xfs
[Top] [All Lists]

[PATCH v2 04/17] xfs: introduce inode record hole mask for sparse inode

To: xfs@xxxxxxxxxxx
Subject: [PATCH v2 04/17] xfs: introduce inode record hole mask for sparse inode chunks
From: Brian Foster <bfoster@xxxxxxxxxx>
Date: Mon, 3 Nov 2014 11:12:13 -0500
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <1415031146-9107-1-git-send-email-bfoster@xxxxxxxxxx>
References: <1415031146-9107-1-git-send-email-bfoster@xxxxxxxxxx>
The inode btrees track 64 inodes per record, regardless of inode size.
Thus, inode chunks on disk vary in size depending on the size of the
inodes. This creates a contiguous allocation requirement for new inode
chunks that can be difficult to satisfy on an aged and fragmented (free
space) filesystem.

The inode record freecount currently uses 4 bytes on disk to track the
free inode count. With a maximum freecount value of 64, only one byte is
required. Convert the freecount field to a single byte and reserve two
of the remaining 3 higher order bytes left to the hole mask field.
The left over byte to the total count field.

The hole mask field tracks holes in the chunks of physical space that
the inode record refers to. This facilitates the sparse allocation of
inode chunks when contiguous chunks are not available and allows the
inode btrees to identify what portions of the chunk contain valid
inodes. The total count field contains the total number of valid inodes
referred to by the record. This can also be deduced from the hole mask.
The count field provides clarity and redundancy for internal record
verification.

Note that both fields are initialized to zero to maintain backwards
compatibility with existing filesystems (e.g., the higher order bytes of
freecount are always 0). Tracking holes means that the hole mask is
initialized to zero and thus remains "valid" in accordance with a
non-sparse inode fs when no sparse chunks are physically allocated.
Update the inode record management functions to handle the new fields
and initialize to zero.

[XXX: The count field breaks this notion of easy backwards compatibility
with !sparseinode fs. Should we reconsider the addition of total count
or the idea of converting back and forth between sparse inode fs with a
simple feature bit? I lean towards the latter at the moment as we could
implement conversion in xfs_repair if there is ever user demand.]

Signed-off-by: Brian Foster <bfoster@xxxxxxxxxx>
---
 fs/xfs/libxfs/xfs_format.h       |  8 ++++++--
 fs/xfs/libxfs/xfs_ialloc.c       | 12 ++++++++++--
 fs/xfs/libxfs/xfs_ialloc_btree.c |  4 +++-
 3 files changed, 19 insertions(+), 5 deletions(-)

diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 7e42bba..873c4d1 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -217,13 +217,17 @@ static inline xfs_inofree_t xfs_inobt_maskn(int i, int n)
  */
 typedef struct xfs_inobt_rec {
        __be32          ir_startino;    /* starting inode number */
-       __be32          ir_freecount;   /* count of free inodes (set bits) */
+       __be16          ir_holemask;    /* hole mask for sparse chunks */
+       __u8            ir_count;       /* total inode count */
+       __u8            ir_freecount;   /* count of free inodes (set bits) */
        __be64          ir_free;        /* free inode mask */
 } xfs_inobt_rec_t;
 
 typedef struct xfs_inobt_rec_incore {
        xfs_agino_t     ir_startino;    /* starting inode number */
-       __int32_t       ir_freecount;   /* count of free inodes (set bits) */
+       __uint16_t      ir_holemask;    /* hole mask for sparse chunks */
+       __uint8_t       ir_count;       /* total inode count */
+       __uint8_t       ir_freecount;   /* count of free inodes (set bits) */
        xfs_inofree_t   ir_free;        /* free inode mask */
 } xfs_inobt_rec_incore_t;
 
diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c
index 89949a3..fcc8baa 100644
--- a/fs/xfs/libxfs/xfs_ialloc.c
+++ b/fs/xfs/libxfs/xfs_ialloc.c
@@ -68,6 +68,8 @@ xfs_inobt_lookup(
        int                     *stat)  /* success/failure */
 {
        cur->bc_rec.i.ir_startino = ino;
+       cur->bc_rec.i.ir_holemask = 0;
+       cur->bc_rec.i.ir_count = 0;
        cur->bc_rec.i.ir_freecount = 0;
        cur->bc_rec.i.ir_free = 0;
        return xfs_btree_lookup(cur, dir, stat);
@@ -85,7 +87,9 @@ xfs_inobt_update(
        union xfs_btree_rec     rec;
 
        rec.inobt.ir_startino = cpu_to_be32(irec->ir_startino);
-       rec.inobt.ir_freecount = cpu_to_be32(irec->ir_freecount);
+       rec.inobt.ir_holemask = cpu_to_be16(irec->ir_holemask);
+       rec.inobt.ir_count = irec->ir_count;
+       rec.inobt.ir_freecount = irec->ir_freecount;
        rec.inobt.ir_free = cpu_to_be64(irec->ir_free);
        return xfs_btree_update(cur, &rec);
 }
@@ -105,7 +109,9 @@ xfs_inobt_get_rec(
        error = xfs_btree_get_rec(cur, &rec, stat);
        if (!error && *stat == 1) {
                irec->ir_startino = be32_to_cpu(rec->inobt.ir_startino);
-               irec->ir_freecount = be32_to_cpu(rec->inobt.ir_freecount);
+               irec->ir_holemask = be16_to_cpu(rec->inobt.ir_holemask);
+               irec->ir_count = rec->inobt.ir_count;
+               irec->ir_freecount = rec->inobt.ir_freecount;
                irec->ir_free = be64_to_cpu(rec->inobt.ir_free);
        }
        return error;
@@ -121,6 +127,8 @@ xfs_inobt_insert_rec(
        xfs_inofree_t           free,
        int                     *stat)
 {
+       cur->bc_rec.i.ir_holemask = 0;
+       cur->bc_rec.i.ir_count = 0; /* zero for backwards compatibility */
        cur->bc_rec.i.ir_freecount = freecount;
        cur->bc_rec.i.ir_free = free;
        return xfs_btree_insert(cur, stat);
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c
index c9b06f3..b2a9117 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.c
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.c
@@ -169,7 +169,9 @@ xfs_inobt_init_rec_from_cur(
        union xfs_btree_rec     *rec)
 {
        rec->inobt.ir_startino = cpu_to_be32(cur->bc_rec.i.ir_startino);
-       rec->inobt.ir_freecount = cpu_to_be32(cur->bc_rec.i.ir_freecount);
+       rec->inobt.ir_holemask = cpu_to_be16(cur->bc_rec.i.ir_holemask);
+       rec->inobt.ir_count = cur->bc_rec.i.ir_count;
+       rec->inobt.ir_freecount = cur->bc_rec.i.ir_freecount;
        rec->inobt.ir_free = cpu_to_be64(cur->bc_rec.i.ir_free);
 }
 
-- 
1.8.3.1

<Prev in Thread] Current Thread [Next in Thread>