X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libbcachefs%2Fbcachefs_format.h;h=8312018e1ed5a5199709446dde6375ec7bef8455;hb=790ca9522a96efe321aae36fb0d7f4e437110b0f;hp=a78988e3ded7347df61c3d13c074438339d15e40;hpb=8bcd38555cb224b7da722bcdeca5c2d15f3ef284;p=bcachefs-tools-debian

diff --git a/libbcachefs/bcachefs_format.h b/libbcachefs/bcachefs_format.h
index a78988e..8312018 100644
--- a/libbcachefs/bcachefs_format.h
+++ b/libbcachefs/bcachefs_format.h
@@ -76,6 +76,22 @@
 #include <asm/byteorder.h>
 #include <linux/kernel.h>
 #include <linux/uuid.h>
+#include "vstructs.h"
+
+#define BITMASK(name, type, field, offset, end)				\
+static const unsigned	name##_OFFSET = offset;				\
+static const unsigned	name##_BITS = (end - offset);			\
+									\
+static inline __u64 name(const type *k)					\
+{									\
+	return (k->field >> offset) & ~(~0ULL << (end - offset));	\
+}									\
+									\
+static inline void SET_##name(type *k, __u64 v)				\
+{									\
+	k->field &= ~(~(~0ULL << (end - offset)) << offset);		\
+	k->field |= (v & ~(~0ULL << (end - offset))) << offset;		\
+}
 
 #define LE_BITMASK(_bits, name, type, field, offset, end)		\
 static const unsigned	name##_OFFSET = offset;				\
@@ -138,19 +154,19 @@ struct bpos {
 #define KEY_SNAPSHOT_MAX		((__u32)~0U)
 #define KEY_SIZE_MAX			((__u32)~0U)
 
-static inline struct bpos POS(__u64 inode, __u64 offset)
+static inline struct bpos SPOS(__u64 inode, __u64 offset, __u32 snapshot)
 {
-	struct bpos ret;
-
-	ret.inode	= inode;
-	ret.offset	= offset;
-	ret.snapshot	= 0;
-
-	return ret;
+	return (struct bpos) {
+		.inode		= inode,
+		.offset		= offset,
+		.snapshot	= snapshot,
+	};
 }
 
-#define POS_MIN				POS(0, 0)
-#define POS_MAX				POS(KEY_INODE_MAX, KEY_OFFSET_MAX)
+#define POS_MIN				SPOS(0, 0, 0)
+#define POS_MAX				SPOS(KEY_INODE_MAX, KEY_OFFSET_MAX, 0)
+#define SPOS_MAX			SPOS(KEY_INODE_MAX, KEY_OFFSET_MAX, KEY_SNAPSHOT_MAX)
+#define POS(_inode, _offset)		SPOS(_inode, _offset, 0)
 
 /* Empty placeholder struct, for container_of() */
 struct bch_val {
@@ -323,10 +339,10 @@ static inline void bkey_init(struct bkey *k)
 */
 #define BCH_BKEY_TYPES()				\
 	x(deleted,		0)			\
-	x(discard,		1)			\
+	x(whiteout,		1)			\
 	x(error,		2)			\
 	x(cookie,		3)			\
-	x(whiteout,		4)			\
+	x(hash_whiteout,	4)			\
 	x(btree_ptr,		5)			\
 	x(extent,		6)			\
 	x(reservation,		7)			\
@@ -340,7 +356,16 @@ static inline void bkey_init(struct bkey *k)
 	x(reflink_p,		15)			\
 	x(reflink_v,		16)			\
 	x(inline_data,		17)			\
-	x(btree_ptr_v2,		18)
+	x(btree_ptr_v2,		18)			\
+	x(indirect_inline_data,	19)			\
+	x(alloc_v2,		20)			\
+	x(subvolume,		21)			\
+	x(snapshot,		22)			\
+	x(inode_v2,		23)			\
+	x(alloc_v3,		24)			\
+	x(set,			25)			\
+	x(lru,			26)			\
+	x(alloc_v4,		27)
 
 enum bch_bkey_type {
 #define x(name, nr) KEY_TYPE_##name	= nr,
@@ -349,11 +374,31 @@ enum bch_bkey_type {
 	KEY_TYPE_MAX,
 };
 
+struct bch_deleted {
+	struct bch_val		v;
+};
+
+struct bch_whiteout {
+	struct bch_val		v;
+};
+
+struct bch_error {
+	struct bch_val		v;
+};
+
 struct bch_cookie {
 	struct bch_val		v;
 	__le64			cookie;
 };
 
+struct bch_hash_whiteout {
+	struct bch_val		v;
+};
+
+struct bch_set {
+	struct bch_val		v;
+};
+
 /* Extents */
 
 /*
@@ -550,9 +595,11 @@ struct bch_extent_stripe_ptr {
 #if defined(__LITTLE_ENDIAN_BITFIELD)
 	__u64			type:5,
 				block:8,
-				idx:51;
+				redundancy:4,
+				idx:47;
 #elif defined (__BIG_ENDIAN_BITFIELD)
-	__u64			idx:51,
+	__u64			idx:47,
+				redundancy:4,
 				block:8,
 				type:5;
 #endif
@@ -602,13 +649,14 @@ struct bch_btree_ptr_v2 {
 	__u64			mem_ptr;
 	__le64			seq;
 	__le16			sectors_written;
-	/* In case we ever decide to do variable size btree nodes: */
-	__le16			sectors;
+	__le16			flags;
 	struct bpos		min_key;
 	struct bch_extent_ptr	start[0];
 	__u64			_data[0];
 } __attribute__((packed, aligned(8)));
 
+LE16_BITMASK(BTREE_PTR_RANGE_UPDATED,	struct bch_btree_ptr_v2, flags, 0, 1);
+
 struct bch_extent {
 	struct bch_val		v;
 
@@ -633,8 +681,6 @@ struct bch_reservation {
 #define BKEY_EXTENT_VAL_U64s_MAX				\
 	(1 + BKEY_EXTENT_PTR_U64s_MAX * (BCH_REPLICAS_MAX + 1))
 
-#define BKEY_PADDED(key)	__BKEY_PADDED(key, BKEY_EXTENT_VAL_U64s_MAX)
-
 /* * Maximum possible size of an entire extent, key + value: */
 #define BKEY_EXTENT_U64s_MAX		(BKEY_U64s + BKEY_EXTENT_VAL_U64s_MAX)
 
@@ -660,6 +706,16 @@ struct bch_inode {
 	__u8			fields[0];
 } __attribute__((packed, aligned(8)));
 
+struct bch_inode_v2 {
+	struct bch_val		v;
+
+	__le64			bi_journal_seq;
+	__le64			bi_hash_seed;
+	__le64			bi_flags;
+	__le16			bi_mode;
+	__u8			fields[0];
+} __attribute__((packed, aligned(8)));
+
 struct bch_inode_generation {
 	struct bch_val		v;
 
@@ -667,11 +723,15 @@ struct bch_inode_generation {
 	__le32			pad;
 } __attribute__((packed, aligned(8)));
 
+/*
+ * bi_subvol and bi_parent_subvol are only set for subvolume roots:
+ */
+
 #define BCH_INODE_FIELDS()			\
-	x(bi_atime,			64)	\
-	x(bi_ctime,			64)	\
-	x(bi_mtime,			64)	\
-	x(bi_otime,			64)	\
+	x(bi_atime,			96)	\
+	x(bi_ctime,			96)	\
+	x(bi_mtime,			96)	\
+	x(bi_otime,			96)	\
 	x(bi_size,			64)	\
 	x(bi_sectors,			64)	\
 	x(bi_uid,			32)	\
@@ -688,7 +748,11 @@ struct bch_inode_generation {
 	x(bi_foreground_target,		16)	\
 	x(bi_background_target,		16)	\
 	x(bi_erasure_code,		16)	\
-	x(bi_fields_set,		16)
+	x(bi_fields_set,		16)	\
+	x(bi_dir,			64)	\
+	x(bi_dir_offset,		64)	\
+	x(bi_subvol,			32)	\
+	x(bi_parent_subvol,		32)
 
 /* subset of BCH_INODE_FIELDS */
 #define BCH_INODE_OPTS()			\
@@ -724,6 +788,7 @@ enum {
 	__BCH_INODE_I_SIZE_DIRTY= 5,
 	__BCH_INODE_I_SECTORS_DIRTY= 6,
 	__BCH_INODE_UNLINKED	= 7,
+	__BCH_INODE_BACKPTR_UNTRUSTED = 8,
 
 	/* bits 20+ reserved for packed fields below: */
 };
@@ -736,9 +801,14 @@ enum {
 #define BCH_INODE_I_SIZE_DIRTY	(1 << __BCH_INODE_I_SIZE_DIRTY)
 #define BCH_INODE_I_SECTORS_DIRTY (1 << __BCH_INODE_I_SECTORS_DIRTY)
 #define BCH_INODE_UNLINKED	(1 << __BCH_INODE_UNLINKED)
+#define BCH_INODE_BACKPTR_UNTRUSTED (1 << __BCH_INODE_BACKPTR_UNTRUSTED)
 
 LE32_BITMASK(INODE_STR_HASH,	struct bch_inode, bi_flags, 20, 24);
-LE32_BITMASK(INODE_NR_FIELDS,	struct bch_inode, bi_flags, 24, 32);
+LE32_BITMASK(INODE_NR_FIELDS,	struct bch_inode, bi_flags, 24, 31);
+LE32_BITMASK(INODE_NEW_VARINT,	struct bch_inode, bi_flags, 31, 32);
+
+LE64_BITMASK(INODEv2_STR_HASH,	struct bch_inode_v2, bi_flags, 20, 24);
+LE64_BITMASK(INODEv2_NR_FIELDS,	struct bch_inode_v2, bi_flags, 24, 31);
 
 /* Dirents */
 
@@ -757,7 +827,13 @@ struct bch_dirent {
 	struct bch_val		v;
 
 	/* Target inode number: */
+	union {
 	__le64			d_inum;
+	struct {		/* DT_SUBVOL */
+	__le32			d_child_subvol;
+	__le32			d_parent_subvol;
+	};
+	};
 
 	/*
 	 * Copy of mode bits 12-15 from the target inode - so userspace can get
@@ -768,6 +844,9 @@ struct bch_dirent {
 	__u8			d_name[];
 } __attribute__((packed, aligned(8)));
 
+#define DT_SUBVOL	16
+#define BCH_DT_MAX	17
+
 #define BCH_NAME_MAX	(U8_MAX * sizeof(u64) -				\
 			 sizeof(struct bkey) -				\
 			 offsetof(struct bch_dirent, d_name))
@@ -798,34 +877,73 @@ struct bch_alloc {
 	__u8			data[];
 } __attribute__((packed, aligned(8)));
 
-#define BCH_ALLOC_FIELDS()			\
+#define BCH_ALLOC_FIELDS_V1()			\
 	x(read_time,		16)		\
 	x(write_time,		16)		\
 	x(data_type,		8)		\
 	x(dirty_sectors,	16)		\
 	x(cached_sectors,	16)		\
-	x(oldest_gen,		8)
+	x(oldest_gen,		8)		\
+	x(stripe,		32)		\
+	x(stripe_redundancy,	8)
 
-enum {
-#define x(name, bytes) BCH_ALLOC_FIELD_##name,
-	BCH_ALLOC_FIELDS()
-#undef x
-	BCH_ALLOC_FIELD_NR
-};
+struct bch_alloc_v2 {
+	struct bch_val		v;
+	__u8			nr_fields;
+	__u8			gen;
+	__u8			oldest_gen;
+	__u8			data_type;
+	__u8			data[];
+} __attribute__((packed, aligned(8)));
 
-static const unsigned BCH_ALLOC_FIELD_BYTES[] = {
-#define x(name, bits) [BCH_ALLOC_FIELD_##name] = bits / 8,
-	BCH_ALLOC_FIELDS()
-#undef x
-};
+#define BCH_ALLOC_FIELDS_V2()			\
+	x(read_time,		64)		\
+	x(write_time,		64)		\
+	x(dirty_sectors,	32)		\
+	x(cached_sectors,	32)		\
+	x(stripe,		32)		\
+	x(stripe_redundancy,	8)
 
-#define x(name, bits) + (bits / 8)
-static const unsigned BKEY_ALLOC_VAL_U64s_MAX =
-	DIV_ROUND_UP(offsetof(struct bch_alloc, data)
-		     BCH_ALLOC_FIELDS(), sizeof(u64));
-#undef x
+struct bch_alloc_v3 {
+	struct bch_val		v;
+	__le64			journal_seq;
+	__le32			flags;
+	__u8			nr_fields;
+	__u8			gen;
+	__u8			oldest_gen;
+	__u8			data_type;
+	__u8			data[];
+} __attribute__((packed, aligned(8)));
+
+struct bch_alloc_v4 {
+	struct bch_val		v;
+	__u64			journal_seq;
+	__u32			flags;
+	__u8			gen;
+	__u8			oldest_gen;
+	__u8			data_type;
+	__u8			stripe_redundancy;
+	__u32			dirty_sectors;
+	__u32			cached_sectors;
+	__u64			io_time[2];
+	__u32			stripe;
+	__u32			nr_external_backpointers;
+	struct bpos		backpointers[0];
+} __attribute__((packed, aligned(8)));
 
-#define BKEY_ALLOC_U64s_MAX	(BKEY_U64s + BKEY_ALLOC_VAL_U64s_MAX)
+LE32_BITMASK(BCH_ALLOC_V3_NEED_DISCARD,struct bch_alloc_v3, flags,  0,  1)
+LE32_BITMASK(BCH_ALLOC_V3_NEED_INC_GEN,struct bch_alloc_v3, flags,  1,  2)
+
+BITMASK(BCH_ALLOC_V4_NEED_DISCARD,	struct bch_alloc_v4, flags,  0,  1)
+BITMASK(BCH_ALLOC_V4_NEED_INC_GEN,	struct bch_alloc_v4, flags,  1,  2)
+BITMASK(BCH_ALLOC_V4_BACKPOINTERS_START,struct bch_alloc_v4, flags,  2,  8)
+BITMASK(BCH_ALLOC_V4_NR_BACKPOINTERS,	struct bch_alloc_v4, flags,  8,  14)
+
+enum {
+#define x(name, _bits) BCH_ALLOC_FIELD_V1_##name,
+	BCH_ALLOC_FIELDS_V1()
+#undef x
+};
 
 /* Quotas: */
 
@@ -873,17 +991,29 @@ struct bch_stripe {
 struct bch_reflink_p {
 	struct bch_val		v;
 	__le64			idx;
-
-	__le32			reservation_generation;
-	__u8			nr_replicas;
-	__u8			pad[3];
-};
+	/*
+	 * A reflink pointer might point to an indirect extent which is then
+	 * later split (by copygc or rebalance). If we only pointed to part of
+	 * the original indirect extent, and then one of the fragments is
+	 * outside the range we point to, we'd leak a refcount: so when creating
+	 * reflink pointers, we need to store pad values to remember the full
+	 * range we were taking a reference on.
+	 */
+	__le32			front_pad;
+	__le32			back_pad;
+} __attribute__((packed, aligned(8)));
 
 struct bch_reflink_v {
 	struct bch_val		v;
 	__le64			refcount;
 	union bch_extent_entry	start[0];
 	__u64			_data[0];
+} __attribute__((packed, aligned(8)));
+
+struct bch_indirect_inline_data {
+	struct bch_val		v;
+	__le64			refcount;
+	u8			data[0];
 };
 
 /* Inline data */
@@ -893,6 +1023,52 @@ struct bch_inline_data {
 	u8			data[0];
 };
 
+/* Subvolumes: */
+
+#define SUBVOL_POS_MIN		POS(0, 1)
+#define SUBVOL_POS_MAX		POS(0, S32_MAX)
+#define BCACHEFS_ROOT_SUBVOL	1
+
+struct bch_subvolume {
+	struct bch_val		v;
+	__le32			flags;
+	__le32			snapshot;
+	__le64			inode;
+};
+
+LE32_BITMASK(BCH_SUBVOLUME_RO,		struct bch_subvolume, flags,  0,  1)
+/*
+ * We need to know whether a subvolume is a snapshot so we can know whether we
+ * can delete it (or whether it should just be rm -rf'd)
+ */
+LE32_BITMASK(BCH_SUBVOLUME_SNAP,	struct bch_subvolume, flags,  1,  2)
+LE32_BITMASK(BCH_SUBVOLUME_UNLINKED,	struct bch_subvolume, flags,  2,  3)
+
+/* Snapshots */
+
+struct bch_snapshot {
+	struct bch_val		v;
+	__le32			flags;
+	__le32			parent;
+	__le32			children[2];
+	__le32			subvol;
+	__le32			pad;
+};
+
+LE32_BITMASK(BCH_SNAPSHOT_DELETED,	struct bch_snapshot, flags,  0,  1)
+
+/* True if a subvolume points to this snapshot node: */
+LE32_BITMASK(BCH_SNAPSHOT_SUBVOL,	struct bch_snapshot, flags,  1,  2)
+
+/* LRU btree: */
+
+struct bch_lru {
+	struct bch_val		v;
+	__le64			idx;
+} __attribute__((packed, aligned(8)));
+
+#define LRU_ID_STRIPES		(1U << 16)
+
 /* Optional/variable size superblock sections: */
 
 struct bch_sb_field {
@@ -901,16 +1077,17 @@ struct bch_sb_field {
 	__le32			type;
 };
 
-#define BCH_SB_FIELDS()		\
-	x(journal,	0)	\
-	x(members,	1)	\
-	x(crypt,	2)	\
-	x(replicas_v0,	3)	\
-	x(quota,	4)	\
-	x(disk_groups,	5)	\
-	x(clean,	6)	\
-	x(replicas,	7)	\
-	x(journal_seq_blacklist, 8)
+#define BCH_SB_FIELDS()				\
+	x(journal,	0)			\
+	x(members,	1)			\
+	x(crypt,	2)			\
+	x(replicas_v0,	3)			\
+	x(quota,	4)			\
+	x(disk_groups,	5)			\
+	x(clean,	6)			\
+	x(replicas,	7)			\
+	x(journal_seq_blacklist, 8)		\
+	x(journal_v2,	9)
 
 enum bch_sb_field_type {
 #define x(f, nr)	BCH_SB_FIELD_##f = nr,
@@ -919,6 +1096,14 @@ enum bch_sb_field_type {
 	BCH_SB_FIELD_NR
 };
 
+/*
+ * Most superblock fields are replicated in all device's superblocks - a few are
+ * not:
+ */
+#define BCH_SINGLE_DEVICE_SB_FIELDS		\
+	((1U << BCH_SB_FIELD_journal)|		\
+	 (1U << BCH_SB_FIELD_journal_v2))
+
 /* BCH_SB_FIELD_journal: */
 
 struct bch_sb_field_journal {
@@ -926,6 +1111,15 @@ struct bch_sb_field_journal {
 	__le64			buckets[0];
 };
 
+struct bch_sb_field_journal_v2 {
+	struct bch_sb_field	field;
+
+	struct bch_sb_field_journal_v2_entry {
+		__le64		start;
+		__le64		nr;
+	}			d[0];
+};
+
 /* BCH_SB_FIELD_members: */
 
 #define BCH_MIN_NR_NBUCKETS	(1 << 6)
@@ -942,33 +1136,30 @@ struct bch_member {
 };
 
 LE64_BITMASK(BCH_MEMBER_STATE,		struct bch_member, flags[0],  0,  4)
-/* 4-10 unused, was TIER, HAS_(META)DATA */
-LE64_BITMASK(BCH_MEMBER_REPLACEMENT,	struct bch_member, flags[0], 10, 14)
+/* 4-14 unused, was TIER, HAS_(META)DATA, REPLACEMENT */
 LE64_BITMASK(BCH_MEMBER_DISCARD,	struct bch_member, flags[0], 14, 15)
 LE64_BITMASK(BCH_MEMBER_DATA_ALLOWED,	struct bch_member, flags[0], 15, 20)
 LE64_BITMASK(BCH_MEMBER_GROUP,		struct bch_member, flags[0], 20, 28)
 LE64_BITMASK(BCH_MEMBER_DURABILITY,	struct bch_member, flags[0], 28, 30)
-
-#define BCH_TIER_MAX			4U
+LE64_BITMASK(BCH_MEMBER_FREESPACE_INITIALIZED,
+					struct bch_member, flags[0], 30, 31)
 
 #if 0
 LE64_BITMASK(BCH_MEMBER_NR_READ_ERRORS,	struct bch_member, flags[1], 0,  20);
 LE64_BITMASK(BCH_MEMBER_NR_WRITE_ERRORS,struct bch_member, flags[1], 20, 40);
 #endif
 
-enum bch_member_state {
-	BCH_MEMBER_STATE_RW		= 0,
-	BCH_MEMBER_STATE_RO		= 1,
-	BCH_MEMBER_STATE_FAILED		= 2,
-	BCH_MEMBER_STATE_SPARE		= 3,
-	BCH_MEMBER_STATE_NR		= 4,
-};
+#define BCH_MEMBER_STATES()			\
+	x(rw,		0)			\
+	x(ro,		1)			\
+	x(failed,	2)			\
+	x(spare,	3)
 
-enum cache_replacement {
-	CACHE_REPLACEMENT_LRU		= 0,
-	CACHE_REPLACEMENT_FIFO		= 1,
-	CACHE_REPLACEMENT_RANDOM	= 2,
-	CACHE_REPLACEMENT_NR		= 3,
+enum bch_member_state {
+#define x(t, n) BCH_MEMBER_STATE_##t = n,
+	BCH_MEMBER_STATES()
+#undef x
+	BCH_MEMBER_STATE_NR
 };
 
 struct bch_sb_field_members {
@@ -1026,14 +1217,20 @@ LE64_BITMASK(BCH_KDF_SCRYPT_P,	struct bch_sb_field_crypt, kdf_flags, 32, 48);
 
 /* BCH_SB_FIELD_replicas: */
 
+#define BCH_DATA_TYPES()		\
+	x(none,		0)		\
+	x(sb,		1)		\
+	x(journal,	2)		\
+	x(btree,	3)		\
+	x(user,		4)		\
+	x(cached,	5)		\
+	x(parity,	6)
+
 enum bch_data_type {
-	BCH_DATA_NONE		= 0,
-	BCH_DATA_SB		= 1,
-	BCH_DATA_JOURNAL	= 2,
-	BCH_DATA_BTREE		= 3,
-	BCH_DATA_USER		= 4,
-	BCH_DATA_CACHED		= 5,
-	BCH_DATA_NR		= 6,
+#define x(t, n) BCH_DATA_##t,
+	BCH_DATA_TYPES()
+#undef x
+	BCH_DATA_NR
 };
 
 struct bch_replicas_entry_v0 {
@@ -1118,8 +1315,8 @@ struct bch_sb_field_clean {
 	struct bch_sb_field	field;
 
 	__le32			flags;
-	__le16			read_clock;
-	__le16			write_clock;
+	__le16			_read_clock; /* no longer used */
+	__le16			_write_clock;
 	__le64			journal_seq;
 
 	union {
@@ -1152,11 +1349,25 @@ struct bch_sb_field_journal_seq_blacklist {
 #define BCH_JSET_VERSION_OLD			2
 #define BCH_BSET_VERSION_OLD			3
 
+#define BCH_METADATA_VERSIONS()				\
+	x(bkey_renumber,		10)		\
+	x(inode_btree_change,		11)		\
+	x(snapshot,			12)		\
+	x(inode_backpointers,		13)		\
+	x(btree_ptr_sectors_written,	14)		\
+	x(snapshot_2,			15)		\
+	x(reflink_p_fix,		16)		\
+	x(subvol_dirent,		17)		\
+	x(inode_v2,			18)		\
+	x(freespace,			19)		\
+	x(alloc_v4,			20)
+
 enum bcachefs_metadata_version {
-	bcachefs_metadata_version_min			= 9,
-	bcachefs_metadata_version_new_versioning	= 10,
-	bcachefs_metadata_version_bkey_renumber		= 10,
-	bcachefs_metadata_version_max			= 11,
+	bcachefs_metadata_version_min = 9,
+#define x(t, n)	bcachefs_metadata_version_##t = n,
+	BCH_METADATA_VERSIONS()
+#undef x
+	bcachefs_metadata_version_max
 };
 
 #define bcachefs_metadata_version_current	(bcachefs_metadata_version_max - 1)
@@ -1260,8 +1471,9 @@ LE64_BITMASK(BCH_SB_GRPQUOTA,		struct bch_sb, flags[0], 58, 59);
 LE64_BITMASK(BCH_SB_PRJQUOTA,		struct bch_sb, flags[0], 59, 60);
 
 LE64_BITMASK(BCH_SB_HAS_ERRORS,		struct bch_sb, flags[0], 60, 61);
+LE64_BITMASK(BCH_SB_HAS_TOPOLOGY_ERRORS,struct bch_sb, flags[0], 61, 62);
 
-/* 61-64 unused */
+LE64_BITMASK(BCH_SB_BIG_ENDIAN,		struct bch_sb, flags[0], 62, 63);
 
 LE64_BITMASK(BCH_SB_STR_HASH_TYPE,	struct bch_sb, flags[1],  0,  4);
 LE64_BITMASK(BCH_SB_COMPRESSION_TYPE,	struct bch_sb, flags[1],  4,  8);
@@ -1289,6 +1501,13 @@ LE64_BITMASK(BCH_SB_BACKGROUND_COMPRESSION_TYPE,
 LE64_BITMASK(BCH_SB_GC_RESERVE_BYTES,	struct bch_sb, flags[2],  4, 64);
 
 LE64_BITMASK(BCH_SB_ERASURE_CODE,	struct bch_sb, flags[3],  0, 16);
+LE64_BITMASK(BCH_SB_METADATA_TARGET,	struct bch_sb, flags[3], 16, 28);
+LE64_BITMASK(BCH_SB_SHARD_INUMS,	struct bch_sb, flags[3], 28, 29);
+LE64_BITMASK(BCH_SB_INODES_USE_KEY_CACHE,struct bch_sb, flags[3], 29, 30);
+LE64_BITMASK(BCH_SB_JOURNAL_FLUSH_DELAY,struct bch_sb, flags[3], 30, 62);
+LE64_BITMASK(BCH_SB_JOURNAL_FLUSH_DISABLED,struct bch_sb, flags[3], 62, 63);
+LE64_BITMASK(BCH_SB_JOURNAL_RECLAIM_DELAY,struct bch_sb, flags[4], 0, 32);
+LE64_BITMASK(BCH_SB_JOURNAL_TRANSACTION_NAMES,struct bch_sb, flags[4], 32, 33);
 
 /*
  * Features:
@@ -1296,7 +1515,7 @@ LE64_BITMASK(BCH_SB_ERASURE_CODE,	struct bch_sb, flags[3],  0, 16);
  * journal_seq_blacklist_v3:	gates BCH_SB_FIELD_journal_seq_blacklist
  * reflink:			gates KEY_TYPE_reflink
  * inline_data:			gates KEY_TYPE_inline_data
- * new_siphash:			gates BCH_STR_HASH_SIPHASH
+ * new_siphash:			gates BCH_STR_HASH_siphash
  * new_extent_overwrite:	gates BTREE_NODE_NEW_EXTENT_OVERWRITE
  */
 #define BCH_SB_FEATURES()			\
@@ -1313,13 +1532,26 @@ LE64_BITMASK(BCH_SB_ERASURE_CODE,	struct bch_sb, flags[3],  0, 16);
 	x(incompressible,		10)	\
 	x(btree_ptr_v2,			11)	\
 	x(extents_above_btree_updates,	12)	\
-	x(btree_updates_journalled,	13)
+	x(btree_updates_journalled,	13)	\
+	x(reflink_inline_data,		14)	\
+	x(new_varint,			15)	\
+	x(journal_no_flush,		16)	\
+	x(alloc_v2,			17)	\
+	x(extents_across_btree_nodes,	18)
+
+#define BCH_SB_FEATURES_ALWAYS				\
+	((1ULL << BCH_FEATURE_new_extent_overwrite)|	\
+	 (1ULL << BCH_FEATURE_extents_above_btree_updates)|\
+	 (1ULL << BCH_FEATURE_btree_updates_journalled)|\
+	 (1ULL << BCH_FEATURE_alloc_v2)|\
+	 (1ULL << BCH_FEATURE_extents_across_btree_nodes))
 
 #define BCH_SB_FEATURES_ALL				\
-	((1ULL << BCH_FEATURE_new_siphash)|		\
-	 (1ULL << BCH_FEATURE_new_extent_overwrite)|	\
+	(BCH_SB_FEATURES_ALWAYS|			\
+	 (1ULL << BCH_FEATURE_new_siphash)|		\
 	 (1ULL << BCH_FEATURE_btree_ptr_v2)|		\
-	 (1ULL << BCH_FEATURE_extents_above_btree_updates))
+	 (1ULL << BCH_FEATURE_new_varint)|		\
+	 (1ULL << BCH_FEATURE_journal_no_flush))
 
 enum bch_sb_feature {
 #define x(f, n) BCH_FEATURE_##f,
@@ -1328,74 +1560,112 @@ enum bch_sb_feature {
 	BCH_FEATURE_NR,
 };
 
+#define BCH_SB_COMPAT()					\
+	x(alloc_info,				0)	\
+	x(alloc_metadata,			1)	\
+	x(extents_above_btree_updates_done,	2)	\
+	x(bformat_overflow_done,		3)
+
 enum bch_sb_compat {
-	BCH_COMPAT_FEAT_ALLOC_INFO	= 0,
-	BCH_COMPAT_FEAT_ALLOC_METADATA	= 1,
+#define x(f, n) BCH_COMPAT_##f,
+	BCH_SB_COMPAT()
+#undef x
+	BCH_COMPAT_NR,
 };
 
 /* options: */
 
 #define BCH_REPLICAS_MAX		4U
 
+#define BCH_BKEY_PTRS_MAX		16U
+
+#define BCH_ERROR_ACTIONS()		\
+	x(continue,		0)	\
+	x(ro,			1)	\
+	x(panic,		2)
+
 enum bch_error_actions {
-	BCH_ON_ERROR_CONTINUE		= 0,
-	BCH_ON_ERROR_RO			= 1,
-	BCH_ON_ERROR_PANIC		= 2,
-	BCH_NR_ERROR_ACTIONS		= 3,
+#define x(t, n) BCH_ON_ERROR_##t = n,
+	BCH_ERROR_ACTIONS()
+#undef x
+	BCH_ON_ERROR_NR
 };
 
+#define BCH_STR_HASH_TYPES()		\
+	x(crc32c,		0)	\
+	x(crc64,		1)	\
+	x(siphash_old,		2)	\
+	x(siphash,		3)
+
 enum bch_str_hash_type {
-	BCH_STR_HASH_CRC32C		= 0,
-	BCH_STR_HASH_CRC64		= 1,
-	BCH_STR_HASH_SIPHASH_OLD	= 2,
-	BCH_STR_HASH_SIPHASH		= 3,
-	BCH_STR_HASH_NR			= 4,
+#define x(t, n) BCH_STR_HASH_##t = n,
+	BCH_STR_HASH_TYPES()
+#undef x
+	BCH_STR_HASH_NR
 };
 
+#define BCH_STR_HASH_OPTS()		\
+	x(crc32c,		0)	\
+	x(crc64,		1)	\
+	x(siphash,		2)
+
 enum bch_str_hash_opts {
-	BCH_STR_HASH_OPT_CRC32C		= 0,
-	BCH_STR_HASH_OPT_CRC64		= 1,
-	BCH_STR_HASH_OPT_SIPHASH	= 2,
-	BCH_STR_HASH_OPT_NR		= 3,
+#define x(t, n) BCH_STR_HASH_OPT_##t = n,
+	BCH_STR_HASH_OPTS()
+#undef x
+	BCH_STR_HASH_OPT_NR
 };
 
+#define BCH_CSUM_TYPES()			\
+	x(none,				0)	\
+	x(crc32c_nonzero,		1)	\
+	x(crc64_nonzero,		2)	\
+	x(chacha20_poly1305_80,		3)	\
+	x(chacha20_poly1305_128,	4)	\
+	x(crc32c,			5)	\
+	x(crc64,			6)	\
+	x(xxhash,			7)
+
 enum bch_csum_type {
-	BCH_CSUM_NONE			= 0,
-	BCH_CSUM_CRC32C_NONZERO		= 1,
-	BCH_CSUM_CRC64_NONZERO		= 2,
-	BCH_CSUM_CHACHA20_POLY1305_80	= 3,
-	BCH_CSUM_CHACHA20_POLY1305_128	= 4,
-	BCH_CSUM_CRC32C			= 5,
-	BCH_CSUM_CRC64			= 6,
-	BCH_CSUM_NR			= 7,
+#define x(t, n) BCH_CSUM_##t = n,
+	BCH_CSUM_TYPES()
+#undef x
+	BCH_CSUM_NR
 };
 
 static const unsigned bch_crc_bytes[] = {
-	[BCH_CSUM_NONE]				= 0,
-	[BCH_CSUM_CRC32C_NONZERO]		= 4,
-	[BCH_CSUM_CRC32C]			= 4,
-	[BCH_CSUM_CRC64_NONZERO]		= 8,
-	[BCH_CSUM_CRC64]			= 8,
-	[BCH_CSUM_CHACHA20_POLY1305_80]		= 10,
-	[BCH_CSUM_CHACHA20_POLY1305_128]	= 16,
+	[BCH_CSUM_none]				= 0,
+	[BCH_CSUM_crc32c_nonzero]		= 4,
+	[BCH_CSUM_crc32c]			= 4,
+	[BCH_CSUM_crc64_nonzero]		= 8,
+	[BCH_CSUM_crc64]			= 8,
+	[BCH_CSUM_xxhash]			= 8,
+	[BCH_CSUM_chacha20_poly1305_80]		= 10,
+	[BCH_CSUM_chacha20_poly1305_128]	= 16,
 };
 
 static inline _Bool bch2_csum_type_is_encryption(enum bch_csum_type type)
 {
 	switch (type) {
-	case BCH_CSUM_CHACHA20_POLY1305_80:
-	case BCH_CSUM_CHACHA20_POLY1305_128:
+	case BCH_CSUM_chacha20_poly1305_80:
+	case BCH_CSUM_chacha20_poly1305_128:
 		return true;
 	default:
 		return false;
 	}
 }
 
+#define BCH_CSUM_OPTS()			\
+	x(none,			0)	\
+	x(crc32c,		1)	\
+	x(crc64,		2)	\
+	x(xxhash,		3)
+
 enum bch_csum_opts {
-	BCH_CSUM_OPT_NONE		= 0,
-	BCH_CSUM_OPT_CRC32C		= 1,
-	BCH_CSUM_OPT_CRC64		= 2,
-	BCH_CSUM_OPT_NR			= 3,
+#define x(t, n) BCH_CSUM_OPT_##t = n,
+	BCH_CSUM_OPTS()
+#undef x
+	BCH_CSUM_OPT_NR
 };
 
 #define BCH_COMPRESSION_TYPES()		\
@@ -1407,7 +1677,7 @@ enum bch_csum_opts {
 	x(incompressible,	5)
 
 enum bch_compression_type {
-#define x(t, n) BCH_COMPRESSION_TYPE_##t,
+#define x(t, n) BCH_COMPRESSION_TYPE_##t = n,
 	BCH_COMPRESSION_TYPES()
 #undef x
 	BCH_COMPRESSION_TYPE_NR
@@ -1420,7 +1690,7 @@ enum bch_compression_type {
 	x(zstd,		3)
 
 enum bch_compression_opts {
-#define x(t, n) BCH_COMPRESSION_OPT_##t,
+#define x(t, n) BCH_COMPRESSION_OPT_##t = n,
 	BCH_COMPRESSION_OPTS()
 #undef x
 	BCH_COMPRESSION_OPT_NR
@@ -1470,7 +1740,10 @@ static inline __u64 __bset_magic(struct bch_sb *sb)
 	x(blacklist,		3)		\
 	x(blacklist_v2,		4)		\
 	x(usage,		5)		\
-	x(data_usage,		6)
+	x(data_usage,		6)		\
+	x(clock,		7)		\
+	x(dev_usage,		8)		\
+	x(log,			9)
 
 enum {
 #define x(f, nr)	BCH_JSET_ENTRY_##f	= nr,
@@ -1500,11 +1773,16 @@ struct jset_entry_blacklist_v2 {
 	__le64			end;
 };
 
+#define BCH_FS_USAGE_TYPES()			\
+	x(reserved,		0)		\
+	x(inodes,		1)		\
+	x(key_version,		2)
+
 enum {
-	FS_USAGE_RESERVED		= 0,
-	FS_USAGE_INODES			= 1,
-	FS_USAGE_KEY_VERSION		= 2,
-	FS_USAGE_NR			= 3
+#define x(f, nr)	BCH_FS_USAGE_##f	= nr,
+	BCH_FS_USAGE_TYPES()
+#undef x
+	BCH_FS_USAGE_NR
 };
 
 struct jset_entry_usage {
@@ -1518,6 +1796,41 @@ struct jset_entry_data_usage {
 	struct bch_replicas_entry r;
 } __attribute__((packed));
 
+struct jset_entry_clock {
+	struct jset_entry	entry;
+	__u8			rw;
+	__u8			pad[7];
+	__le64			time;
+} __attribute__((packed));
+
+struct jset_entry_dev_usage_type {
+	__le64			buckets;
+	__le64			sectors;
+	__le64			fragmented;
+} __attribute__((packed));
+
+struct jset_entry_dev_usage {
+	struct jset_entry	entry;
+	__le32			dev;
+	__u32			pad;
+
+	__le64			buckets_ec;
+	__le64			buckets_unavailable;
+
+	struct jset_entry_dev_usage_type d[];
+} __attribute__((packed));
+
+static inline unsigned jset_entry_dev_usage_nr_types(struct jset_entry_dev_usage *u)
+{
+	return (vstruct_bytes(&u->entry) - sizeof(struct jset_entry_dev_usage)) /
+		sizeof(struct jset_entry_dev_usage_type);
+}
+
+struct jset_entry_log {
+	struct jset_entry	entry;
+	u8			d[];
+} __attribute__((packed));
+
 /*
  * On disk format for a journal entry:
  * seq is monotonically increasing; every journal entry has its own unique
@@ -1540,8 +1853,8 @@ struct jset {
 
 	__u8			encrypted_start[0];
 
-	__le16			read_clock;
-	__le16			write_clock;
+	__le16			_read_clock; /* no longer used */
+	__le16			_write_clock;
 
 	/* Sequence number of oldest dirty journal entry */
 	__le64			last_seq;
@@ -1555,23 +1868,30 @@ struct jset {
 
 LE32_BITMASK(JSET_CSUM_TYPE,	struct jset, flags, 0, 4);
 LE32_BITMASK(JSET_BIG_ENDIAN,	struct jset, flags, 4, 5);
+LE32_BITMASK(JSET_NO_FLUSH,	struct jset, flags, 5, 6);
 
 #define BCH_JOURNAL_BUCKETS_MIN		8
 
 /* Btree: */
 
-#define BCH_BTREE_IDS()					\
-	x(EXTENTS,	0, "extents")			\
-	x(INODES,	1, "inodes")			\
-	x(DIRENTS,	2, "dirents")			\
-	x(XATTRS,	3, "xattrs")			\
-	x(ALLOC,	4, "alloc")			\
-	x(QUOTAS,	5, "quotas")			\
-	x(EC,		6, "stripes")			\
-	x(REFLINK,	7, "reflink")
+#define BCH_BTREE_IDS()				\
+	x(extents,	0)			\
+	x(inodes,	1)			\
+	x(dirents,	2)			\
+	x(xattrs,	3)			\
+	x(alloc,	4)			\
+	x(quotas,	5)			\
+	x(stripes,	6)			\
+	x(reflink,	7)			\
+	x(subvolumes,	8)			\
+	x(snapshots,	9)			\
+	x(lru,		10)			\
+	x(freespace,	11)			\
+	x(need_discard,	12)			\
+	x(backpointers,	13)
 
 enum btree_id {
-#define x(kwd, val, name) BTREE_ID_##kwd = val,
+#define x(kwd, val) BTREE_ID_##kwd = val,
 	BCH_BTREE_IDS()
 #undef x
 	BTREE_ID_NR
@@ -1615,6 +1935,9 @@ LE32_BITMASK(BSET_BIG_ENDIAN,	struct bset, flags, 4, 5);
 LE32_BITMASK(BSET_SEPARATE_WHITEOUTS,
 				struct bset, flags, 5, 6);
 
+/* Sector offset within the btree node: */
+LE32_BITMASK(BSET_OFFSET,	struct bset, flags, 16, 32);
+
 struct btree_node {
 	struct bch_csum		csum;
 	__le64			magic;
@@ -1625,7 +1948,7 @@ struct btree_node {
 	/* Closed interval: */
 	struct bpos		min_key;
 	struct bpos		max_key;
-	struct bch_extent_ptr	ptr;
+	struct bch_extent_ptr	_ptr; /* not used anymore */
 	struct bkey_format	format;
 
 	union {