			"ntfs3: Cannot use different iocharset when remounting!");
 
-	sync_filesystem(sb);
-
 	if (ro_rw && (sbi->volume.flags & VOLUME_FLAG_DIRTY) &&
 	    !new_opts->force) {
 		errorf(fc,
@@ -436,6 +440,8 @@ static int ntfs_fs_reconfigure(struct fs_context *fc)
 		return -EINVAL;
 	}
 
+out:
+	sync_filesystem(sb);
 	swap(sbi->options, fc->fs_private);
 
 	return 0;
@@ -1613,6 +1619,8 @@ load_root:
 	}
 #endif
 
+	if (is_legacy_ntfs(sb))
+		sb->s_flags |= SB_RDONLY;
 	return 0;
 
 put_inode_out:
@@ -1730,7 +1738,7 @@ static const struct fs_context_operations ntfs_context_ops = {
  * This will called when mount/remount. We will first initialize
  * options so that if remount we can use just that.
  */
-static int ntfs_init_fs_context(struct fs_context *fc)
+static int __ntfs_init_fs_context(struct fs_context *fc)
 {
 	struct ntfs_mount_options *opts;
 	struct ntfs_sb_info *sbi;
@@ -1778,6 +1786,11 @@ free_opts:
 	return -ENOMEM;
 }
 
+static int ntfs_init_fs_context(struct fs_context *fc)
+{
+	return __ntfs_init_fs_context(fc);
+}
+
 static void ntfs3_kill_sb(struct super_block *sb)
 {
 	struct ntfs_sb_info *sbi = sb->s_fs_info;
@@ -1798,6 +1811,50 @@ static struct file_system_type ntfs_fs_type = {
 	.kill_sb		= ntfs3_kill_sb,
 	.fs_flags		= FS_REQUIRES_DEV | FS_ALLOW_IDMAP,
 };
+
+#if IS_ENABLED(CONFIG_NTFS_FS)
+static int ntfs_legacy_init_fs_context(struct fs_context *fc)
+{
+	int ret;
+
+	ret = __ntfs_init_fs_context(fc);
+	/* If ntfs3 is used as legacy ntfs enforce read-only mode. */
+	fc->sb_flags |= SB_RDONLY;
+	return ret;
+}
+
+static struct file_system_type ntfs_legacy_fs_type = {
+	.owner			= THIS_MODULE,
+	.name			= "ntfs",
+	.init_fs_context	= ntfs_legacy_init_fs_context,
+	.parameters		= ntfs_fs_parameters,
+	.kill_sb		= ntfs3_kill_sb,
+	.fs_flags		= FS_REQUIRES_DEV | FS_ALLOW_IDMAP,
+};
+MODULE_ALIAS_FS("ntfs");
+
+static inline void register_as_ntfs_legacy(void)
+{
+	int err = register_filesystem(&ntfs_legacy_fs_type);
+	if (err)
+		pr_warn("ntfs3: Failed to register legacy ntfs filesystem driver: %d\n", err);
+}
+
+static inline void unregister_as_ntfs_legacy(void)
+{
+	unregister_filesystem(&ntfs_legacy_fs_type);
+}
+bool is_legacy_ntfs(struct super_block *sb)
+{
+	return sb->s_type == &ntfs_legacy_fs_type;
+}
+#else
+static inline void register_as_ntfs_legacy(void) {}
+static inline void unregister_as_ntfs_legacy(void) {}
+bool is_legacy_ntfs(struct super_block *sb) { return false; }
+#endif
+
+
 // clang-format on
 
 static int __init init_ntfs_fs(void)
@@ -1832,6 +1889,7 @@ static int __init init_ntfs_fs(void)
 		goto out1;
 	}
 
+	register_as_ntfs_legacy();
 	err = register_filesystem(&ntfs_fs_type);
 	if (err)
 		goto out;
@@ -1849,6 +1907,7 @@ static void __exit exit_ntfs_fs(void)
 	rcu_barrier();
 	kmem_cache_destroy(ntfs_inode_cachep);
 	unregister_filesystem(&ntfs_fs_type);
+	unregister_as_ntfs_legacy();
 	ntfs3_exit_bitmap();
 
 #ifdef CONFIG_PROC_FS
diff --git a/fs/proc/bootconfig.c b/fs/proc/bootconfig.c
index 902b326e1e56..87dcaae32ff8 100644
--- a/fs/proc/bootconfig.c
+++ b/fs/proc/bootconfig.c
@@ -62,12 +62,12 @@ static int __init copy_xbc_key_value_list(char *dst, size_t size)
 				break;
 			dst += ret;
 		}
-		if (ret >= 0 && boot_command_line[0]) {
-			ret = snprintf(dst, rest(dst, end), "# Parameters from bootloader:\n# %s\n",
-				       boot_command_line);
-			if (ret > 0)
-				dst += ret;
-		}
+	}
+	if (cmdline_has_extra_options() && ret >= 0 && boot_command_line[0]) {
+		ret = snprintf(dst, rest(dst, end), "# Parameters from bootloader:\n# %s\n",
+			       boot_command_line);
+		if (ret > 0)
+			dst += ret;
 	}
 out:
 	kfree(key);
diff --git a/fs/proc/page.c b/fs/proc/page.c
index 195b077c0fac..9223856c934b 100644
--- a/fs/proc/page.c
+++ b/fs/proc/page.c
@@ -67,7 +67,7 @@ static ssize_t kpagecount_read(struct file *file, char __user *buf,
 		 */
 		ppage = pfn_to_online_page(pfn);
 
-		if (!ppage || PageSlab(ppage) || page_has_type(ppage))
+		if (!ppage)
 			pcount = 0;
 		else
 			pcount = page_mapcount(ppage);
@@ -124,11 +124,8 @@ u64 stable_page_flags(struct page *page)
 
 	/*
 	 * pseudo flags for the well known (anonymous) memory mapped pages
-	 *
-	 * Note that page->_mapcount is overloaded in SLAB, so the
-	 * simple test in page_mapped() is not enough.
 	 */
-	if (!PageSlab(page) && page_mapped(page))
+	if (page_mapped(page))
 		u |= 1 << KPF_MMAP;
 	if (PageAnon(page))
 		u |= 1 << KPF_ANON;
diff --git a/fs/smb/client/cached_dir.c b/fs/smb/client/cached_dir.c
index 13a9d7acf8f8..0ff2491c311d 100644
--- a/fs/smb/client/cached_dir.c
+++ b/fs/smb/client/cached_dir.c
@@ -433,8 +433,8 @@ smb2_close_cached_fid(struct kref *ref)
 	if (cfid->is_open) {
 		rc = SMB2_close(0, cfid->tcon, cfid->fid.persistent_fid,
 			   cfid->fid.volatile_fid);
-		if (rc != -EBUSY && rc != -EAGAIN)
-			atomic_dec(&cfid->tcon->num_remote_opens);
+		if (rc) /* should we retry on -EBUSY or -EAGAIN? */
+			cifs_dbg(VFS, "close cached dir rc %d\n", rc);
 	}
 
 	free_cached_dir(cfid);
diff --git a/fs/smb/client/cifsfs.c b/fs/smb/client/cifsfs.c
index d41eedbff674..39277c37185c 100644
--- a/fs/smb/client/cifsfs.c
+++ b/fs/smb/client/cifsfs.c
@@ -389,6 +389,7 @@ cifs_alloc_inode(struct super_block *sb)
 	 * server, can not assume caching of file data or metadata.
 	 */
 	cifs_set_oplock_level(cifs_inode, 0);
+	cifs_inode->lease_granted = false;
 	cifs_inode->flags = 0;
 	spin_lock_init(&cifs_inode->writers_lock);
 	cifs_inode->writers = 0;
@@ -739,6 +740,8 @@ static void cifs_umount_begin(struct super_block *sb)
 
 	spin_lock(&cifs_tcp_ses_lock);
 	spin_lock(&tcon->tc_lock);
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+			    netfs_trace_tcon_ref_see_umount);
 	if ((tcon->tc_count > 1) || (tcon->status == TID_EXITING)) {
 		/* we have other mounts to same share or we have
 		   already tried to umount this and woken up
diff --git a/fs/smb/client/cifsglob.h b/fs/smb/client/cifsglob.h
index f6a302205f89..6ff35570db81 100644
--- a/fs/smb/client/cifsglob.h
+++ b/fs/smb/client/cifsglob.h
@@ -1077,6 +1077,7 @@ struct cifs_ses {
 				   and after mount option parsing we fill it */
 	char *domainName;
 	char *password;
+	char *password2; /* When key rotation used, new password may be set before it expires */
 	char workstation_name[CIFS_MAX_WORKSTATION_LEN];
 	struct session_key auth_key;
 	struct ntlmssp_auth *ntlmssp; /* ciphertext, flags, server challenge */
@@ -1189,6 +1190,7 @@ struct cifs_fattr {
  */
 struct cifs_tcon {
 	struct list_head tcon_list;
+	int debug_id;		/* Debugging for tracing */
 	int tc_count;
 	struct list_head rlist; /* reconnect list */
 	spinlock_t tc_lock;  /* protect anything here that is not protected */
@@ -1275,7 +1277,9 @@ struct cifs_tcon {
 	__u32 max_cached_dirs;
 #ifdef CONFIG_CIFS_FSCACHE
 	u64 resource_id;		/* server resource id */
+	bool fscache_acquired;		/* T if we've tried acquiring a cookie */
 	struct fscache_volume *fscache;	/* cookie for share */
+	struct mutex fscache_lock;	/* Prevent regetting a cookie */
 #endif
 	struct list_head pending_opens;	/* list of incomplete opens */
 	struct cached_fids *cfids;
diff --git a/fs/smb/client/cifspdu.h b/fs/smb/client/cifspdu.h
index c0513fbb8a59..c46d418c1c0c 100644
--- a/fs/smb/client/cifspdu.h
+++ b/fs/smb/client/cifspdu.h
@@ -882,7 +882,7 @@ typedef struct smb_com_open_rsp {
 	__u8 OplockLevel;
 	__u16 Fid;
 	__le32 CreateAction;
-	struct_group(common_attributes,
+	struct_group_attr(common_attributes, __packed,
 		__le64 CreationTime;
 		__le64 LastAccessTime;
 		__le64 LastWriteTime;
@@ -2266,7 +2266,7 @@ typedef struct {
 /* QueryFileInfo/QueryPathinfo (also for SetPath/SetFile) data buffer formats */
 /******************************************************************************/
 typedef struct { /* data block encoding of response to level 263 QPathInfo */
-	struct_group(common_attributes,
+	struct_group_attr(common_attributes, __packed,
 		__le64 CreationTime;
 		__le64 LastAccessTime;
 		__le64 LastWriteTime;
diff --git a/fs/smb/client/cifsproto.h b/fs/smb/client/cifsproto.h
index 8e0a348f1f66..fbc358c09da3 100644
--- a/fs/smb/client/cifsproto.h
+++ b/fs/smb/client/cifsproto.h
@@ -303,7 +303,7 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx,
 		     struct TCP_Server_Info *primary_server);
 extern void cifs_put_tcp_session(struct TCP_Server_Info *server,
 				 int from_reconnect);
-extern void cifs_put_tcon(struct cifs_tcon *tcon);
+extern void cifs_put_tcon(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace);
 
 extern void cifs_release_automount_timer(void);
 
@@ -530,8 +530,9 @@ extern int CIFSSMBLogoff(const unsigned int xid, struct cifs_ses *ses);
 
 extern struct cifs_ses *sesInfoAlloc(void);
 extern void sesInfoFree(struct cifs_ses *);
-extern struct cifs_tcon *tcon_info_alloc(bool dir_leases_enabled);
-extern void tconInfoFree(struct cifs_tcon *);
+extern struct cifs_tcon *tcon_info_alloc(bool dir_leases_enabled,
+					 enum smb3_tcon_ref_trace trace);
+extern void tconInfoFree(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace);
 
 extern int cifs_sign_rqst(struct smb_rqst *rqst, struct TCP_Server_Info *server,
 		   __u32 *pexpected_response_sequence_number);
@@ -721,8 +722,6 @@ static inline int cifs_create_options(struct cifs_sb_info *cifs_sb, int options)
 		return options;
 }
 
-struct super_block *cifs_get_tcon_super(struct cifs_tcon *tcon);
-void cifs_put_tcon_super(struct super_block *sb);
 int cifs_wait_for_server_reconnect(struct TCP_Server_Info *server, bool retry);
 
 /* Put references of @ses and its children */
diff --git a/fs/smb/client/connect.c b/fs/smb/client/connect.c
index 85679ae106fd..7a16e12f5da8 100644
--- a/fs/smb/client/connect.c
+++ b/fs/smb/client/connect.c
@@ -1943,7 +1943,7 @@ cifs_setup_ipc(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 	}
 
 	/* no need to setup directory caching on IPC share, so pass in false */
-	tcon = tcon_info_alloc(false);
+	tcon = tcon_info_alloc(false, netfs_trace_tcon_ref_new_ipc);
 	if (tcon == NULL)
 		return -ENOMEM;
 
@@ -1960,7 +1960,7 @@ cifs_setup_ipc(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 
 	if (rc) {
 		cifs_server_dbg(VFS, "failed to connect to IPC (rc=%d)\n", rc);
-		tconInfoFree(tcon);
+		tconInfoFree(tcon, netfs_trace_tcon_ref_free_ipc_fail);
 		goto out;
 	}
 
@@ -2043,7 +2043,7 @@ void __cifs_put_smb_ses(struct cifs_ses *ses)
 	 * files on session close, as specified in MS-SMB2 3.3.5.6 Receiving an
 	 * SMB2 LOGOFF Request.
 	 */
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free_ipc);
 	if (do_logoff) {
 		xid = get_xid();
 		rc = server->ops->logoff(xid, ses);
@@ -2183,6 +2183,7 @@ cifs_set_cifscreds(struct smb3_fs_context *ctx, struct cifs_ses *ses)
 	}
 
 	++delim;
+	/* BB consider adding support for password2 (Key Rotation) for multiuser in future */
 	ctx->password = kstrndup(delim, len, GFP_KERNEL);
 	if (!ctx->password) {
 		cifs_dbg(FYI, "Unable to allocate %zd bytes for password\n",
@@ -2206,6 +2207,7 @@ cifs_set_cifscreds(struct smb3_fs_context *ctx, struct cifs_ses *ses)
 			kfree(ctx->username);
 			ctx->username = NULL;
 			kfree_sensitive(ctx->password);
+			/* no need to free ctx->password2 since not allocated in this path */
 			ctx->password = NULL;
 			goto out_key_put;
 		}
@@ -2317,6 +2319,12 @@ cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb3_fs_context *ctx)
 		if (!ses->password)
 			goto get_ses_fail;
 	}
+	/* ctx->password freed at unmount */
+	if (ctx->password2) {
+		ses->password2 = kstrdup(ctx->password2, GFP_KERNEL);
+		if (!ses->password2)
+			goto get_ses_fail;
+	}
 	if (ctx->domainname) {
 		ses->domainName = kstrdup(ctx->domainname, GFP_KERNEL);
 		if (!ses->domainName)
@@ -2424,6 +2432,8 @@ cifs_find_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 			continue;
 		}
 		++tcon->tc_count;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_get_find);
 		spin_unlock(&tcon->tc_lock);
 		spin_unlock(&cifs_tcp_ses_lock);
 		return tcon;
@@ -2433,7 +2443,7 @@ cifs_find_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 }
 
 void
-cifs_put_tcon(struct cifs_tcon *tcon)
+cifs_put_tcon(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace)
 {
 	unsigned int xid;
 	struct cifs_ses *ses;
@@ -2449,6 +2459,7 @@ cifs_put_tcon(struct cifs_tcon *tcon)
 	cifs_dbg(FYI, "%s: tc_count=%d\n", __func__, tcon->tc_count);
 	spin_lock(&cifs_tcp_ses_lock);
 	spin_lock(&tcon->tc_lock);
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count - 1, trace);
 	if (--tcon->tc_count > 0) {
 		spin_unlock(&tcon->tc_lock);
 		spin_unlock(&cifs_tcp_ses_lock);
@@ -2485,7 +2496,7 @@ cifs_put_tcon(struct cifs_tcon *tcon)
 	_free_xid(xid);
 
 	cifs_fscache_release_super_cookie(tcon);
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free);
 	cifs_put_smb_ses(ses);
 }
 
@@ -2539,7 +2550,7 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 		nohandlecache = ctx->nohandlecache;
 	else
 		nohandlecache = true;
-	tcon = tcon_info_alloc(!nohandlecache);
+	tcon = tcon_info_alloc(!nohandlecache, netfs_trace_tcon_ref_new);
 	if (tcon == NULL) {
 		rc = -ENOMEM;
 		goto out_fail;
@@ -2729,7 +2740,7 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb3_fs_context *ctx)
 	return tcon;
 
 out_fail:
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free_fail);
 	return ERR_PTR(rc);
 }
 
@@ -2746,7 +2757,7 @@ cifs_put_tlink(struct tcon_link *tlink)
 	}
 
 	if (!IS_ERR(tlink_tcon(tlink)))
-		cifs_put_tcon(tlink_tcon(tlink));
+		cifs_put_tcon(tlink_tcon(tlink), netfs_trace_tcon_ref_put_tlink);
 	kfree(tlink);
 }
 
@@ -3311,7 +3322,7 @@ void cifs_mount_put_conns(struct cifs_mount_ctx *mnt_ctx)
 	int rc = 0;
 
 	if (mnt_ctx->tcon)
-		cifs_put_tcon(mnt_ctx->tcon);
+		cifs_put_tcon(mnt_ctx->tcon, netfs_trace_tcon_ref_put_mnt_ctx);
 	else if (mnt_ctx->ses)
 		cifs_put_smb_ses(mnt_ctx->ses);
 	else if (mnt_ctx->server)
diff --git a/fs/smb/client/fs_context.c b/fs/smb/client/fs_context.c
index b7bfe705b2c4..3bbac925d076 100644
--- a/fs/smb/client/fs_context.c
+++ b/fs/smb/client/fs_context.c
@@ -162,6 +162,7 @@ const struct fs_parameter_spec smb3_fs_parameters[] = {
 	fsparam_string("username", Opt_user),
 	fsparam_string("pass", Opt_pass),
 	fsparam_string("password", Opt_pass),
+	fsparam_string("password2", Opt_pass2),
 	fsparam_string("ip", Opt_ip),
 	fsparam_string("addr", Opt_ip),
 	fsparam_string("domain", Opt_domain),
@@ -345,6 +346,7 @@ smb3_fs_context_dup(struct smb3_fs_context *new_ctx, struct smb3_fs_context *ctx
 	new_ctx->nodename = NULL;
 	new_ctx->username = NULL;
 	new_ctx->password = NULL;
+	new_ctx->password2 = NULL;
 	new_ctx->server_hostname = NULL;
 	new_ctx->domainname = NULL;
 	new_ctx->UNC = NULL;
@@ -357,6 +359,7 @@ smb3_fs_context_dup(struct smb3_fs_context *new_ctx, struct smb3_fs_context *ctx
 	DUP_CTX_STR(prepath);
 	DUP_CTX_STR(username);
 	DUP_CTX_STR(password);
+	DUP_CTX_STR(password2);
 	DUP_CTX_STR(server_hostname);
 	DUP_CTX_STR(UNC);
 	DUP_CTX_STR(source);
@@ -745,6 +748,16 @@ static int smb3_fs_context_validate(struct fs_context *fc)
 	/* set the port that we got earlier */
 	cifs_set_port((struct sockaddr *)&ctx->dstaddr, ctx->port);
 
+	if (ctx->uid_specified && !ctx->forceuid_specified) {
+		ctx->override_uid = 1;
+		pr_notice("enabling forceuid mount option implicitly because uid= option is specified\n");
+	}
+
+	if (ctx->gid_specified && !ctx->forcegid_specified) {
+		ctx->override_gid = 1;
+		pr_notice("enabling forcegid mount option implicitly because gid= option is specified\n");
+	}
+
 	if (ctx->override_uid && !ctx->uid_specified) {
 		ctx->override_uid = 0;
 		pr_notice("ignoring forceuid mount option specified with no uid= option\n");
@@ -905,6 +918,8 @@ static int smb3_reconfigure(struct fs_context *fc)
 	else  {
 		kfree_sensitive(ses->password);
 		ses->password = kstrdup(ctx->password, GFP_KERNEL);
+		kfree_sensitive(ses->password2);
+		ses->password2 = kstrdup(ctx->password2, GFP_KERNEL);
 	}
 	STEAL_STRING(cifs_sb, ctx, domainname);
 	STEAL_STRING(cifs_sb, ctx, nodename);
@@ -1014,12 +1029,14 @@ static int smb3_fs_context_parse_param(struct fs_context *fc,
 			ctx->override_uid = 0;
 		else
 			ctx->override_uid = 1;
+		ctx->forceuid_specified = true;
 		break;
 	case Opt_forcegid:
 		if (result.negated)
 			ctx->override_gid = 0;
 		else
 			ctx->override_gid = 1;
+		ctx->forcegid_specified = true;
 		break;
 	case Opt_perm:
 		if (result.negated)
@@ -1305,6 +1322,18 @@ static int smb3_fs_context_parse_param(struct fs_context *fc,
 			goto cifs_parse_mount_err;
 		}
 		break;
+	case Opt_pass2:
+		kfree_sensitive(ctx->password2);
+		ctx->password2 = NULL;
+		if (strlen(param->string) == 0)
+			break;
+
+		ctx->password2 = kstrdup(param->string, GFP_KERNEL);
+		if (ctx->password2 == NULL) {
+			cifs_errorf(fc, "OOM when copying password2 string\n");
+			goto cifs_parse_mount_err;
+		}
+		break;
 	case Opt_ip:
 		if (strlen(param->string) == 0) {
 			ctx->got_ip = false;
@@ -1608,6 +1637,8 @@ static int smb3_fs_context_parse_param(struct fs_context *fc,
  cifs_parse_mount_err:
 	kfree_sensitive(ctx->password);
 	ctx->password = NULL;
+	kfree_sensitive(ctx->password2);
+	ctx->password2 = NULL;
 	return -EINVAL;
 }
 
@@ -1713,6 +1744,8 @@ smb3_cleanup_fs_context_contents(struct smb3_fs_context *ctx)
 	ctx->username = NULL;
 	kfree_sensitive(ctx->password);
 	ctx->password = NULL;
+	kfree_sensitive(ctx->password2);
+	ctx->password2 = NULL;
 	kfree(ctx->server_hostname);
 	ctx->server_hostname = NULL;
 	kfree(ctx->UNC);
diff --git a/fs/smb/client/fs_context.h b/fs/smb/client/fs_context.h
index 8a35645e0b65..cf577ec0dd0a 100644
--- a/fs/smb/client/fs_context.h
+++ b/fs/smb/client/fs_context.h
@@ -145,6 +145,7 @@ enum cifs_param {
 	Opt_source,
 	Opt_user,
 	Opt_pass,
+	Opt_pass2,
 	Opt_ip,
 	Opt_domain,
 	Opt_srcaddr,
@@ -164,6 +165,8 @@ enum cifs_param {
 };
 
 struct smb3_fs_context {
+	bool forceuid_specified;
+	bool forcegid_specified;
 	bool uid_specified;
 	bool cruid_specified;
 	bool gid_specified;
@@ -177,6 +180,7 @@ struct smb3_fs_context {
 
 	char *username;
 	char *password;
+	char *password2;
 	char *domainname;
 	char *source;
 	char *server_hostname;
diff --git a/fs/smb/client/fscache.c b/fs/smb/client/fscache.c
index 340efce8f052..1a895e6243ee 100644
--- a/fs/smb/client/fscache.c
+++ b/fs/smb/client/fscache.c
@@ -43,12 +43,23 @@ int cifs_fscache_get_super_cookie(struct cifs_tcon *tcon)
 	char *key;
 	int ret = -ENOMEM;
 
+	if (tcon->fscache_acquired)
+		return 0;
+
+	mutex_lock(&tcon->fscache_lock);
+	if (tcon->fscache_acquired) {
+		mutex_unlock(&tcon->fscache_lock);
+		return 0;
+	}
+	tcon->fscache_acquired = true;
+
 	tcon->fscache = NULL;
 	switch (sa->sa_family) {
 	case AF_INET:
 	case AF_INET6:
 		break;
 	default:
+		mutex_unlock(&tcon->fscache_lock);
 		cifs_dbg(VFS, "Unknown network family '%d'\n", sa->sa_family);
 		return -EINVAL;
 	}
@@ -57,6 +68,7 @@ int cifs_fscache_get_super_cookie(struct cifs_tcon *tcon)
 
 	sharename = extract_sharename(tcon->tree_name);
 	if (IS_ERR(sharename)) {
+		mutex_unlock(&tcon->fscache_lock);
 		cifs_dbg(FYI, "%s: couldn't extract sharename\n", __func__);
 		return PTR_ERR(sharename);
 	}
@@ -82,6 +94,11 @@ int cifs_fscache_get_super_cookie(struct cifs_tcon *tcon)
 		}
 		pr_err("Cache volume key already in use (%s)\n", key);
 		vcookie = NULL;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_see_fscache_collision);
+	} else {
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_see_fscache_okay);
 	}
 
 	tcon->fscache = vcookie;
@@ -90,6 +107,7 @@ out_2:
 	kfree(key);
 out:
 	kfree(sharename);
+	mutex_unlock(&tcon->fscache_lock);
 	return ret;
 }
 
@@ -102,6 +120,8 @@ void cifs_fscache_release_super_cookie(struct cifs_tcon *tcon)
 	cifs_fscache_fill_volume_coherency(tcon, &cd);
 	fscache_relinquish_volume(tcon->fscache, &cd, false);
 	tcon->fscache = NULL;
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+			    netfs_trace_tcon_ref_see_fscache_relinq);
 }
 
 void cifs_fscache_get_inode_cookie(struct inode *inode)
diff --git a/fs/smb/client/inode.c b/fs/smb/client/inode.c
index 91b07ef9e25c..60afab5c83d4 100644
--- a/fs/smb/client/inode.c
+++ b/fs/smb/client/inode.c
@@ -1105,7 +1105,8 @@ static int cifs_get_fattr(struct cifs_open_info_data *data,
 		} else {
 			cifs_open_info_to_fattr(fattr, data, sb);
 		}
-		if (!rc && fattr->cf_flags & CIFS_FATTR_DELETE_PENDING)
+		if (!rc && *inode &&
+		    (fattr->cf_flags & CIFS_FATTR_DELETE_PENDING))
 			cifs_mark_open_handles_for_deleted_file(*inode, full_path);
 		break;
 	case -EREMOTE:
diff --git a/fs/smb/client/misc.c b/fs/smb/client/misc.c
index 33ac4f8f5050..07c468ddb88a 100644
--- a/fs/smb/client/misc.c
+++ b/fs/smb/client/misc.c
@@ -98,6 +98,7 @@ sesInfoFree(struct cifs_ses *buf_to_free)
 	kfree(buf_to_free->serverDomain);
 	kfree(buf_to_free->serverNOS);
 	kfree_sensitive(buf_to_free->password);
+	kfree_sensitive(buf_to_free->password2);
 	kfree(buf_to_free->user_name);
 	kfree(buf_to_free->domainName);
 	kfree_sensitive(buf_to_free->auth_key.response);
@@ -110,9 +111,10 @@ sesInfoFree(struct cifs_ses *buf_to_free)
 }
 
 struct cifs_tcon *
-tcon_info_alloc(bool dir_leases_enabled)
+tcon_info_alloc(bool dir_leases_enabled, enum smb3_tcon_ref_trace trace)
 {
 	struct cifs_tcon *ret_buf;
+	static atomic_t tcon_debug_id;
 
 	ret_buf = kzalloc(sizeof(*ret_buf), GFP_KERNEL);
 	if (!ret_buf)
@@ -129,7 +131,8 @@ tcon_info_alloc(bool dir_leases_enabled)
 
 	atomic_inc(&tconInfoAllocCount);
 	ret_buf->status = TID_NEW;
-	++ret_buf->tc_count;
+	ret_buf->debug_id = atomic_inc_return(&tcon_debug_id);
+	ret_buf->tc_count = 1;
 	spin_lock_init(&ret_buf->tc_lock);
 	INIT_LIST_HEAD(&ret_buf->openFileList);
 	INIT_LIST_HEAD(&ret_buf->tcon_list);
@@ -138,17 +141,22 @@ tcon_info_alloc(bool dir_leases_enabled)
 	atomic_set(&ret_buf->num_local_opens, 0);
 	atomic_set(&ret_buf->num_remote_opens, 0);
 	ret_buf->stats_from_time = ktime_get_real_seconds();
+#ifdef CONFIG_CIFS_FSCACHE
+	mutex_init(&ret_buf->fscache_lock);
+#endif
+	trace_smb3_tcon_ref(ret_buf->debug_id, ret_buf->tc_count, trace);
 
 	return ret_buf;
 }
 
 void
-tconInfoFree(struct cifs_tcon *tcon)
+tconInfoFree(struct cifs_tcon *tcon, enum smb3_tcon_ref_trace trace)
 {
 	if (tcon == NULL) {
 		cifs_dbg(FYI, "Null buffer passed to tconInfoFree\n");
 		return;
 	}
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count, trace);
 	free_cached_dirs(tcon->cfids);
 	atomic_dec(&tconInfoAllocCount);
 	kfree(tcon->nativeFileSystem);
diff --git a/fs/smb/client/smb2misc.c b/fs/smb/client/smb2misc.c
index cc72be5a93a9..677ef6f99a5b 100644
--- a/fs/smb/client/smb2misc.c
+++ b/fs/smb/client/smb2misc.c
@@ -767,7 +767,7 @@ smb2_cancelled_close_fid(struct work_struct *work)
 	if (rc)
 		cifs_tcon_dbg(VFS, "Close cancelled mid failed rc:%d\n", rc);
 
-	cifs_put_tcon(tcon);
+	cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_cancelled_close_fid);
 	kfree(cancelled);
 }
 
@@ -811,6 +811,8 @@ smb2_handle_cancelled_close(struct cifs_tcon *tcon, __u64 persistent_fid,
 	if (tcon->tc_count <= 0) {
 		struct TCP_Server_Info *server = NULL;
 
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_see_cancelled_close);
 		WARN_ONCE(tcon->tc_count < 0, "tcon refcount is negative");
 		spin_unlock(&cifs_tcp_ses_lock);
 
@@ -823,12 +825,14 @@ smb2_handle_cancelled_close(struct cifs_tcon *tcon, __u64 persistent_fid,
 		return 0;
 	}
 	tcon->tc_count++;
+	trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+			    netfs_trace_tcon_ref_get_cancelled_close);
 	spin_unlock(&cifs_tcp_ses_lock);
 
 	rc = __smb2_handle_cancelled_cmd(tcon, SMB2_CLOSE_HE, 0,
 					 persistent_fid, volatile_fid);
 	if (rc)
-		cifs_put_tcon(tcon);
+		cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_cancelled_close);
 
 	return rc;
 }
@@ -856,7 +860,7 @@ smb2_handle_cancelled_mid(struct mid_q_entry *mid, struct TCP_Server_Info *serve
 					 rsp->PersistentFileId,
 					 rsp->VolatileFileId);
 	if (rc)
-		cifs_put_tcon(tcon);
+		cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_cancelled_mid);
 
 	return rc;
 }
diff --git a/fs/smb/client/smb2ops.c b/fs/smb/client/smb2ops.c
index b156eefa75d7..28f0b7d19d53 100644
--- a/fs/smb/client/smb2ops.c
+++ b/fs/smb/client/smb2ops.c
@@ -2915,8 +2915,11 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 		tcon = list_first_entry_or_null(&ses->tcon_list,
 						struct cifs_tcon,
 						tcon_list);
-		if (tcon)
+		if (tcon) {
 			tcon->tc_count++;
+			trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+					    netfs_trace_tcon_ref_get_dfs_refer);
+		}
 		spin_unlock(&cifs_tcp_ses_lock);
 	}
 
@@ -2980,6 +2983,8 @@ smb2_get_dfs_refer(const unsigned int xid, struct cifs_ses *ses,
 		/* ipc tcons are not refcounted */
 		spin_lock(&cifs_tcp_ses_lock);
 		tcon->tc_count--;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_dec_dfs_refer);
 		/* tc_count can never go negative */
 		WARN_ON(tcon->tc_count < 0);
 		spin_unlock(&cifs_tcp_ses_lock);
@@ -4964,68 +4969,84 @@ static int smb2_next_header(struct TCP_Server_Info *server, char *buf,
 	return 0;
 }
 
-int cifs_sfu_make_node(unsigned int xid, struct inode *inode,
-		       struct dentry *dentry, struct cifs_tcon *tcon,
-		       const char *full_path, umode_t mode, dev_t dev)
+static int __cifs_sfu_make_node(unsigned int xid, struct inode *inode,
+				struct dentry *dentry, struct cifs_tcon *tcon,
+				const char *full_path, umode_t mode, dev_t dev)
 {
-	struct cifs_open_info_data buf = {};
 	struct TCP_Server_Info *server = tcon->ses->server;
 	struct cifs_open_parms oparms;
 	struct cifs_io_parms io_parms = {};
 	struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
 	struct cifs_fid fid;
 	unsigned int bytes_written;
-	struct win_dev *pdev;
+	struct win_dev pdev = {};
 	struct kvec iov[2];
 	__u32 oplock = server->oplocks ? REQ_OPLOCK : 0;
 	int rc;
 
-	if (!S_ISCHR(mode) && !S_ISBLK(mode) && !S_ISFIFO(mode))
+	switch (mode & S_IFMT) {
+	case S_IFCHR:
+		strscpy(pdev.type, "IntxCHR");
+		pdev.major = cpu_to_le64(MAJOR(dev));
+		pdev.minor = cpu_to_le64(MINOR(dev));
+		break;
+	case S_IFBLK:
+		strscpy(pdev.type, "IntxBLK");
+		pdev.major = cpu_to_le64(MAJOR(dev));
+		pdev.minor = cpu_to_le64(MINOR(dev));
+		break;
+	case S_IFIFO:
+		strscpy(pdev.type, "LnxFIFO");
+		break;
+	default:
 		return -EPERM;
+	}
 
-	oparms = (struct cifs_open_parms) {
-		.tcon = tcon,
-		.cifs_sb = cifs_sb,
-		.desired_access = GENERIC_WRITE,
-		.create_options = cifs_create_options(cifs_sb, CREATE_NOT_DIR |
-						      CREATE_OPTION_SPECIAL),
-		.disposition = FILE_CREATE,
-		.path = full_path,
-		.fid = &fid,
-	};
+	oparms = CIFS_OPARMS(cifs_sb, tcon, full_path, GENERIC_WRITE,
+			     FILE_CREATE, CREATE_NOT_DIR |
+			     CREATE_OPTION_SPECIAL, ACL_NO_MODE);
+	oparms.fid = &fid;
 
-	rc = server->ops->open(xid, &oparms, &oplock, &buf);
+	rc = server->ops->open(xid, &oparms, &oplock, NULL);
 	if (rc)
 		return rc;
 
-	/*
-	 * BB Do not bother to decode buf since no local inode yet to put
-	 * timestamps in, but we can reuse it safely.
-	 */
-	pdev = (struct win_dev *)&buf.fi;
 	io_parms.pid = current->tgid;
 	io_parms.tcon = tcon;
-	io_parms.length = sizeof(*pdev);
-	iov[1].iov_base = pdev;
-	iov[1].iov_len = sizeof(*pdev);
-	if (S_ISCHR(mode)) {
-		memcpy(pdev->type, "IntxCHR", 8);
-		pdev->major = cpu_to_le64(MAJOR(dev));
-		pdev->minor = cpu_to_le64(MINOR(dev));
-	} else if (S_ISBLK(mode)) {
-		memcpy(pdev->type, "IntxBLK", 8);
-		pdev->major = cpu_to_le64(MAJOR(dev));
-		pdev->minor = cpu_to_le64(MINOR(dev));
-	} else if (S_ISFIFO(mode)) {
-		memcpy(pdev->type, "LnxFIFO", 8);
-	}
+	io_parms.length = sizeof(pdev);
+	iov[1].iov_base = &pdev;
+	iov[1].iov_len = sizeof(pdev);
 
 	rc = server->ops->sync_write(xid, &fid, &io_parms,
 				     &bytes_written, iov, 1);
 	server->ops->close(xid, tcon, &fid);
-	d_drop(dentry);
-	/* FIXME: add code here to set EAs */
-	cifs_free_open_info(&buf);
+	return rc;
+}
+
+int cifs_sfu_make_node(unsigned int xid, struct inode *inode,
+		       struct dentry *dentry, struct cifs_tcon *tcon,
+		       const char *full_path, umode_t mode, dev_t dev)
+{
+	struct inode *new = NULL;
+	int rc;
+
+	rc = __cifs_sfu_make_node(xid, inode, dentry, tcon,
+				  full_path, mode, dev);
+	if (rc)
+		return rc;
+
+	if (tcon->posix_extensions) {
+		rc = smb311_posix_get_inode_info(&new, full_path, NULL,
+						 inode->i_sb, xid);
+	} else if (tcon->unix_ext) {
+		rc = cifs_get_inode_info_unix(&new, full_path,
+					      inode->i_sb, xid);
+	} else {
+		rc = cifs_get_inode_info(&new, full_path, NULL,
+					 inode->i_sb, xid, NULL);
+	}
+	if (!rc)
+		d_instantiate(dentry, new);
 	return rc;
 }
 
diff --git a/fs/smb/client/smb2pdu.c b/fs/smb/client/smb2pdu.c
index c0c4933af5fc..a5efce03cb58 100644
--- a/fs/smb/client/smb2pdu.c
+++ b/fs/smb/client/smb2pdu.c
@@ -367,6 +367,17 @@ again:
 		}
 
 		rc = cifs_setup_session(0, ses, server, nls_codepage);
+		if ((rc == -EACCES) || (rc == -EKEYEXPIRED) || (rc == -EKEYREVOKED)) {
+			/*
+			 * Try alternate password for next reconnect (key rotation
+			 * could be enabled on the server e.g.) if an alternate
+			 * password is available and the current password is expired,
+			 * but do not swap on non pwd related errors like host down
+			 */
+			if (ses->password2)
+				swap(ses->password2, ses->password);
+		}
+
 		if ((rc == -EACCES) && !tcon->retry) {
 			mutex_unlock(&ses->session_mutex);
 			rc = -EHOSTDOWN;
@@ -4127,6 +4138,8 @@ void smb2_reconnect_server(struct work_struct *work)
 		list_for_each_entry(tcon, &ses->tcon_list, tcon_list) {
 			if (tcon->need_reconnect || tcon->need_reopen_files) {
 				tcon->tc_count++;
+				trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+						    netfs_trace_tcon_ref_get_reconnect_server);
 				list_add_tail(&tcon->rlist, &tmp_list);
 				tcon_selected = true;
 			}
@@ -4165,14 +4178,14 @@ void smb2_reconnect_server(struct work_struct *work)
 		if (tcon->ipc)
 			cifs_put_smb_ses(tcon->ses);
 		else
-			cifs_put_tcon(tcon);
+			cifs_put_tcon(tcon, netfs_trace_tcon_ref_put_reconnect_server);
 	}
 
 	if (!ses_exist)
 		goto done;
 
 	/* allocate a dummy tcon struct used for reconnect */
-	tcon = tcon_info_alloc(false);
+	tcon = tcon_info_alloc(false, netfs_trace_tcon_ref_new_reconnect_server);
 	if (!tcon) {
 		resched = true;
 		list_for_each_entry_safe(ses, ses2, &tmp_ses_list, rlist) {
@@ -4195,7 +4208,7 @@ void smb2_reconnect_server(struct work_struct *work)
 		list_del_init(&ses->rlist);
 		cifs_put_smb_ses(ses);
 	}
-	tconInfoFree(tcon);
+	tconInfoFree(tcon, netfs_trace_tcon_ref_free_reconnect_server);
 
 done:
 	cifs_dbg(FYI, "Reconnecting tcons and channels finished\n");
diff --git a/fs/smb/client/smb2pdu.h b/fs/smb/client/smb2pdu.h
index c72a3b2886b7..2fccf0d4f53d 100644
--- a/fs/smb/client/smb2pdu.h
+++ b/fs/smb/client/smb2pdu.h
@@ -320,7 +320,7 @@ struct smb2_file_reparse_point_info {
 } __packed;
 
 struct smb2_file_network_open_info {
-	struct_group(network_open_info,
+	struct_group_attr(network_open_info, __packed,
 		__le64 CreationTime;
 		__le64 LastAccessTime;
 		__le64 LastWriteTime;
diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c
index 1d6e54f7879e..02135a605305 100644
--- a/fs/smb/client/smb2transport.c
+++ b/fs/smb/client/smb2transport.c
@@ -189,6 +189,8 @@ smb2_find_smb_sess_tcon_unlocked(struct cifs_ses *ses, __u32  tid)
 		if (tcon->tid != tid)
 			continue;
 		++tcon->tc_count;
+		trace_smb3_tcon_ref(tcon->debug_id, tcon->tc_count,
+				    netfs_trace_tcon_ref_get_find_sess_tcon);
 		return tcon;
 	}
 
diff --git a/fs/smb/client/trace.h b/fs/smb/client/trace.h
index 5e83cb9da902..604e52876cd2 100644
--- a/fs/smb/client/trace.h
+++ b/fs/smb/client/trace.h
@@ -3,6 +3,9 @@
  *   Copyright (C) 2018, Microsoft Corporation.
  *
  *   Author(s): Steve French <stfrench@microsoft.com>
+ *
+ * Please use this 3-part article as a reference for writing new tracepoints:
+ * https://lwn.net/Articles/379903/
  */
 #undef TRACE_SYSTEM
 #define TRACE_SYSTEM cifs
@@ -15,9 +18,70 @@
 #include <linux/inet.h>
 
 /*
- * Please use this 3-part article as a reference for writing new tracepoints:
- * https://lwn.net/Articles/379903/
+ * Specify enums for tracing information.
+ */
+#define smb3_tcon_ref_traces					      \
+	EM(netfs_trace_tcon_ref_dec_dfs_refer,		"DEC DfsRef") \
+	EM(netfs_trace_tcon_ref_free,			"FRE       ") \
+	EM(netfs_trace_tcon_ref_free_fail,		"FRE Fail  ") \
+	EM(netfs_trace_tcon_ref_free_ipc,		"FRE Ipc   ") \
+	EM(netfs_trace_tcon_ref_free_ipc_fail,		"FRE Ipc-F ") \
+	EM(netfs_trace_tcon_ref_free_reconnect_server,	"FRE Reconn") \
+	EM(netfs_trace_tcon_ref_get_cancelled_close,	"GET Cn-Cls") \
+	EM(netfs_trace_tcon_ref_get_dfs_refer,		"GET DfsRef") \
+	EM(netfs_trace_tcon_ref_get_find,		"GET Find  ") \
+	EM(netfs_trace_tcon_ref_get_find_sess_tcon,	"GET FndSes") \
+	EM(netfs_trace_tcon_ref_get_reconnect_server,	"GET Reconn") \
+	EM(netfs_trace_tcon_ref_new,			"NEW       ") \
+	EM(netfs_trace_tcon_ref_new_ipc,		"NEW Ipc   ") \
+	EM(netfs_trace_tcon_ref_new_reconnect_server,	"NEW Reconn") \
+	EM(netfs_trace_tcon_ref_put_cancelled_close,	"PUT Cn-Cls") \
+	EM(netfs_trace_tcon_ref_put_cancelled_close_fid, "PUT Cn-Fid") \
+	EM(netfs_trace_tcon_ref_put_cancelled_mid,	"PUT Cn-Mid") \
+	EM(netfs_trace_tcon_ref_put_mnt_ctx,		"PUT MntCtx") \
+	EM(netfs_trace_tcon_ref_put_reconnect_server,	"PUT Reconn") \
+	EM(netfs_trace_tcon_ref_put_tlink,		"PUT Tlink ") \
+	EM(netfs_trace_tcon_ref_see_cancelled_close,	"SEE Cn-Cls") \
+	EM(netfs_trace_tcon_ref_see_fscache_collision,	"SEE FV-CO!") \
+	EM(netfs_trace_tcon_ref_see_fscache_okay,	"SEE FV-Ok ") \
+	EM(netfs_trace_tcon_ref_see_fscache_relinq,	"SEE FV-Rlq") \
+	E_(netfs_trace_tcon_ref_see_umount,		"SEE Umount")
+
+#undef EM
+#undef E_
+
+/*
+ * Define those tracing enums.
+ */
+#ifndef __SMB3_DECLARE_TRACE_ENUMS_ONCE_ONLY
+#define __SMB3_DECLARE_TRACE_ENUMS_ONCE_ONLY
+
+#define EM(a, b) a,
+#define E_(a, b) a
+
+enum smb3_tcon_ref_trace { smb3_tcon_ref_traces } __mode(byte);
+
+#undef EM
+#undef E_
+#endif
+
+/*
+ * Export enum symbols via userspace.
+ */
+#define EM(a, b) TRACE_DEFINE_ENUM(a);
+#define E_(a, b) TRACE_DEFINE_ENUM(a);
+
+smb3_tcon_ref_traces;
+
+#undef EM
+#undef E_
+
+/*
+ * Now redefine the EM() and E_() macros to map the enums to the strings that
+ * will be printed in the output.
  */
+#define EM(a, b)	{ a, b },
+#define E_(a, b)	{ a, b }
 
 /* For logging errors in read or write */
 DECLARE_EVENT_CLASS(smb3_rw_err_class,
@@ -1125,6 +1189,30 @@ DEFINE_SMB3_CREDIT_EVENT(waitff_credits);
 DEFINE_SMB3_CREDIT_EVENT(overflow_credits);
 DEFINE_SMB3_CREDIT_EVENT(set_credits);
 
+
+TRACE_EVENT(smb3_tcon_ref,
+	    TP_PROTO(unsigned int tcon_debug_id, int ref,
+		     enum smb3_tcon_ref_trace trace),
+	    TP_ARGS(tcon_debug_id, ref, trace),
+	    TP_STRUCT__entry(
+		    __field(unsigned int,		tcon)
+		    __field(int,			ref)
+		    __field(enum smb3_tcon_ref_trace,	trace)
+			     ),
+	    TP_fast_assign(
+		    __entry->tcon	= tcon_debug_id;
+		    __entry->ref	= ref;
+		    __entry->trace	= trace;
+			   ),
+	    TP_printk("TC=%08x %s r=%u",
+		      __entry->tcon,
+		      __print_symbolic(__entry->trace, smb3_tcon_ref_traces),
+		      __entry->ref)
+	    );
+
+
+#undef EM
+#undef E_
 #endif /* _CIFS_TRACE_H */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/fs/smb/client/transport.c b/fs/smb/client/transport.c
index 994d70193432..ddf1a3aafee5 100644
--- a/fs/smb/client/transport.c
+++ b/fs/smb/client/transport.c
@@ -909,12 +909,15 @@ cifs_sync_mid_result(struct mid_q_entry *mid, struct TCP_Server_Info *server)
 			list_del_init(&mid->qhead);
 			mid->mid_flags |= MID_DELETED;
 		}
+		spin_unlock(&server->mid_lock);
 		cifs_server_dbg(VFS, "%s: invalid mid state mid=%llu state=%d\n",
 			 __func__, mid->mid, mid->mid_state);
 		rc = -EIO;
+		goto sync_mid_done;
 	}
 	spin_unlock(&server->mid_lock);
 
+sync_mid_done:
 	release_mid(mid);
 	return rc;
 }
@@ -1057,9 +1060,11 @@ struct TCP_Server_Info *cifs_pick_channel(struct cifs_ses *ses)
 		index = (uint)atomic_inc_return(&ses->chan_seq);
 		index %= ses->chan_count;
 	}
+
+	server = ses->chans[index].server;
 	spin_unlock(&ses->chan_lock);
 
-	return ses->chans[index].server;
+	return server;
 }
 
 int
diff --git a/fs/smb/common/smb2pdu.h b/fs/smb/common/smb2pdu.h
index 1b594307c9d5..202ff9128156 100644
--- a/fs/smb/common/smb2pdu.h
+++ b/fs/smb/common/smb2pdu.h
@@ -711,7 +711,7 @@ struct smb2_close_rsp {
 	__le16 StructureSize; /* 60 */
 	__le16 Flags;
 	__le32 Reserved;
-	struct_group(network_open_info,
+	struct_group_attr(network_open_info, __packed,
 		__le64 CreationTime;
 		__le64 LastAccessTime;
 		__le64 LastWriteTime;
diff --git a/fs/smb/server/ksmbd_netlink.h b/fs/smb/server/ksmbd_netlink.h
index 686b321c5a8b..f4e55199938d 100644
--- a/fs/smb/server/ksmbd_netlink.h
+++ b/fs/smb/server/ksmbd_netlink.h
@@ -340,23 +340,24 @@ enum KSMBD_TREE_CONN_STATUS {
 /*
  * Share config flags.
  */
-#define KSMBD_SHARE_FLAG_INVALID		(0)
-#define KSMBD_SHARE_FLAG_AVAILABLE		BIT(0)
-#define KSMBD_SHARE_FLAG_BROWSEABLE		BIT(1)
-#define KSMBD_SHARE_FLAG_WRITEABLE		BIT(2)
-#define KSMBD_SHARE_FLAG_READONLY		BIT(3)
-#define KSMBD_SHARE_FLAG_GUEST_OK		BIT(4)
-#define KSMBD_SHARE_FLAG_GUEST_ONLY		BIT(5)
-#define KSMBD_SHARE_FLAG_STORE_DOS_ATTRS	BIT(6)
-#define KSMBD_SHARE_FLAG_OPLOCKS		BIT(7)
-#define KSMBD_SHARE_FLAG_PIPE			BIT(8)
-#define KSMBD_SHARE_FLAG_HIDE_DOT_FILES		BIT(9)
-#define KSMBD_SHARE_FLAG_INHERIT_OWNER		BIT(10)
-#define KSMBD_SHARE_FLAG_STREAMS		BIT(11)
-#define KSMBD_SHARE_FLAG_FOLLOW_SYMLINKS	BIT(12)
-#define KSMBD_SHARE_FLAG_ACL_XATTR		BIT(13)
-#define KSMBD_SHARE_FLAG_UPDATE			BIT(14)
-#define KSMBD_SHARE_FLAG_CROSSMNT		BIT(15)
+#define KSMBD_SHARE_FLAG_INVALID			(0)
+#define KSMBD_SHARE_FLAG_AVAILABLE			BIT(0)
+#define KSMBD_SHARE_FLAG_BROWSEABLE			BIT(1)
+#define KSMBD_SHARE_FLAG_WRITEABLE			BIT(2)
+#define KSMBD_SHARE_FLAG_READONLY			BIT(3)
+#define KSMBD_SHARE_FLAG_GUEST_OK			BIT(4)
+#define KSMBD_SHARE_FLAG_GUEST_ONLY			BIT(5)
+#define KSMBD_SHARE_FLAG_STORE_DOS_ATTRS		BIT(6)
+#define KSMBD_SHARE_FLAG_OPLOCKS			BIT(7)
+#define KSMBD_SHARE_FLAG_PIPE				BIT(8)
+#define KSMBD_SHARE_FLAG_HIDE_DOT_FILES			BIT(9)
+#define KSMBD_SHARE_FLAG_INHERIT_OWNER			BIT(10)
+#define KSMBD_SHARE_FLAG_STREAMS			BIT(11)
+#define KSMBD_SHARE_FLAG_FOLLOW_SYMLINKS		BIT(12)
+#define KSMBD_SHARE_FLAG_ACL_XATTR			BIT(13)
+#define KSMBD_SHARE_FLAG_UPDATE				BIT(14)
+#define KSMBD_SHARE_FLAG_CROSSMNT			BIT(15)
+#define KSMBD_SHARE_FLAG_CONTINUOUS_AVAILABILITY	BIT(16)
 
 /*
  * Tree connect request flags.
diff --git a/fs/smb/server/server.c b/fs/smb/server/server.c
index c0788188aa82..c67fbc8d6683 100644
--- a/fs/smb/server/server.c
+++ b/fs/smb/server/server.c
@@ -167,20 +167,17 @@ static void __handle_ksmbd_work(struct ksmbd_work *work,
 	int rc;
 	bool is_chained = false;
 
-	if (conn->ops->allocate_rsp_buf(work))
-		return;
-
 	if (conn->ops->is_transform_hdr &&
 	    conn->ops->is_transform_hdr(work->request_buf)) {
 		rc = conn->ops->decrypt_req(work);
-		if (rc < 0) {
-			conn->ops->set_rsp_status(work, STATUS_DATA_ERROR);
-			goto send;
-		}
-
+		if (rc < 0)
+			return;
 		work->encrypted = true;
 	}
 
+	if (conn->ops->allocate_rsp_buf(work))
+		return;
+
 	rc = conn->ops->init_rsp_hdr(work);
 	if (rc) {
 		/* either uid or tid is not correct */
diff --git a/fs/smb/server/smb2pdu.c b/fs/smb/server/smb2pdu.c
index 5723bbf372d7..355824151c2d 100644
--- a/fs/smb/server/smb2pdu.c
+++ b/fs/smb/server/smb2pdu.c
@@ -535,6 +535,10 @@ int smb2_allocate_rsp_buf(struct ksmbd_work *work)
 	if (cmd == SMB2_QUERY_INFO_HE) {
 		struct smb2_query_info_req *req;
 
+		if (get_rfc1002_len(work->request_buf) <
+		    offsetof(struct smb2_query_info_req, OutputBufferLength))
+			return -EINVAL;
+
 		req = smb2_get_msg(work->request_buf);
 		if ((req->InfoType == SMB2_O_INFO_FILE &&
 		     (req->FileInfoClass == FILE_FULL_EA_INFORMATION ||
@@ -1984,7 +1988,12 @@ int smb2_tree_connect(struct ksmbd_work *work)
 	write_unlock(&sess->tree_conns_lock);
 	rsp->StructureSize = cpu_to_le16(16);
 out_err1:
-	rsp->Capabilities = 0;
+	if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE &&
+	    test_share_config_flag(share,
+				   KSMBD_SHARE_FLAG_CONTINUOUS_AVAILABILITY))
+		rsp->Capabilities = SMB2_SHARE_CAP_CONTINUOUS_AVAILABILITY;
+	else
+		rsp->Capabilities = 0;
 	rsp->Reserved = 0;
 	/* default manual caching */
 	rsp->ShareFlags = SMB2_SHAREFLAG_MANUAL_CACHING;
@@ -3498,7 +3507,9 @@ int smb2_open(struct ksmbd_work *work)
 	memcpy(fp->client_guid, conn->ClientGUID, SMB2_CLIENT_GUID_SIZE);
 
 	if (dh_info.type == DURABLE_REQ_V2 || dh_info.type == DURABLE_REQ) {
-		if (dh_info.type == DURABLE_REQ_V2 && dh_info.persistent)
+		if (dh_info.type == DURABLE_REQ_V2 && dh_info.persistent &&
+		    test_share_config_flag(work->tcon->share_conf,
+					   KSMBD_SHARE_FLAG_CONTINUOUS_AVAILABILITY))
 			fp->is_persistent = true;
 		else
 			fp->is_durable = true;
diff --git a/fs/smb/server/vfs.c b/fs/smb/server/vfs.c
index 22f0f3db3ac9..51b1b0bed616 100644
--- a/fs/smb/server/vfs.c
+++ b/fs/smb/server/vfs.c
@@ -754,10 +754,15 @@ retry:
 		goto out4;
 	}
 
+	/*
+	 * explicitly handle file overwrite case, for compatibility with
+	 * filesystems that may not support rename flags (e.g: fuse)
+	 */
 	if ((flags & RENAME_NOREPLACE) && d_is_positive(new_dentry)) {
 		err = -EEXIST;
 		goto out4;
 	}
+	flags &= ~(RENAME_NOREPLACE);
 
 	if (old_child == trap) {
 		err = -EINVAL;
diff --git a/fs/squashfs/inode.c b/fs/squashfs/inode.c
index aa3411354e66..16bd693d0b3a 100644
--- a/fs/squashfs/inode.c
+++ b/fs/squashfs/inode.c
@@ -48,6 +48,10 @@ static int squashfs_new_inode(struct super_block *sb, struct inode *inode,
 	gid_t i_gid;
 	int err;
 
+	inode->i_ino = le32_to_cpu(sqsh_ino->inode_number);
+	if (inode->i_ino == 0)
+		return -EINVAL;
+
 	err = squashfs_get_id(sb, le16_to_cpu(sqsh_ino->uid), &i_uid);
 	if (err)
 		return err;
@@ -58,7 +62,6 @@ static int squashfs_new_inode(struct super_block *sb, struct inode *inode,
 
 	i_uid_write(inode, i_uid);
 	i_gid_write(inode, i_gid);
-	inode->i_ino = le32_to_cpu(sqsh_ino->inode_number);
 	inode_set_mtime(inode, le32_to_cpu(sqsh_ino->mtime), 0);
 	inode_set_atime(inode, inode_get_mtime_sec(inode), 0);
 	inode_set_ctime(inode, inode_get_mtime_sec(inode), 0);
diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
index 6b7652fb8050..7cd64021d453 100644
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -463,6 +463,8 @@ struct kernfs_node *sysfs_break_active_protection(struct kobject *kobj,
 	kn = kernfs_find_and_get(kobj->sd, attr->name);
 	if (kn)
 		kernfs_break_active_protection(kn);
+	else
+		kobject_put(kobj);
 	return kn;
 }
 EXPORT_SYMBOL_GPL(sysfs_break_active_protection);
diff --git a/fs/tracefs/event_inode.c b/fs/tracefs/event_inode.c
index dc067eeb6387..a878cea70f4c 100644
--- a/fs/tracefs/event_inode.c
+++ b/fs/tracefs/event_inode.c
@@ -37,6 +37,7 @@ static DEFINE_MUTEX(eventfs_mutex);
 
 struct eventfs_root_inode {
 	struct eventfs_inode		ei;
+	struct inode			*parent_inode;
 	struct dentry			*events_dir;
 };
 
@@ -68,11 +69,25 @@ enum {
 	EVENTFS_SAVE_MODE	= BIT(16),
 	EVENTFS_SAVE_UID	= BIT(17),
 	EVENTFS_SAVE_GID	= BIT(18),
-	EVENTFS_TOPLEVEL	= BIT(19),
 };
 
 #define EVENTFS_MODE_MASK	(EVENTFS_SAVE_MODE - 1)
 
+static void free_ei_rcu(struct rcu_head *rcu)
+{
+	struct eventfs_inode *ei = container_of(rcu, struct eventfs_inode, rcu);
+	struct eventfs_root_inode *rei;
+
+	kfree(ei->entry_attrs);
+	kfree_const(ei->name);
+	if (ei->is_events) {
+		rei = get_root_inode(ei);
+		kfree(rei);
+	} else {
+		kfree(ei);
+	}
+}
+
 /*
  * eventfs_inode reference count management.
  *
@@ -84,18 +99,17 @@ enum {
 static void release_ei(struct kref *ref)
 {
 	struct eventfs_inode *ei = container_of(ref, struct eventfs_inode, kref);
-	struct eventfs_root_inode *rei;
+	const struct eventfs_entry *entry;
 
 	WARN_ON_ONCE(!ei->is_freed);
 
-	kfree(ei->entry_attrs);
-	kfree_const(ei->name);
-	if (ei->is_events) {
-		rei = get_root_inode(ei);
-		kfree_rcu(rei, ei.rcu);
-	} else {
-		kfree_rcu(ei, rcu);
+	for (int i = 0; i < ei->nr_entries; i++) {
+		entry = &ei->entries[i];
+		if (entry->release)
+			entry->release(entry->name, ei->data);
 	}
+
+	call_rcu(&ei->rcu, free_ei_rcu);
 }
 
 static inline void put_ei(struct eventfs_inode *ei)
@@ -112,6 +126,18 @@ static inline void free_ei(struct eventfs_inode *ei)
 	}
 }
 
+/*
+ * Called when creation of an ei fails, do not call release() functions.
+ */
+static inline void cleanup_ei(struct eventfs_inode *ei)
+{
+	if (ei) {
+		/* Set nr_entries to 0 to prevent release() function being called */
+		ei->nr_entries = 0;
+		free_ei(ei);
+	}
+}
+
 static inline struct eventfs_inode *get_ei(struct eventfs_inode *ei)
 {
 	if (ei)
@@ -181,21 +207,7 @@ static int eventfs_set_attr(struct mnt_idmap *idmap, struct dentry *dentry,
 	 * determined by the parent directory.
 	 */
 	if (dentry->d_inode->i_mode & S_IFDIR) {
-		/*
-		 * The events directory dentry is never freed, unless its
-		 * part of an instance that is deleted. It's attr is the
-		 * default for its child files and directories.
-		 * Do not update it. It's not used for its own mode or ownership.
-		 */
-		if (ei->is_events) {
-			/* But it still needs to know if it was modified */
-			if (iattr->ia_valid & ATTR_UID)
-				ei->attr.mode |= EVENTFS_SAVE_UID;
-			if (iattr->ia_valid & ATTR_GID)
-				ei->attr.mode |= EVENTFS_SAVE_GID;
-		} else {
-			update_attr(&ei->attr, iattr);
-		}
+		update_attr(&ei->attr, iattr);
 
 	} else {
 		name = dentry->d_name.name;
@@ -213,18 +225,25 @@ static int eventfs_set_attr(struct mnt_idmap *idmap, struct dentry *dentry,
 	return ret;
 }
 
-static void update_top_events_attr(struct eventfs_inode *ei, struct super_block *sb)
+static void update_events_attr(struct eventfs_inode *ei, struct super_block *sb)
 {
-	struct inode *root;
+	struct eventfs_root_inode *rei;
+	struct inode *parent;
 
-	/* Only update if the "events" was on the top level */
-	if (!ei || !(ei->attr.mode & EVENTFS_TOPLEVEL))
-		return;
+	rei = get_root_inode(ei);
+
+	/* Use the parent inode permissions unless root set its permissions */
+	parent = rei->parent_inode;
 
-	/* Get the tracefs root inode. */
-	root = d_inode(sb->s_root);
-	ei->attr.uid = root->i_uid;
-	ei->attr.gid = root->i_gid;
+	if (rei->ei.attr.mode & EVENTFS_SAVE_UID)
+		ei->attr.uid = rei->ei.attr.uid;
+	else
+		ei->attr.uid = parent->i_uid;
+
+	if (rei->ei.attr.mode & EVENTFS_SAVE_GID)
+		ei->attr.gid = rei->ei.attr.gid;
+	else
+		ei->attr.gid = parent->i_gid;
 }
 
 static void set_top_events_ownership(struct inode *inode)
@@ -233,10 +252,10 @@ static void set_top_events_ownership(struct inode *inode)
 	struct eventfs_inode *ei = ti->private;
 
 	/* The top events directory doesn't get automatically updated */
-	if (!ei || !ei->is_events || !(ei->attr.mode & EVENTFS_TOPLEVEL))
+	if (!ei || !ei->is_events)
 		return;
 
-	update_top_events_attr(ei, inode->i_sb);
+	update_events_attr(ei, inode->i_sb);
 
 	if (!(ei->attr.mode & EVENTFS_SAVE_UID))
 		inode->i_uid = ei->attr.uid;
@@ -265,7 +284,7 @@ static int eventfs_permission(struct mnt_idmap *idmap,
 	return generic_permission(idmap, inode, mask);
 }
 
-static const struct inode_operations eventfs_root_dir_inode_operations = {
+static const struct inode_operations eventfs_dir_inode_operations = {
 	.lookup		= eventfs_root_lookup,
 	.setattr	= eventfs_set_attr,
 	.getattr	= eventfs_get_attr,
@@ -282,6 +301,35 @@ static const struct file_operations eventfs_file_operations = {
 	.llseek		= generic_file_llseek,
 };
 
+/*
+ * On a remount of tracefs, if UID or GID options are set, then
+ * the mount point inode permissions should be used.
+ * Reset the saved permission flags appropriately.
+ */
+void eventfs_remount(struct tracefs_inode *ti, bool update_uid, bool update_gid)
+{
+	struct eventfs_inode *ei = ti->private;
+
+	if (!ei)
+		return;
+
+	if (update_uid)
+		ei->attr.mode &= ~EVENTFS_SAVE_UID;
+
+	if (update_gid)
+		ei->attr.mode &= ~EVENTFS_SAVE_GID;
+
+	if (!ei->entry_attrs)
+		return;
+
+	for (int i = 0; i < ei->nr_entries; i++) {
+		if (update_uid)
+			ei->entry_attrs[i].mode &= ~EVENTFS_SAVE_UID;
+		if (update_gid)
+			ei->entry_attrs[i].mode &= ~EVENTFS_SAVE_GID;
+	}
+}
+
 /* Return the evenfs_inode of the "events" directory */
 static struct eventfs_inode *eventfs_find_events(struct dentry *dentry)
 {
@@ -304,7 +352,7 @@ static struct eventfs_inode *eventfs_find_events(struct dentry *dentry)
 		// Walk upwards until you find the events inode
 	} while (!ei->is_events);
 
-	update_top_events_attr(ei, dentry->d_sb);
+	update_events_attr(ei, dentry->d_sb);
 
 	return ei;
 }
@@ -336,6 +384,7 @@ static void update_inode_attr(struct dentry *dentry, struct inode *inode,
 
 /**
  * lookup_file - look up a file in the tracefs filesystem
+ * @parent_ei: Pointer to the eventfs_inode that represents parent of the file
  * @dentry: the dentry to look up
  * @mode: the permission that the file should have.
  * @attr: saved attributes changed by user
@@ -389,6 +438,7 @@ static struct dentry *lookup_file(struct eventfs_inode *parent_ei,
 /**
  * lookup_dir_entry - look up a dir in the tracefs filesystem
  * @dentry: the directory to look up
+ * @pei: Pointer to the parent eventfs_inode if available
  * @ei: the eventfs_inode that represents the directory to create
  *
  * This function will look up a dentry for a directory represented by
@@ -408,7 +458,7 @@ static struct dentry *lookup_dir_entry(struct dentry *dentry,
 	update_inode_attr(dentry, inode, &ei->attr,
 			  S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO);
 
-	inode->i_op = &eventfs_root_dir_inode_operations;
+	inode->i_op = &eventfs_dir_inode_operations;
 	inode->i_fop = &eventfs_file_operations;
 
 	/* All directories will have the same inode number */
@@ -478,16 +528,20 @@ void eventfs_d_release(struct dentry *dentry)
 
 /**
  * lookup_file_dentry - create a dentry for a file of an eventfs_inode
+ * @dentry: The parent dentry under which the new file's dentry will be created
  * @ei: the eventfs_inode that the file will be created under
  * @idx: the index into the entry_attrs[] of the @ei
- * @parent: The parent dentry of the created file.
- * @name: The name of the file to create
  * @mode: The mode of the file.
  * @data: The data to use to set the inode of the file with on open()
  * @fops: The fops of the file to be created.
  *
- * Create a dentry for a file of an eventfs_inode @ei and place it into the
- * address located at @e_dentry.
+ * This function creates a dentry for a file associated with an
+ * eventfs_inode @ei. It uses the entry attributes specified by @idx,
+ * if available. The file will have the specified @mode and its inode will be
+ * set up with @data upon open. The file operations will be set to @fops.
+ *
+ * Return: Returns a pointer to the newly created file's dentry or an error
+ * pointer.
  */
 static struct dentry *
 lookup_file_dentry(struct dentry *dentry,
@@ -728,7 +782,7 @@ struct eventfs_inode *eventfs_create_dir(const char *name, struct eventfs_inode
 
 	/* Was the parent freed? */
 	if (list_empty(&ei->list)) {
-		free_ei(ei);
+		cleanup_ei(ei);
 		ei = NULL;
 	}
 	return ei;
@@ -775,6 +829,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	// Note: we have a ref to the dentry from tracefs_start_creating()
 	rei = get_root_inode(ei);
 	rei->events_dir = dentry;
+	rei->parent_inode = d_inode(dentry->d_sb->s_root);
 
 	ei->entries = entries;
 	ei->nr_entries = size;
@@ -784,29 +839,26 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	uid = d_inode(dentry->d_parent)->i_uid;
 	gid = d_inode(dentry->d_parent)->i_gid;
 
-	/*
-	 * If the events directory is of the top instance, then parent
-	 * is NULL. Set the attr.mode to reflect this and its permissions will
-	 * default to the tracefs root dentry.
-	 */
-	if (!parent)
-		ei->attr.mode = EVENTFS_TOPLEVEL;
-
-	/* This is used as the default ownership of the files and directories */
 	ei->attr.uid = uid;
 	ei->attr.gid = gid;
 
+	/*
+	 * When the "events" directory is created, it takes on the
+	 * permissions of its parent. But can be reset on remount.
+	 */
+	ei->attr.mode |= EVENTFS_SAVE_UID | EVENTFS_SAVE_GID;
+
 	INIT_LIST_HEAD(&ei->children);
 	INIT_LIST_HEAD(&ei->list);
 
 	ti = get_tracefs(inode);
-	ti->flags |= TRACEFS_EVENT_INODE | TRACEFS_EVENT_TOP_INODE;
+	ti->flags |= TRACEFS_EVENT_INODE;
 	ti->private = ei;
 
 	inode->i_mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
 	inode->i_uid = uid;
 	inode->i_gid = gid;
-	inode->i_op = &eventfs_root_dir_inode_operations;
+	inode->i_op = &eventfs_dir_inode_operations;
 	inode->i_fop = &eventfs_file_operations;
 
 	dentry->d_fsdata = get_ei(ei);
@@ -829,7 +881,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry
 	return ei;
 
  fail:
-	free_ei(ei);
+	cleanup_ei(ei);
 	tracefs_failed_creating(dentry);
 	return ERR_PTR(-ENOMEM);
 }
diff --git a/fs/tracefs/inode.c b/fs/tracefs/inode.c
index 5545e6bf7d26..417c840e6403 100644
--- a/fs/tracefs/inode.c
+++ b/fs/tracefs/inode.c
@@ -30,20 +30,47 @@ static struct vfsmount *tracefs_mount;
 static int tracefs_mount_count;
 static bool tracefs_registered;
 
+/*
+ * Keep track of all tracefs_inodes in order to update their
+ * flags if necessary on a remount.
+ */
+static DEFINE_SPINLOCK(tracefs_inode_lock);
+static LIST_HEAD(tracefs_inodes);
+
 static struct inode *tracefs_alloc_inode(struct super_block *sb)
 {
 	struct tracefs_inode *ti;
+	unsigned long flags;
 
 	ti = kmem_cache_alloc(tracefs_inode_cachep, GFP_KERNEL);
 	if (!ti)
 		return NULL;
 
+	spin_lock_irqsave(&tracefs_inode_lock, flags);
+	list_add_rcu(&ti->list, &tracefs_inodes);
+	spin_unlock_irqrestore(&tracefs_inode_lock, flags);
+
 	return &ti->vfs_inode;
 }
 
+static void tracefs_free_inode_rcu(struct rcu_head *rcu)
+{
+	struct tracefs_inode *ti;
+
+	ti = container_of(rcu, struct tracefs_inode, rcu);
+	kmem_cache_free(tracefs_inode_cachep, ti);
+}
+
 static void tracefs_free_inode(struct inode *inode)
 {
-	kmem_cache_free(tracefs_inode_cachep, get_tracefs(inode));
+	struct tracefs_inode *ti = get_tracefs(inode);
+	unsigned long flags;
+
+	spin_lock_irqsave(&tracefs_inode_lock, flags);
+	list_del_rcu(&ti->list);
+	spin_unlock_irqrestore(&tracefs_inode_lock, flags);
+
+	call_rcu(&ti->rcu, tracefs_free_inode_rcu);
 }
 
 static ssize_t default_read_file(struct file *file, char __user *buf,
@@ -153,16 +180,39 @@ static void set_tracefs_inode_owner(struct inode *inode)
 {
 	struct tracefs_inode *ti = get_tracefs(inode);
 	struct inode *root_inode = ti->private;
+	kuid_t uid;
+	kgid_t gid;
+
+	uid = root_inode->i_uid;
+	gid = root_inode->i_gid;
+
+	/*
+	 * If the root is not the mount point, then check the root's
+	 * permissions. If it was never set, then default to the
+	 * mount point.
+	 */
+	if (root_inode != d_inode(root_inode->i_sb->s_root)) {
+		struct tracefs_inode *rti;
+
+		rti = get_tracefs(root_inode);
+		root_inode = d_inode(root_inode->i_sb->s_root);
+
+		if (!(rti->flags & TRACEFS_UID_PERM_SET))
+			uid = root_inode->i_uid;
+
+		if (!(rti->flags & TRACEFS_GID_PERM_SET))
+			gid = root_inode->i_gid;
+	}
 
 	/*
 	 * If this inode has never been referenced, then update
 	 * the permissions to the superblock.
 	 */
 	if (!(ti->flags & TRACEFS_UID_PERM_SET))
-		inode->i_uid = root_inode->i_uid;
+		inode->i_uid = uid;
 
 	if (!(ti->flags & TRACEFS_GID_PERM_SET))
-		inode->i_gid = root_inode->i_gid;
+		inode->i_gid = gid;
 }
 
 static int tracefs_permission(struct mnt_idmap *idmap,
@@ -313,6 +363,8 @@ static int tracefs_apply_options(struct super_block *sb, bool remount)
 	struct tracefs_fs_info *fsi = sb->s_fs_info;
 	struct inode *inode = d_inode(sb->s_root);
 	struct tracefs_mount_opts *opts = &fsi->mount_opts;
+	struct tracefs_inode *ti;
+	bool update_uid, update_gid;
 	umode_t tmp_mode;
 
 	/*
@@ -332,6 +384,25 @@ static int tracefs_apply_options(struct super_block *sb, bool remount)
 	if (!remount || opts->opts & BIT(Opt_gid))
 		inode->i_gid = opts->gid;
 
+	if (remount && (opts->opts & BIT(Opt_uid) || opts->opts & BIT(Opt_gid))) {
+
+		update_uid = opts->opts & BIT(Opt_uid);
+		update_gid = opts->opts & BIT(Opt_gid);
+
+		rcu_read_lock();
+		list_for_each_entry_rcu(ti, &tracefs_inodes, list) {
+			if (update_uid)
+				ti->flags &= ~TRACEFS_UID_PERM_SET;
+
+			if (update_gid)
+				ti->flags &= ~TRACEFS_GID_PERM_SET;
+
+			if (ti->flags & TRACEFS_EVENT_INODE)
+				eventfs_remount(ti, update_uid, update_gid);
+		}
+		rcu_read_unlock();
+	}
+
 	return 0;
 }
 
@@ -398,7 +469,22 @@ static int tracefs_d_revalidate(struct dentry *dentry, unsigned int flags)
 	return !(ei && ei->is_freed);
 }
 
+static void tracefs_d_iput(struct dentry *dentry, struct inode *inode)
+{
+	struct tracefs_inode *ti = get_tracefs(inode);
+
+	/*
+	 * This inode is being freed and cannot be used for
+	 * eventfs. Clear the flag so that it doesn't call into
+	 * eventfs during the remount flag updates. The eventfs_inode
+	 * gets freed after an RCU cycle, so the content will still
+	 * be safe if the iteration is going on now.
+	 */
+	ti->flags &= ~TRACEFS_EVENT_INODE;
+}
+
 static const struct dentry_operations tracefs_dentry_operations = {
+	.d_iput = tracefs_d_iput,
 	.d_revalidate = tracefs_d_revalidate,
 	.d_release = tracefs_d_release,
 };
diff --git a/fs/tracefs/internal.h b/fs/tracefs/internal.h
index 15c26f9aaad4..f704d8348357 100644
--- a/fs/tracefs/internal.h
+++ b/fs/tracefs/internal.h
@@ -4,15 +4,18 @@
 
 enum {
 	TRACEFS_EVENT_INODE		= BIT(1),
-	TRACEFS_EVENT_TOP_INODE		= BIT(2),
-	TRACEFS_GID_PERM_SET		= BIT(3),
-	TRACEFS_UID_PERM_SET		= BIT(4),
-	TRACEFS_INSTANCE_INODE		= BIT(5),
+	TRACEFS_GID_PERM_SET		= BIT(2),
+	TRACEFS_UID_PERM_SET		= BIT(3),
+	TRACEFS_INSTANCE_INODE		= BIT(4),
 };
 
 struct tracefs_inode {
-	struct inode            vfs_inode;
+	union {
+		struct inode            vfs_inode;
+		struct rcu_head		rcu;
+	};
 	/* The below gets initialized with memset_after(ti, 0, vfs_inode) */
+	struct list_head	list;
 	unsigned long           flags;
 	void                    *private;
 };
@@ -73,6 +76,7 @@ struct dentry *tracefs_end_creating(struct dentry *dentry);
 struct dentry *tracefs_failed_creating(struct dentry *dentry);
 struct inode *tracefs_get_inode(struct super_block *sb);
 
+void eventfs_remount(struct tracefs_inode *ti, bool update_uid, bool update_gid);
 void eventfs_d_release(struct dentry *dentry);
 
 #endif /* _TRACEFS_INTERNAL_H */
diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index c6a124e8d565..964fa7f24003 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -1048,7 +1048,7 @@ static int zonefs_init_zgroup(struct super_block *sb,
 	zonefs_info(sb, "Zone group \"%s\" has %u file%s\n",
 		    zonefs_zgroup_name(ztype),
 		    zgroup->g_nr_zones,
-		    zgroup->g_nr_zones > 1 ? "s" : "");
+		    str_plural(zgroup->g_nr_zones));
 
 	return 0;
 }
author	Paolo Bonzini <pbonzini@redhat.com>	2024-05-10 10:49:02 -0400
committer	Paolo Bonzini <pbonzini@redhat.com>	2024-05-10 13:20:18 -0400
commit	4232da23d75d173195c6766729e51947b64f83cd (patch)
tree	7eb1420b118d452c0b497bf6267ecdfb0a6d5931 /fs
parent	Merge branch 'kvm-sev-es-ghcbv2' into HEAD (diff)
parent	LoongArch: KVM: Add mmio trace events support (diff)
download	wireguard-linux-4232da23d75d173195c6766729e51947b64f83cd.tar.xz wireguard-linux-4232da23d75d173195c6766729e51947b64f83cd.zip