+/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _BCACHEFS_FORMAT_H
#define _BCACHEFS_FORMAT_H
/*
* bcachefs on disk data structures
+ *
+ * OVERVIEW:
+ *
+ * There are three main types of on disk data structures in bcachefs (this is
+ * reduced from 5 in bcache)
+ *
+ * - superblock
+ * - journal
+ * - btree
+ *
+ * The btree is the primary structure; most metadata exists as keys in the
+ * various btrees. There are only a small number of btrees, they're not
+ * sharded - we have one btree for extents, another for inodes, et cetera.
+ *
+ * SUPERBLOCK:
+ *
+ * The superblock contains the location of the journal, the list of devices in
+ * the filesystem, and in general any metadata we need in order to decide
+ * whether we can start a filesystem or prior to reading the journal/btree
+ * roots.
+ *
+ * The superblock is extensible, and most of the contents of the superblock are
+ * in variable length, type tagged fields; see struct bch_sb_field.
+ *
+ * Backup superblocks do not reside in a fixed location; also, superblocks do
+ * not have a fixed size. To locate backup superblocks we have struct
+ * bch_sb_layout; we store a copy of this inside every superblock, and also
+ * before the first superblock.
+ *
+ * JOURNAL:
+ *
+ * The journal primarily records btree updates in the order they occurred;
+ * journal replay consists of just iterating over all the keys in the open
+ * journal entries and re-inserting them into the btrees.
+ *
+ * The journal also contains entry types for the btree roots, and blacklisted
+ * journal sequence numbers (see journal_seq_blacklist.c).
+ *
+ * BTREE:
+ *
+ * bcachefs btrees are copy on write b+ trees, where nodes are big (typically
+ * 128k-256k) and log structured. We use struct btree_node for writing the first
+ * entry in a given node (offset 0), and struct btree_node_entry for all
+ * subsequent writes.
+ *
+ * After the header, btree node entries contain a list of keys in sorted order.
+ * Values are stored inline with the keys; since values are variable length (and
+ * keys effectively are variable length too, due to packing) we can't do random
+ * access without building up additional in memory tables in the btree node read
+ * path.
+ *
+ * BTREE KEYS (struct bkey):
+ *
+ * The various btrees share a common format for the key - so as to avoid
+ * switching in fastpath lookup/comparison code - but define their own
+ * structures for the key values.
+ *
+ * The size of a key/value pair is stored as a u8 in units of u64s, so the max
+ * size is just under 2k. The common part also contains a type tag for the
+ * value, and a format field indicating whether the key is packed or not (and
+ * also meant to allow adding new key fields in the future, if desired).
+ *
+ * bkeys, when stored within a btree node, may also be packed. In that case, the
+ * bkey_format in that node is used to unpack it. Packed bkeys mean that we can
+ * be generous with field sizes in the common part of the key format (64 bit
+ * inode number, 64 bit offset, 96 bit version field, etc.) for negligible cost.
*/
#include <asm/types.h>
#include <asm/byteorder.h>
+#include <linux/kernel.h>
#include <linux/uuid.h>
#define LE_BITMASK(_bits, name, type, field, offset, end) \
/* Btree keys - all units are in sectors */
struct bpos {
- /* Word order matches machine byte order */
-#if defined(__LITTLE_ENDIAN)
+ /*
+ * Word order matches machine byte order - btree code treats a bpos as a
+ * single large integer, for search/comparison purposes
+ *
+ * Note that wherever a bpos is embedded in another on disk data
+ * structure, it has to be byte swabbed when reading in metadata that
+ * wasn't written in native endian order:
+ */
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
__u32 snapshot;
__u64 offset;
__u64 inode;
-#elif defined(__BIG_ENDIAN)
+#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
__u64 inode;
__u64 offset; /* Points to end of extent - sectors */
__u32 snapshot;
#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, KEY_SNAPSHOT_MAX)
+#define POS(_inode, _offset) SPOS(_inode, _offset, 0)
/* Empty placeholder struct, for container_of() */
struct bch_val {
};
struct bversion {
-#if defined(__LITTLE_ENDIAN)
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
__u64 lo;
__u32 hi;
-#elif defined(__BIG_ENDIAN)
+#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
__u32 hi;
__u64 lo;
#endif
/* Type of the value */
__u8 type;
-#if defined(__LITTLE_ENDIAN)
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
__u8 pad[1];
struct bversion version;
__u32 size; /* extent size, in sectors */
struct bpos p;
-#elif defined(__BIG_ENDIAN)
+#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
struct bpos p;
__u32 size; /* extent size, in sectors */
struct bversion version;
} __attribute__((packed, aligned(8)));
#define BKEY_U64s (sizeof(struct bkey) / sizeof(__u64))
+#define BKEY_U64s_MAX U8_MAX
+#define BKEY_VAL_U64s_MAX (BKEY_U64s_MAX - BKEY_U64s)
+
#define KEY_PACKED_BITS_START 24
#define KEY_FORMAT_LOCAL_BTREE 0
#define __BKEY_PADDED(key, pad) \
struct { struct bkey_i key; __u64 key ## _pad[pad]; }
-#define BKEY_VAL_TYPE(name, nr) \
-struct bkey_i_##name { \
- union { \
- struct bkey k; \
- struct bkey_i k_i; \
- }; \
- struct bch_##name v; \
-}
-
/*
* - DELETED keys are used internally to mark keys that should be ignored but
* override keys in composition order. Their version number is ignored.
* by new writes or cluster-wide GC. Node repair can also overwrite them with
* the same or a more recent version number, but not with an older version
* number.
+ *
+ * - WHITEOUT: for hash table btrees
*/
-#define KEY_TYPE_DELETED 0
-#define KEY_TYPE_DISCARD 1
-#define KEY_TYPE_ERROR 2
-#define KEY_TYPE_COOKIE 3
-#define KEY_TYPE_PERSISTENT_DISCARD 4
-#define KEY_TYPE_GENERIC_NR 128
+#define BCH_BKEY_TYPES() \
+ x(deleted, 0) \
+ x(discard, 1) \
+ x(error, 2) \
+ x(cookie, 3) \
+ x(hash_whiteout, 4) \
+ x(btree_ptr, 5) \
+ x(extent, 6) \
+ x(reservation, 7) \
+ x(inode, 8) \
+ x(inode_generation, 9) \
+ x(dirent, 10) \
+ x(xattr, 11) \
+ x(alloc, 12) \
+ x(quota, 13) \
+ x(stripe, 14) \
+ x(reflink_p, 15) \
+ x(reflink_v, 16) \
+ x(inline_data, 17) \
+ x(btree_ptr_v2, 18) \
+ x(indirect_inline_data, 19) \
+ x(alloc_v2, 20)
+
+enum bch_bkey_type {
+#define x(name, nr) KEY_TYPE_##name = nr,
+ BCH_BKEY_TYPES()
+#undef x
+ KEY_TYPE_MAX,
+};
+
+struct bch_deleted {
+ struct bch_val v;
+};
+
+struct bch_discard {
+ struct bch_val v;
+};
+
+struct bch_error {
+ struct bch_val v;
+};
struct bch_cookie {
struct bch_val v;
__le64 cookie;
};
-BKEY_VAL_TYPE(cookie, KEY_TYPE_COOKIE);
+
+struct bch_hash_whiteout {
+ struct bch_val v;
+};
/* Extents */
*
* If an extent is not checksummed or compressed, when the extent is trimmed we
* don't have to remember the extent we originally allocated and wrote: we can
- * merely adjust ptr->offset to point to the start of the start of the data that
- * is currently live. The size field in struct bkey records the current (live)
- * size of the extent, and is also used to mean "size of region on disk that we
- * point to" in this case.
+ * merely adjust ptr->offset to point to the start of the data that is currently
+ * live. The size field in struct bkey records the current (live) size of the
+ * extent, and is also used to mean "size of region on disk that we point to" in
+ * this case.
*
* Thus an extent that is not checksummed or compressed will consist only of a
* list of bch_extent_ptrs, with none of the fields in
__le64 hi;
} __attribute__((packed, aligned(8)));
-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,
-};
-
-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:
- return true;
- default:
- return false;
- }
-}
-
-enum bch_compression_type {
- BCH_COMPRESSION_NONE = 0,
- BCH_COMPRESSION_LZ4_OLD = 1,
- BCH_COMPRESSION_GZIP = 2,
- BCH_COMPRESSION_LZ4 = 3,
- BCH_COMPRESSION_ZSTD = 4,
- BCH_COMPRESSION_NR = 5,
-};
+#define BCH_EXTENT_ENTRY_TYPES() \
+ x(ptr, 0) \
+ x(crc32, 1) \
+ x(crc64, 2) \
+ x(crc128, 3) \
+ x(stripe_ptr, 4)
+#define BCH_EXTENT_ENTRY_MAX 5
enum bch_extent_entry_type {
- BCH_EXTENT_ENTRY_ptr = 0,
- BCH_EXTENT_ENTRY_crc32 = 1,
- BCH_EXTENT_ENTRY_crc64 = 2,
- BCH_EXTENT_ENTRY_crc128 = 3,
+#define x(f, n) BCH_EXTENT_ENTRY_##f = n,
+ BCH_EXTENT_ENTRY_TYPES()
+#undef x
};
-#define BCH_EXTENT_ENTRY_MAX 4
-
/* Compressed/uncompressed size are stored biased by 1: */
struct bch_extent_crc32 {
#if defined(__LITTLE_ENDIAN_BITFIELD)
#elif defined (__BIG_ENDIAN_BITFIELD)
__u64 compression_type:4,
csum_type:4,
- nonce:14,
+ nonce:13,
offset:13,
_uncompressed_size:13,
_compressed_size:13,
- type:3;
+ type:4;
#endif
struct bch_csum csum;
} __attribute__((packed, aligned(8)));
#if defined(__LITTLE_ENDIAN_BITFIELD)
__u64 type:1,
cached:1,
- erasure_coded:1,
+ unused:1,
reservation:1,
offset:44, /* 8 petabytes */
dev:8,
dev:8,
offset:44,
reservation:1,
- erasure_coded:1,
+ unused:1,
cached:1,
type:1;
#endif
} __attribute__((packed, aligned(8)));
-struct bch_extent_reservation {
+struct bch_extent_stripe_ptr {
#if defined(__LITTLE_ENDIAN_BITFIELD)
__u64 type:5,
- unused:23,
+ block:8,
+ redundancy:4,
+ idx:47;
+#elif defined (__BIG_ENDIAN_BITFIELD)
+ __u64 idx:47,
+ redundancy:4,
+ block:8,
+ type:5;
+#endif
+};
+
+struct bch_extent_reservation {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+ __u64 type:6,
+ unused:22,
replicas:4,
generation:32;
#elif defined (__BIG_ENDIAN_BITFIELD)
__u64 generation:32,
replicas:4,
- unused:23,
- type:5;
+ unused:22,
+ type:6;
#endif
};
union bch_extent_entry {
-#if defined(__LITTLE_ENDIAN) || __BITS_PER_LONG == 64
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__ || __BITS_PER_LONG == 64
unsigned long type;
#elif __BITS_PER_LONG == 32
struct {
#else
#error edit for your odd byteorder.
#endif
- struct bch_extent_crc32 crc32;
- struct bch_extent_crc64 crc64;
- struct bch_extent_crc128 crc128;
- struct bch_extent_ptr ptr;
+
+#define x(f, n) struct bch_extent_##f f;
+ BCH_EXTENT_ENTRY_TYPES()
+#undef x
};
-enum {
- BCH_EXTENT = 128,
+struct bch_btree_ptr {
+ struct bch_val v;
- /*
- * This is kind of a hack, we're overloading the type for a boolean that
- * really should be part of the value - BCH_EXTENT and BCH_EXTENT_CACHED
- * have the same value type:
- */
- BCH_EXTENT_CACHED = 129,
+ struct bch_extent_ptr start[0];
+ __u64 _data[0];
+} __attribute__((packed, aligned(8)));
- /*
- * Persistent reservation:
- */
- BCH_RESERVATION = 130,
-};
+struct bch_btree_ptr_v2 {
+ struct bch_val v;
+
+ __u64 mem_ptr;
+ __le64 seq;
+ __le16 sectors_written;
+ __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;
union bch_extent_entry start[0];
__u64 _data[0];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(extent, BCH_EXTENT);
struct bch_reservation {
struct bch_val v;
__u8 nr_replicas;
__u8 pad[3];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(reservation, BCH_RESERVATION);
/* Maximum size (in u64s) a single pointer could be: */
#define BKEY_EXTENT_PTR_U64s_MAX\
sizeof(struct bch_extent_ptr)) / sizeof(u64))
/* Maximum possible size of an entire extent value: */
-/* There's a hack in the keylist code that needs to be fixed.. */
#define BKEY_EXTENT_VAL_U64s_MAX \
- (BKEY_EXTENT_PTR_U64s_MAX * (BCH_REPLICAS_MAX + 1))
+ (1 + BKEY_EXTENT_PTR_U64s_MAX * (BCH_REPLICAS_MAX + 1))
/* * Maximum possible size of an entire extent, key + value: */
#define BKEY_EXTENT_U64s_MAX (BKEY_U64s + BKEY_EXTENT_VAL_U64s_MAX)
/* Btree pointers don't carry around checksums: */
#define BKEY_BTREE_PTR_VAL_U64s_MAX \
- ((sizeof(struct bch_extent_ptr)) / sizeof(u64) * BCH_REPLICAS_MAX)
+ ((sizeof(struct bch_btree_ptr_v2) + \
+ sizeof(struct bch_extent_ptr) * BCH_REPLICAS_MAX) / sizeof(u64))
#define BKEY_BTREE_PTR_U64s_MAX \
(BKEY_U64s + BKEY_BTREE_PTR_VAL_U64s_MAX)
#define BCACHEFS_ROOT_INO 4096
-enum bch_inode_types {
- BCH_INODE_FS = 128,
- BCH_INODE_BLOCKDEV = 129,
- BCH_INODE_GENERATION = 130,
-};
-
struct bch_inode {
struct bch_val v;
__le16 bi_mode;
__u8 fields[0];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(inode, BCH_INODE_FS);
struct bch_inode_generation {
struct bch_val v;
__le32 bi_generation;
__le32 pad;
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(inode_generation, BCH_INODE_GENERATION);
-
-#define BCH_INODE_FIELDS() \
- BCH_INODE_FIELD(bi_atime, 64) \
- BCH_INODE_FIELD(bi_ctime, 64) \
- BCH_INODE_FIELD(bi_mtime, 64) \
- BCH_INODE_FIELD(bi_otime, 64) \
- BCH_INODE_FIELD(bi_size, 64) \
- BCH_INODE_FIELD(bi_sectors, 64) \
- BCH_INODE_FIELD(bi_uid, 32) \
- BCH_INODE_FIELD(bi_gid, 32) \
- BCH_INODE_FIELD(bi_nlink, 32) \
- BCH_INODE_FIELD(bi_generation, 32) \
- BCH_INODE_FIELD(bi_dev, 32) \
- BCH_INODE_FIELD(bi_data_checksum, 8) \
- BCH_INODE_FIELD(bi_compression, 8) \
- BCH_INODE_FIELD(bi_project, 32) \
- BCH_INODE_FIELD(bi_background_compression, 8) \
- BCH_INODE_FIELD(bi_data_replicas, 8) \
- BCH_INODE_FIELD(bi_promote_target, 16) \
- BCH_INODE_FIELD(bi_foreground_target, 16) \
- BCH_INODE_FIELD(bi_background_target, 16)
-
-#define BCH_INODE_FIELDS_INHERIT() \
- BCH_INODE_FIELD(bi_data_checksum) \
- BCH_INODE_FIELD(bi_compression) \
- BCH_INODE_FIELD(bi_project) \
- BCH_INODE_FIELD(bi_background_compression) \
- BCH_INODE_FIELD(bi_data_replicas) \
- BCH_INODE_FIELD(bi_promote_target) \
- BCH_INODE_FIELD(bi_foreground_target) \
- BCH_INODE_FIELD(bi_background_target)
+
+#define BCH_INODE_FIELDS() \
+ 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) \
+ x(bi_gid, 32) \
+ x(bi_nlink, 32) \
+ x(bi_generation, 32) \
+ x(bi_dev, 32) \
+ x(bi_data_checksum, 8) \
+ x(bi_compression, 8) \
+ x(bi_project, 32) \
+ x(bi_background_compression, 8) \
+ x(bi_data_replicas, 8) \
+ x(bi_promote_target, 16) \
+ x(bi_foreground_target, 16) \
+ x(bi_background_target, 16) \
+ x(bi_erasure_code, 16) \
+ x(bi_fields_set, 16) \
+ x(bi_dir, 64) \
+ x(bi_dir_offset, 64)
+
+/* subset of BCH_INODE_FIELDS */
+#define BCH_INODE_OPTS() \
+ x(data_checksum, 8) \
+ x(compression, 8) \
+ x(project, 32) \
+ x(background_compression, 8) \
+ x(data_replicas, 8) \
+ x(promote_target, 16) \
+ x(foreground_target, 16) \
+ x(background_target, 16) \
+ x(erasure_code, 16)
+
+enum inode_opt_id {
+#define x(name, ...) \
+ Inode_opt_##name,
+ BCH_INODE_OPTS()
+#undef x
+ Inode_opt_nr,
+};
enum {
/*
__BCH_INODE_I_SIZE_DIRTY= 5,
__BCH_INODE_I_SECTORS_DIRTY= 6,
-
- /* not implemented yet: */
- __BCH_INODE_HAS_XATTRS = 7, /* has xattrs in xattr btree */
+ __BCH_INODE_UNLINKED = 7,
+ __BCH_INODE_BACKPTR_UNTRUSTED = 8,
/* bits 20+ reserved for packed fields below: */
};
#define BCH_INODE_NOATIME (1 << __BCH_INODE_NOATIME)
#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_HAS_XATTRS (1 << __BCH_INODE_HAS_XATTRS)
+#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);
-
-struct bch_inode_blockdev {
- struct bch_val v;
-
- __le64 i_size;
- __le64 i_flags;
-
- /* Seconds: */
- __le64 i_ctime;
- __le64 i_mtime;
-
- uuid_le i_uuid;
- __u8 i_label[32];
-} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(inode_blockdev, BCH_INODE_BLOCKDEV);
-
-/* Thin provisioned volume, or cache for another block device? */
-LE64_BITMASK(CACHED_DEV, struct bch_inode_blockdev, i_flags, 0, 1)
+LE32_BITMASK(INODE_NR_FIELDS, struct bch_inode, bi_flags, 24, 31);
+LE32_BITMASK(INODE_NEW_VARINT, struct bch_inode, bi_flags, 31, 32);
/* Dirents */
* collision:
*/
-enum {
- BCH_DIRENT = 128,
- BCH_DIRENT_WHITEOUT = 129,
-};
-
struct bch_dirent {
struct bch_val v;
__u8 d_name[];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(dirent, BCH_DIRENT);
-/* Xattrs */
+#define BCH_NAME_MAX (U8_MAX * sizeof(u64) - \
+ sizeof(struct bkey) - \
+ offsetof(struct bch_dirent, d_name))
-enum {
- BCH_XATTR = 128,
- BCH_XATTR_WHITEOUT = 129,
-};
-#define BCH_XATTR_INDEX_USER 0
-#define BCH_XATTR_INDEX_POSIX_ACL_ACCESS 1
-#define BCH_XATTR_INDEX_POSIX_ACL_DEFAULT 2
-#define BCH_XATTR_INDEX_TRUSTED 3
-#define BCH_XATTR_INDEX_SECURITY 4
+/* Xattrs */
+
+#define KEY_TYPE_XATTR_INDEX_USER 0
+#define KEY_TYPE_XATTR_INDEX_POSIX_ACL_ACCESS 1
+#define KEY_TYPE_XATTR_INDEX_POSIX_ACL_DEFAULT 2
+#define KEY_TYPE_XATTR_INDEX_TRUSTED 3
+#define KEY_TYPE_XATTR_INDEX_SECURITY 4
struct bch_xattr {
struct bch_val v;
__le16 x_val_len;
__u8 x_name[];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(xattr, BCH_XATTR);
/* Bucket/allocation information: */
-enum {
- BCH_ALLOC = 128,
-};
-
-enum {
- BCH_ALLOC_FIELD_READ_TIME = 0,
- BCH_ALLOC_FIELD_WRITE_TIME = 1,
-};
-
struct bch_alloc {
struct bch_val v;
__u8 fields;
__u8 gen;
__u8 data[];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(alloc, BCH_ALLOC);
-/* Quotas: */
+#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(stripe, 32) \
+ x(stripe_redundancy, 8)
+
+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)));
+
+#define BCH_ALLOC_FIELDS_V2() \
+ x(read_time, 64) \
+ x(write_time, 64) \
+ x(dirty_sectors, 16) \
+ x(cached_sectors, 16) \
+ x(stripe, 32) \
+ x(stripe_redundancy, 8)
enum {
- BCH_QUOTA = 128,
+#define x(name, _bits) BCH_ALLOC_FIELD_V1_##name,
+ BCH_ALLOC_FIELDS_V1()
+#undef x
+ BCH_ALLOC_FIELD_NR
};
+/* Quotas: */
+
enum quota_types {
QTYP_USR = 0,
QTYP_GRP = 1,
struct bch_val v;
struct bch_quota_counter c[Q_COUNTERS];
} __attribute__((packed, aligned(8)));
-BKEY_VAL_TYPE(quota, BCH_QUOTA);
+
+/* Erasure coding */
+
+struct bch_stripe {
+ struct bch_val v;
+ __le16 sectors;
+ __u8 algorithm;
+ __u8 nr_blocks;
+ __u8 nr_redundant;
+
+ __u8 csum_granularity_bits;
+ __u8 csum_type;
+ __u8 pad;
+
+ struct bch_extent_ptr ptrs[0];
+} __attribute__((packed, aligned(8)));
+
+/* Reflink: */
+
+struct bch_reflink_p {
+ struct bch_val v;
+ __le64 idx;
+
+ __le32 reservation_generation;
+ __u8 nr_replicas;
+ __u8 pad[3];
+};
+
+struct bch_reflink_v {
+ struct bch_val v;
+ __le64 refcount;
+ union bch_extent_entry start[0];
+ __u64 _data[0];
+};
+
+struct bch_indirect_inline_data {
+ struct bch_val v;
+ __le64 refcount;
+ u8 data[0];
+};
+
+/* Inline data */
+
+struct bch_inline_data {
+ struct bch_val v;
+ u8 data[0];
+};
/* Optional/variable size superblock sections: */
x(journal, 0) \
x(members, 1) \
x(crypt, 2) \
- x(replicas, 3) \
+ x(replicas_v0, 3) \
x(quota, 4) \
- x(disk_groups, 5)
+ x(disk_groups, 5) \
+ x(clean, 6) \
+ x(replicas, 7) \
+ x(journal_seq_blacklist, 8)
enum bch_sb_field_type {
#define x(f, nr) BCH_SB_FIELD_##f = nr,
/* BCH_SB_FIELD_members: */
+#define BCH_MIN_NR_NBUCKETS (1 << 6)
+
struct bch_member {
uuid_le uuid;
__le64 nbuckets; /* device size */
LE64_BITMASK(BCH_MEMBER_NR_WRITE_ERRORS,struct bch_member, flags[1], 20, 40);
#endif
+#define BCH_MEMBER_STATES() \
+ x(rw, 0) \
+ x(ro, 1) \
+ x(failed, 2) \
+ x(spare, 3)
+
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 x(t, n) BCH_MEMBER_STATE_##t = n,
+ BCH_MEMBER_STATES()
+#undef x
+ BCH_MEMBER_STATE_NR
};
-enum cache_replacement {
- CACHE_REPLACEMENT_LRU = 0,
- CACHE_REPLACEMENT_FIFO = 1,
- CACHE_REPLACEMENT_RANDOM = 2,
- CACHE_REPLACEMENT_NR = 3,
+#define BCH_CACHE_REPLACEMENT_POLICIES() \
+ x(lru, 0) \
+ x(fifo, 1) \
+ x(random, 2)
+
+enum bch_cache_replacement_policies {
+#define x(t, n) BCH_CACHE_REPLACEMENT_##t = n,
+ BCH_CACHE_REPLACEMENT_POLICIES()
+#undef x
+ BCH_CACHE_REPLACEMENT_NR
};
struct bch_sb_field_members {
/* 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 {
+ __u8 data_type;
+ __u8 nr_devs;
+ __u8 devs[0];
+} __attribute__((packed));
+
+struct bch_sb_field_replicas_v0 {
+ struct bch_sb_field field;
+ struct bch_replicas_entry_v0 entries[0];
+} __attribute__((packed, aligned(8)));
+
struct bch_replicas_entry {
- u8 data_type;
- u8 nr;
- u8 devs[0];
-};
+ __u8 data_type;
+ __u8 nr_devs;
+ __u8 nr_required;
+ __u8 devs[0];
+} __attribute__((packed));
+
+#define replicas_entry_bytes(_i) \
+ (offsetof(typeof(*(_i)), devs) + (_i)->nr_devs)
struct bch_sb_field_replicas {
struct bch_sb_field field;
struct bch_replicas_entry entries[0];
-};
+} __attribute__((packed, aligned(8)));
/* BCH_SB_FIELD_quota: */
struct bch_disk_group {
__u8 label[BCH_SB_LABEL_SIZE];
__le64 flags[2];
-};
+} __attribute__((packed, aligned(8)));
LE64_BITMASK(BCH_GROUP_DELETED, struct bch_disk_group, flags[0], 0, 1)
LE64_BITMASK(BCH_GROUP_DATA_ALLOWED, struct bch_disk_group, flags[0], 1, 6)
struct bch_sb_field_disk_groups {
struct bch_sb_field field;
struct bch_disk_group entries[0];
+} __attribute__((packed, aligned(8)));
+
+/*
+ * On clean shutdown, store btree roots and current journal sequence number in
+ * the superblock:
+ */
+struct jset_entry {
+ __le16 u64s;
+ __u8 btree_id;
+ __u8 level;
+ __u8 type; /* designates what this jset holds */
+ __u8 pad[3];
+
+ union {
+ struct bkey_i start[0];
+ __u64 _data[0];
+ };
+};
+
+struct bch_sb_field_clean {
+ struct bch_sb_field field;
+
+ __le32 flags;
+ __le16 _read_clock; /* no longer used */
+ __le16 _write_clock;
+ __le64 journal_seq;
+
+ union {
+ struct jset_entry start[0];
+ __u64 _data[0];
+ };
+};
+
+struct journal_seq_blacklist_entry {
+ __le64 start;
+ __le64 end;
+};
+
+struct bch_sb_field_journal_seq_blacklist {
+ struct bch_sb_field field;
+
+ union {
+ struct journal_seq_blacklist_entry start[0];
+ __u64 _data[0];
+ };
};
/* Superblock: */
/*
- * Version 8: BCH_SB_ENCODED_EXTENT_MAX_BITS
- * BCH_MEMBER_DATA_ALLOWED
- * Version 9: incompatible extent nonce change
+ * New versioning scheme:
+ * One common version number for all on disk data structures - superblock, btree
+ * nodes, journal entries
*/
+#define BCH_JSET_VERSION_OLD 2
+#define BCH_BSET_VERSION_OLD 3
+
+enum bcachefs_metadata_version {
+ bcachefs_metadata_version_min = 9,
+ bcachefs_metadata_version_new_versioning = 10,
+ bcachefs_metadata_version_bkey_renumber = 10,
+ bcachefs_metadata_version_inode_btree_change = 11,
+ bcachefs_metadata_version_snapshot = 12,
+ bcachefs_metadata_version_inode_backpointers = 13,
+ bcachefs_metadata_version_max = 14,
+};
-#define BCH_SB_VERSION_MIN 7
-#define BCH_SB_VERSION_EXTENT_MAX 8
-#define BCH_SB_VERSION_EXTENT_NONCE_V1 9
-#define BCH_SB_VERSION_MAX 9
+#define bcachefs_metadata_version_current (bcachefs_metadata_version_max - 1)
#define BCH_SB_SECTOR 8
#define BCH_SB_MEMBERS_MAX 64 /* XXX kill */
/*
* @offset - sector where this sb was written
* @version - on disk format version
+ * @version_min - Oldest metadata version this filesystem contains; so we can
+ * safely drop compatibility code and refuse to mount filesystems
+ * we'd need it for
* @magic - identifies as a bcachefs superblock (BCACHE_MAGIC)
* @seq - incremented each time superblock is written
* @uuid - used for generating various magic numbers and identifying
*/
struct bch_sb {
struct bch_csum csum;
- __le64 version;
+ __le16 version;
+ __le16 version_min;
+ __le16 pad[2];
uuid_le magic;
uuid_le uuid;
uuid_le user_uuid;
LE64_BITMASK(BCH_SB_GRPQUOTA, struct bch_sb, flags[0], 58, 59);
LE64_BITMASK(BCH_SB_PRJQUOTA, struct bch_sb, flags[0], 59, 60);
-/* 60-64 unused */
+LE64_BITMASK(BCH_SB_HAS_ERRORS, struct bch_sb, flags[0], 60, 61);
+
+/* bit 61 was reflink option */
+LE64_BITMASK(BCH_SB_BIG_ENDIAN, struct bch_sb, flags[0], 62, 63);
+
+/* 61-64 unused */
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);
LE64_BITMASK(BCH_SB_BACKGROUND_COMPRESSION_TYPE,
struct bch_sb, flags[2], 0, 4);
+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);
-/* Features: */
-enum bch_sb_features {
- BCH_FEATURE_LZ4 = 0,
- BCH_FEATURE_GZIP = 1,
- BCH_FEATURE_ZSTD = 2,
+/*
+ * Features:
+ *
+ * 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_extent_overwrite: gates BTREE_NODE_NEW_EXTENT_OVERWRITE
+ */
+#define BCH_SB_FEATURES() \
+ x(lz4, 0) \
+ x(gzip, 1) \
+ x(zstd, 2) \
+ x(atomic_nlink, 3) \
+ x(ec, 4) \
+ x(journal_seq_blacklist_v3, 5) \
+ x(reflink, 6) \
+ x(new_siphash, 7) \
+ x(inline_data, 8) \
+ x(new_extent_overwrite, 9) \
+ x(incompressible, 10) \
+ x(btree_ptr_v2, 11) \
+ x(extents_above_btree_updates, 12) \
+ 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 \
+ (BCH_SB_FEATURES_ALWAYS| \
+ (1ULL << BCH_FEATURE_new_siphash)| \
+ (1ULL << BCH_FEATURE_btree_ptr_v2)| \
+ (1ULL << BCH_FEATURE_new_varint)| \
+ (1ULL << BCH_FEATURE_journal_no_flush))
+
+enum bch_sb_feature {
+#define x(f, n) BCH_FEATURE_##f,
+ BCH_SB_FEATURES()
+#undef x
+ BCH_FEATURE_NR,
+};
+
+enum bch_sb_compat {
+ BCH_COMPAT_FEAT_ALLOC_INFO = 0,
+ BCH_COMPAT_FEAT_ALLOC_METADATA = 1,
+ BCH_COMPAT_FEAT_EXTENTS_ABOVE_BTREE_UPDATES_DONE = 2,
+ BCH_COMPAT_FEAT_BFORMAT_OVERFLOW_DONE = 3,
};
/* 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
};
-enum bch_csum_opts {
- BCH_CSUM_OPT_NONE = 0,
- BCH_CSUM_OPT_CRC32C = 1,
- BCH_CSUM_OPT_CRC64 = 2,
- BCH_CSUM_OPT_NR = 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 BCH_STR_HASH_OPTS() \
+ x(crc32c, 0) \
+ x(crc64, 1) \
+ x(siphash, 2)
+
enum bch_str_hash_opts {
- BCH_STR_HASH_CRC32C = 0,
- BCH_STR_HASH_CRC64 = 1,
- BCH_STR_HASH_SIPHASH = 2,
- BCH_STR_HASH_NR = 3,
+#define x(t, n) BCH_STR_HASH_OPT_##t = n,
+ BCH_STR_HASH_OPTS()
+#undef x
+ BCH_STR_HASH_OPT_NR
+};
+
+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,
+};
+
+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,
+};
+
+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:
+ return true;
+ default:
+ return false;
+ }
+}
+
+#define BCH_CSUM_OPTS() \
+ x(none, 0) \
+ x(crc32c, 1) \
+ x(crc64, 2)
+
+enum bch_csum_opts {
+#define x(t, n) BCH_CSUM_OPT_##t = n,
+ BCH_CSUM_OPTS()
+#undef x
+ BCH_CSUM_OPT_NR
};
#define BCH_COMPRESSION_TYPES() \
- x(NONE) \
- x(LZ4) \
- x(GZIP) \
- x(ZSTD)
+ x(none, 0) \
+ x(lz4_old, 1) \
+ x(gzip, 2) \
+ x(lz4, 3) \
+ x(zstd, 4) \
+ x(incompressible, 5)
-enum bch_compression_opts {
-#define x(t) BCH_COMPRESSION_OPT_##t,
+enum bch_compression_type {
+#define x(t, n) BCH_COMPRESSION_TYPE_##t = n,
BCH_COMPRESSION_TYPES()
+#undef x
+ BCH_COMPRESSION_TYPE_NR
+};
+
+#define BCH_COMPRESSION_OPTS() \
+ x(none, 0) \
+ x(lz4, 1) \
+ x(gzip, 2) \
+ x(zstd, 3)
+
+enum bch_compression_opts {
+#define x(t, n) BCH_COMPRESSION_OPT_##t = n,
+ BCH_COMPRESSION_OPTS()
#undef x
BCH_COMPRESSION_OPT_NR
};
/* Journal */
-#define BCACHE_JSET_VERSION_UUIDv1 1
-#define BCACHE_JSET_VERSION_UUID 1 /* Always latest UUID format */
-#define BCACHE_JSET_VERSION_JKEYS 2
-#define BCACHE_JSET_VERSION 2
-
-struct jset_entry {
- __le16 u64s;
- __u8 btree_id;
- __u8 level;
- __u8 type; /* designates what this jset holds */
- __u8 pad[3];
-
- union {
- struct bkey_i start[0];
- __u64 _data[0];
- };
-};
-
#define JSET_KEYS_U64s (sizeof(struct jset_entry) / sizeof(__u64))
#define BCH_JSET_ENTRY_TYPES() \
x(btree_root, 1) \
x(prio_ptrs, 2) \
x(blacklist, 3) \
- x(blacklist_v2, 4)
+ x(blacklist_v2, 4) \
+ x(usage, 5) \
+ x(data_usage, 6) \
+ x(clock, 7) \
+ x(dev_usage, 8)
enum {
#define x(f, nr) BCH_JSET_ENTRY_##f = nr,
__le64 end;
};
+enum {
+ FS_USAGE_RESERVED = 0,
+ FS_USAGE_INODES = 1,
+ FS_USAGE_KEY_VERSION = 2,
+ FS_USAGE_NR = 3
+};
+
+struct jset_entry_usage {
+ struct jset_entry entry;
+ __le64 v;
+} __attribute__((packed));
+
+struct jset_entry_data_usage {
+ struct jset_entry entry;
+ __le64 v;
+ 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));
+
/*
* On disk format for a journal entry:
* seq is monotonically increasing; every journal entry has its own unique
__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;
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 20
+#define BCH_JOURNAL_BUCKETS_MIN 8
/* Btree: */
-#define DEFINE_BCH_BTREE_IDS() \
- DEF_BTREE_ID(EXTENTS, 0, "extents") \
- DEF_BTREE_ID(INODES, 1, "inodes") \
- DEF_BTREE_ID(DIRENTS, 2, "dirents") \
- DEF_BTREE_ID(XATTRS, 3, "xattrs") \
- DEF_BTREE_ID(ALLOC, 4, "alloc") \
- DEF_BTREE_ID(QUOTAS, 5, "quotas")
-
-#define DEF_BTREE_ID(kwd, val, name) BTREE_ID_##kwd = val,
+#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)
enum btree_id {
- DEFINE_BCH_BTREE_IDS()
+#define x(kwd, val) BTREE_ID_##kwd = val,
+ BCH_BTREE_IDS()
+#undef x
BTREE_ID_NR
};
-#undef DEF_BTREE_ID
-
#define BTREE_MAX_DEPTH 4U
/* Btree nodes */
-/* Version 1: Seed pointer into btree node checksum
- */
-#define BCACHE_BSET_CSUM 1
-#define BCACHE_BSET_KEY_v1 2
-#define BCACHE_BSET_JOURNAL_SEQ 3
-#define BCACHE_BSET_VERSION 3
-
/*
* Btree nodes
*
/* 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 {
LE64_BITMASK(BTREE_NODE_ID, struct btree_node, flags, 0, 4);
LE64_BITMASK(BTREE_NODE_LEVEL, struct btree_node, flags, 4, 8);
-/* 8-32 unused */
+LE64_BITMASK(BTREE_NODE_NEW_EXTENT_OVERWRITE,
+ struct btree_node, flags, 8, 9);
+/* 9-32 unused */
LE64_BITMASK(BTREE_NODE_SEQ, struct btree_node, flags, 32, 64);
struct btree_node_entry {
};
} __attribute__((packed, aligned(8)));
-/* Obsolete: */
-
-struct prio_set {
- struct bch_csum csum;
-
- __le64 magic;
- __le32 nonce[3];
- __le16 version;
- __le16 flags;
-
- __u8 encrypted_start[0];
-
- __le64 next_bucket;
-
- struct bucket_disk {
- __le16 prio[2];
- __u8 gen;
- } __attribute__((packed)) data[];
-} __attribute__((packed, aligned(8)));
-
-LE32_BITMASK(PSET_CSUM_TYPE, struct prio_set, flags, 0, 4);
-
-#define PSET_MAGIC __cpu_to_le64(0x6750e15f87337f91ULL)
-
-static inline __u64 __pset_magic(struct bch_sb *sb)
-{
- return __le64_to_cpu(__bch2_sb_magic(sb) ^ PSET_MAGIC);
-}
-
#endif /* _BCACHEFS_FORMAT_H */
projects / bcachefs-tools-debian / blobdiff