OSDN Git Service

[XFS] add helpers for addressing entities inside a btree block
authorChristoph Hellwig <hch@infradead.org>
Thu, 30 Oct 2008 05:55:34 +0000 (16:55 +1100)
committerLachlan McIlroy <lachlan@sgi.com>
Thu, 30 Oct 2008 05:55:34 +0000 (16:55 +1100)
Add new helpers in xfs_btree.c to find the record, key and block pointer
entries inside a btree block. To implement this genericly the
->get_maxrecs methods and two new xfs_btree_ops entries for the key and
record sizes are used. Also add a big comment describing how the
addressing inside a btree block works.

Note that these helpers are unused until users are introduced in the next
patches and this patch will thus cause some harmless compiler warnings.

SGI-PV: 985583

SGI-Modid: xfs-linux-melb:xfs-kern:32189a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Bill O'Donnell <billodo@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
fs/xfs/xfs_alloc_btree.c
fs/xfs/xfs_bmap_btree.c
fs/xfs/xfs_btree.c
fs/xfs/xfs_btree.h
fs/xfs/xfs_ialloc_btree.c

index 1f268b6..9e2421c 100644 (file)
@@ -2294,6 +2294,9 @@ xfs_allocbt_trace_record(
 #endif /* XFS_BTREE_TRACE */
 
 static const struct xfs_btree_ops xfs_allocbt_ops = {
+       .rec_len                = sizeof(xfs_alloc_rec_t),
+       .key_len                = sizeof(xfs_alloc_key_t),
+
        .dup_cursor             = xfs_allocbt_dup_cursor,
        .get_maxrecs            = xfs_allocbt_get_maxrecs,
 
index bdcfbea..a71010a 100644 (file)
@@ -2509,6 +2509,9 @@ xfs_bmbt_trace_record(
 #endif /* XFS_BTREE_TRACE */
 
 static const struct xfs_btree_ops xfs_bmbt_ops = {
+       .rec_len                = sizeof(xfs_bmbt_rec_t),
+       .key_len                = sizeof(xfs_bmbt_key_t),
+
        .dup_cursor             = xfs_bmbt_dup_cursor,
        .get_maxrecs            = xfs_bmbt_get_maxrecs,
 
index 893e86f..4aec7c7 100644 (file)
@@ -403,6 +403,136 @@ xfs_btree_dup_cursor(
 }
 
 /*
+ * XFS btree block layout and addressing:
+ *
+ * There are two types of blocks in the btree: leaf and non-leaf blocks.
+ *
+ * The leaf record start with a header then followed by records containing
+ * the values.  A non-leaf block also starts with the same header, and
+ * then first contains lookup keys followed by an equal number of pointers
+ * to the btree blocks at the previous level.
+ *
+ *             +--------+-------+-------+-------+-------+-------+-------+
+ * Leaf:       | header | rec 1 | rec 2 | rec 3 | rec 4 | rec 5 | rec N |
+ *             +--------+-------+-------+-------+-------+-------+-------+
+ *
+ *             +--------+-------+-------+-------+-------+-------+-------+
+ * Non-Leaf:   | header | key 1 | key 2 | key N | ptr 1 | ptr 2 | ptr N |
+ *             +--------+-------+-------+-------+-------+-------+-------+
+ *
+ * The header is called struct xfs_btree_block for reasons better left unknown
+ * and comes in different versions for short (32bit) and long (64bit) block
+ * pointers.  The record and key structures are defined by the btree instances
+ * and opaque to the btree core.  The block pointers are simple disk endian
+ * integers, available in a short (32bit) and long (64bit) variant.
+ *
+ * The helpers below calculate the offset of a given record, key or pointer
+ * into a btree block (xfs_btree_*_offset) or return a pointer to the given
+ * record, key or pointer (xfs_btree_*_addr).  Note that all addressing
+ * inside the btree block is done using indices starting at one, not zero!
+ */
+
+/*
+ * Return size of the btree block header for this btree instance.
+ */
+static inline size_t xfs_btree_block_len(struct xfs_btree_cur *cur)
+{
+       return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+               sizeof(struct xfs_btree_lblock) :
+               sizeof(struct xfs_btree_sblock);
+}
+
+/*
+ * Return size of btree block pointers for this btree instance.
+ */
+static inline size_t xfs_btree_ptr_len(struct xfs_btree_cur *cur)
+{
+       return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+               sizeof(__be64) : sizeof(__be32);
+}
+
+/*
+ * Calculate offset of the n-th record in a btree block.
+ */
+STATIC size_t
+xfs_btree_rec_offset(
+       struct xfs_btree_cur    *cur,
+       int                     n)
+{
+       return xfs_btree_block_len(cur) +
+               (n - 1) * cur->bc_ops->rec_len;
+}
+
+/*
+ * Calculate offset of the n-th key in a btree block.
+ */
+STATIC size_t
+xfs_btree_key_offset(
+       struct xfs_btree_cur    *cur,
+       int                     n)
+{
+       return xfs_btree_block_len(cur) +
+               (n - 1) * cur->bc_ops->key_len;
+}
+
+/*
+ * Calculate offset of the n-th block pointer in a btree block.
+ */
+STATIC size_t
+xfs_btree_ptr_offset(
+       struct xfs_btree_cur    *cur,
+       int                     n,
+       int                     level)
+{
+       return xfs_btree_block_len(cur) +
+               cur->bc_ops->get_maxrecs(cur, level) * cur->bc_ops->key_len +
+               (n - 1) * xfs_btree_ptr_len(cur);
+}
+
+/*
+ * Return a pointer to the n-th record in the btree block.
+ */
+STATIC union xfs_btree_rec *
+xfs_btree_rec_addr(
+       struct xfs_btree_cur    *cur,
+       int                     n,
+       struct xfs_btree_block  *block)
+{
+       return (union xfs_btree_rec *)
+               ((char *)block + xfs_btree_rec_offset(cur, n));
+}
+
+/*
+ * Return a pointer to the n-th key in the btree block.
+ */
+STATIC union xfs_btree_key *
+xfs_btree_key_addr(
+       struct xfs_btree_cur    *cur,
+       int                     n,
+       struct xfs_btree_block  *block)
+{
+       return (union xfs_btree_key *)
+               ((char *)block + xfs_btree_key_offset(cur, n));
+}
+
+/*
+ * Return a pointer to the n-th block pointer in the btree block.
+ */
+STATIC union xfs_btree_ptr *
+xfs_btree_ptr_addr(
+       struct xfs_btree_cur    *cur,
+       int                     n,
+       struct xfs_btree_block  *block)
+{
+       int                     level = xfs_btree_get_level(block);
+
+       ASSERT(block->bb_level != 0);
+
+       return (union xfs_btree_ptr *)
+               ((char *)block + xfs_btree_ptr_offset(cur, n, level));
+}
+
+/*
  * Get a the root block which is stored in the inode.
  *
  * For now this btree implementation assumes the btree root is always
index 5398cd0..593f82b 100644 (file)
@@ -180,6 +180,10 @@ do {    \
 #define        XFS_BTREE_MAXLEVELS     8       /* max of all btrees */
 
 struct xfs_btree_ops {
+       /* size of the key and record structures */
+       size_t  key_len;
+       size_t  rec_len;
+
        /* cursor operations */
        struct xfs_btree_cur *(*dup_cursor)(struct xfs_btree_cur *);
 
@@ -497,6 +501,15 @@ xfs_btree_setbuf(
        int                     lev,    /* level in btree */
        struct xfs_buf          *bp);   /* new buffer to set */
 
+
+/*
+ * Helpers.
+ */
+static inline int xfs_btree_get_level(struct xfs_btree_block *block)
+{
+       return be16_to_cpu(block->bb_level);
+}
+
 #endif /* __KERNEL__ */
 
 
index 18867f1..fc6db94 100644 (file)
@@ -2160,6 +2160,9 @@ xfs_inobt_trace_record(
 #endif /* XFS_BTREE_TRACE */
 
 static const struct xfs_btree_ops xfs_inobt_ops = {
+       .rec_len                = sizeof(xfs_inobt_rec_t),
+       .key_len                = sizeof(xfs_inobt_key_t),
+
        .dup_cursor             = xfs_inobt_dup_cursor,
        .get_maxrecs            = xfs_inobt_get_maxrecs,