Introduce a new tree, dedup tree to record on-disk dedup hash.
As a persist hash storage instead of in-memeory only implement.
Unlike Liu Bo's implement, in this version we won't do hack for
bytenr -> hash search, but add a new type, DEDUP_BYTENR_ITEM for such
search case, just like in-memory backend.
Signed-off-by: Liu Bo <bo.li....@oracle.com>
Signed-off-by: Wang Xiaoguang <wangxg.f...@cn.fujitsu.com>
Signed-off-by: Qu Wenruo <quwen...@cn.fujitsu.com>
---
fs/btrfs/ctree.h | 67 +++++++++++++++++++++++++++++++++++++++++++-
fs/btrfs/dedup.h | 8 ++++++
fs/btrfs/disk-io.c | 1 +
include/trace/events/btrfs.h | 3 +-
4 files changed, 77 insertions(+), 2 deletions(-)
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 62fed1d..60a10e4 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -96,6 +96,9 @@ struct btrfs_ordered_sum;
/* for storing items that use the BTRFS_UUID_KEY* types */
#define BTRFS_UUID_TREE_OBJECTID 9ULL
+/* on-disk dedup tree (EXPERIMENTAL) */
+#define BTRFS_DEDUP_TREE_OBJECTID 10ULL
+
/* for storing balance parameters in the root tree */
#define BTRFS_BALANCE_OBJECTID -4ULL
@@ -523,10 +526,12 @@ struct btrfs_super_block {
#define BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA (1ULL << 8)
#define BTRFS_FEATURE_INCOMPAT_NO_HOLES (1ULL << 9)
+#define BTRFS_FEATURE_COMPAT_RO_DEDUP (1ULL << 0)
#define BTRFS_FEATURE_COMPAT_SUPP 0ULL
#define BTRFS_FEATURE_COMPAT_SAFE_SET 0ULL
#define BTRFS_FEATURE_COMPAT_SAFE_CLEAR 0ULL
-#define BTRFS_FEATURE_COMPAT_RO_SUPP 0ULL
+#define BTRFS_FEATURE_COMPAT_RO_SUPP \
+ (BTRFS_FEATURE_COMPAT_RO_DEDUP)
#define BTRFS_FEATURE_COMPAT_RO_SAFE_SET 0ULL
#define BTRFS_FEATURE_COMPAT_RO_SAFE_CLEAR 0ULL
@@ -954,6 +959,46 @@ struct btrfs_csum_item {
u8 csum;
} __attribute__ ((__packed__));
+/*
+ * Objectid: 0
+ * Type: BTRFS_DEDUP_STATUS_ITEM_KEY
+ * Offset: 0
+ */
+struct btrfs_dedup_status_item {
+ __le64 blocksize;
+ __le64 limit_nr;
+ __le16 hash_type;
+ __le16 backend;
+} __attribute__ ((__packed__));
+
+/*
+ * Objectid: Last 64 bit of the hash
+ * Type: BTRFS_DEDUP_HASH_ITEM_KEY
+ * Offset: Bytenr of the hash
+ *
+ * Used for hash <-> bytenr search
+ * XXX: On-disk format not stable yet, see the unsed one
+ */
+struct btrfs_dedup_hash_item {
+ /* on disk length of dedup range */
+ __le64 len;
+
+ /* Spare space */
+ u8 __unused[16];
+
+ /* Hash follows */
+} __attribute__ ((__packed__));
+
+/*
+ * Objectid: bytenr
+ * Type: BTRFS_DEDUP_BYTENR_ITEM_KEY
+ * offset: Last 64 bit of the hash
+ *
+ * Used for bytenr <-> hash search (for free_extent)
+ * all its content is hash.
+ * So no special item struct is needed.
+ */
+
struct btrfs_dev_stats_item {
/*
* grow this item struct at the end for future enhancements and keep
@@ -2100,6 +2145,13 @@ struct btrfs_ioctl_defrag_range_args {
#define BTRFS_CHUNK_ITEM_KEY 228
/*
+ * Dedup item and status
+ */
+#define BTRFS_DEDUP_STATUS_ITEM_KEY 230
+#define BTRFS_DEDUP_HASH_ITEM_KEY 231
+#define BTRFS_DEDUP_BYTENR_ITEM_KEY 232
+
+/*
* Records the overall state of the qgroups.
* There's only one instance of this key present,
* (0, BTRFS_QGROUP_STATUS_KEY, 0)
@@ -3157,6 +3209,19 @@ static inline unsigned long btrfs_leaf_data(struct
extent_buffer *l)
return offsetof(struct btrfs_leaf, items);
}
+/* btrfs_dedup_status */
+BTRFS_SETGET_FUNCS(dedup_status_blocksize, struct btrfs_dedup_status_item,
+ blocksize, 64);
+BTRFS_SETGET_FUNCS(dedup_status_limit, struct btrfs_dedup_status_item,
+ limit_nr, 64);
+BTRFS_SETGET_FUNCS(dedup_status_hash_type, struct btrfs_dedup_status_item,
+ hash_type, 16);
+BTRFS_SETGET_FUNCS(dedup_status_backend, struct btrfs_dedup_status_item,
+ backend, 16);
+
+/* btrfs_dedup_hash_item */
+BTRFS_SETGET_FUNCS(dedup_hash_len, struct btrfs_dedup_hash_item, len, 64);
+
/* struct btrfs_file_extent_item */
BTRFS_SETGET_FUNCS(file_extent_type, struct btrfs_file_extent_item, type, 8);
BTRFS_SETGET_STACK_FUNCS(stack_file_extent_disk_bytenr,
diff --git a/fs/btrfs/dedup.h b/fs/btrfs/dedup.h
index 3cd6d14..bc37a8d 100644
--- a/fs/btrfs/dedup.h
+++ b/fs/btrfs/dedup.h
@@ -59,6 +59,8 @@ struct btrfs_dedup_hash {
u8 hash[];
};
+struct btrfs_root;
+
struct btrfs_dedup_info {
/* dedup blocksize */
u64 blocksize;
@@ -75,6 +77,12 @@ struct btrfs_dedup_info {
spinlock_t lock;
u64 limit_nr;
u64 current_nr;
+
+ /* for persist data like dedup-hash and dedup status */
+ struct btrfs_root *dedup_root;
+
+ /* on-disk mode only mutex */
+ struct mutex ondisk_lock;
};
struct btrfs_trans_handle;
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index b213439..cf9dbff 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -175,6 +175,7 @@ static struct btrfs_lockdep_keyset {
{ .id = BTRFS_TREE_RELOC_OBJECTID, .name_stem = "treloc" },
{ .id = BTRFS_DATA_RELOC_TREE_OBJECTID, .name_stem = "dreloc" },
{ .id = BTRFS_UUID_TREE_OBJECTID, .name_stem = "uuid" },
+ { .id = BTRFS_DEDUP_TREE_OBJECTID, .name_stem = "dedup" },
{ .id = 0, .name_stem = "tree" },
};
diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h
index b4473da..655b630 100644
--- a/include/trace/events/btrfs.h
+++ b/include/trace/events/btrfs.h
@@ -46,12 +46,13 @@ struct btrfs_qgroup_operation;
{ BTRFS_QUOTA_TREE_OBJECTID, "QUOTA_TREE" }, \
{ BTRFS_TREE_RELOC_OBJECTID, "TREE_RELOC" }, \
{ BTRFS_UUID_TREE_OBJECTID, "UUID_RELOC" }, \
+ { BTRFS_DEDUP_TREE_OBJECTID, "DEDUP_TREE" }, \
{ BTRFS_DATA_RELOC_TREE_OBJECTID, "DATA_RELOC_TREE" })
#define show_root_type(obj) \
obj, ((obj >= BTRFS_DATA_RELOC_TREE_OBJECTID) || \
(obj >= BTRFS_ROOT_TREE_OBJECTID && \
- obj <= BTRFS_QUOTA_TREE_OBJECTID)) ? __show_root_type(obj) : "-"
+ obj <= BTRFS_DEDUP_TREE_OBJECTID)) ? __show_root_type(obj) : "-"
#define BTRFS_GROUP_FLAGS \
{ BTRFS_BLOCK_GROUP_DATA, "DATA"}, \
--
2.6.4
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
