On Fri, Nov 01, 2019 at 03:06:50PM -0700, Christoph Hellwig wrote:
> All but two callers of the ->node_tree_p dir operation already have a
> xfs_da3_icnode_hdr from a previous call to xfs_da3_node_hdr_from_disk at
> hand. Add a pointer to the btree entries to struct xfs_da3_icnode_hdr
> to clean up this pattern. The two remaining callers now expand the
> whole header as well, but that isn't very expensive and not in a super
> hot path anyway.
>
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> ---
> fs/xfs/libxfs/xfs_attr_leaf.c | 6 ++--
> fs/xfs/libxfs/xfs_da_btree.c | 68 ++++++++++++++++-------------------
> fs/xfs/libxfs/xfs_da_btree.h | 1 +
> fs/xfs/libxfs/xfs_da_format.c | 21 -----------
> fs/xfs/libxfs/xfs_dir2.h | 2 --
> fs/xfs/scrub/dabtree.c | 6 ++--
> fs/xfs/xfs_attr_inactive.c | 34 +++++++++---------
> fs/xfs/xfs_attr_list.c | 2 +-
> 8 files changed, 55 insertions(+), 85 deletions(-)
>
<snip>
> diff --git a/fs/xfs/libxfs/xfs_da_btree.h b/fs/xfs/libxfs/xfs_da_btree.h
> index 69ebf6a50d85..63ed45057fa5 100644
> --- a/fs/xfs/libxfs/xfs_da_btree.h
> +++ b/fs/xfs/libxfs/xfs_da_btree.h
> @@ -135,6 +135,7 @@ struct xfs_da3_icnode_hdr {
> uint16_t magic;
> uint16_t count;
> uint16_t level;
> + struct xfs_da_node_entry *btree;
This adds to the incore node header structure a pointer to raw disk
structures, right? Can we make this a little more explicit by naming
the field "raw_entries" or something?
--D
> };
>
> /*
> diff --git a/fs/xfs/libxfs/xfs_da_format.c b/fs/xfs/libxfs/xfs_da_format.c
> index 912096416a86..f896d37c845f 100644
> --- a/fs/xfs/libxfs/xfs_da_format.c
> +++ b/fs/xfs/libxfs/xfs_da_format.c
> @@ -494,22 +494,6 @@ xfs_dir3_leaf_hdr_to_disk(
> hdr3->stale = cpu_to_be16(from->stale);
> }
>
> -
> -/*
> - * Directory/Attribute Node block operations
> - */
> -static struct xfs_da_node_entry *
> -xfs_da2_node_tree_p(struct xfs_da_intnode *dap)
> -{
> - return dap->__btree;
> -}
> -
> -static struct xfs_da_node_entry *
> -xfs_da3_node_tree_p(struct xfs_da_intnode *dap)
> -{
> - return ((struct xfs_da3_intnode *)dap)->__btree;
> -}
> -
> /*
> * Directory free space block operations
> */
> @@ -669,7 +653,6 @@ static const struct xfs_dir_ops xfs_dir2_ops = {
> .leaf_ents_p = xfs_dir2_leaf_ents_p,
>
> .node_hdr_size = sizeof(struct xfs_da_node_hdr),
> - .node_tree_p = xfs_da2_node_tree_p,
>
> .free_hdr_size = sizeof(struct xfs_dir2_free_hdr),
> .free_hdr_to_disk = xfs_dir2_free_hdr_to_disk,
> @@ -717,7 +700,6 @@ static const struct xfs_dir_ops xfs_dir2_ftype_ops = {
> .leaf_ents_p = xfs_dir2_leaf_ents_p,
>
> .node_hdr_size = sizeof(struct xfs_da_node_hdr),
> - .node_tree_p = xfs_da2_node_tree_p,
>
> .free_hdr_size = sizeof(struct xfs_dir2_free_hdr),
> .free_hdr_to_disk = xfs_dir2_free_hdr_to_disk,
> @@ -765,7 +747,6 @@ static const struct xfs_dir_ops xfs_dir3_ops = {
> .leaf_ents_p = xfs_dir3_leaf_ents_p,
>
> .node_hdr_size = sizeof(struct xfs_da3_node_hdr),
> - .node_tree_p = xfs_da3_node_tree_p,
>
> .free_hdr_size = sizeof(struct xfs_dir3_free_hdr),
> .free_hdr_to_disk = xfs_dir3_free_hdr_to_disk,
> @@ -778,12 +759,10 @@ static const struct xfs_dir_ops xfs_dir3_ops = {
>
> static const struct xfs_dir_ops xfs_dir2_nondir_ops = {
> .node_hdr_size = sizeof(struct xfs_da_node_hdr),
> - .node_tree_p = xfs_da2_node_tree_p,
> };
>
> static const struct xfs_dir_ops xfs_dir3_nondir_ops = {
> .node_hdr_size = sizeof(struct xfs_da3_node_hdr),
> - .node_tree_p = xfs_da3_node_tree_p,
> };
>
> /*
> diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h
> index c16efeae0f2b..6eee4c1b20da 100644
> --- a/fs/xfs/libxfs/xfs_dir2.h
> +++ b/fs/xfs/libxfs/xfs_dir2.h
> @@ -82,8 +82,6 @@ struct xfs_dir_ops {
> (*leaf_ents_p)(struct xfs_dir2_leaf *lp);
>
> int node_hdr_size;
> - struct xfs_da_node_entry *
> - (*node_tree_p)(struct xfs_da_intnode *dap);
>
> int free_hdr_size;
> void (*free_hdr_to_disk)(struct xfs_dir2_free *to,
> diff --git a/fs/xfs/scrub/dabtree.c b/fs/xfs/scrub/dabtree.c
> index be19a48716b7..52aa69743dfa 100644
> --- a/fs/xfs/scrub/dabtree.c
> +++ b/fs/xfs/scrub/dabtree.c
> @@ -83,10 +83,12 @@ xchk_da_btree_node_entry(
> int level)
> {
> struct xfs_da_state_blk *blk = &ds->state->path.blk[level];
> + struct xfs_da3_icnode_hdr hdr;
>
> ASSERT(blk->magic == XFS_DA_NODE_MAGIC);
>
> - return (void *)ds->dargs.dp->d_ops->node_tree_p(blk->bp->b_addr) +
> + xfs_da3_node_hdr_from_disk(ds->sc->mp, &hdr, blk->bp->b_addr);
> + return (void *)hdr.btree +
> (blk->index * sizeof(struct xfs_da_node_entry));
> }
>
> @@ -411,7 +413,7 @@ xchk_da_btree_block(
> blk->magic = XFS_DA_NODE_MAGIC;
> node = blk->bp->b_addr;
> xfs_da3_node_hdr_from_disk(ip->i_mount, &nodehdr, node);
> - btree = ip->d_ops->node_tree_p(node);
> + btree = nodehdr.btree;
> *pmaxrecs = nodehdr.count;
> blk->hashval = be32_to_cpu(btree[*pmaxrecs - 1].hashval);
> if (level == 0) {
> diff --git a/fs/xfs/xfs_attr_inactive.c b/fs/xfs/xfs_attr_inactive.c
> index c1ce06ccfab0..f6d8967f0a8e 100644
> --- a/fs/xfs/xfs_attr_inactive.c
> +++ b/fs/xfs/xfs_attr_inactive.c
> @@ -190,19 +190,17 @@ xfs_attr3_leaf_inactive(
> */
> STATIC int
> xfs_attr3_node_inactive(
> - struct xfs_trans **trans,
> - struct xfs_inode *dp,
> - struct xfs_buf *bp,
> - int level)
> + struct xfs_trans **trans,
> + struct xfs_inode *dp,
> + struct xfs_buf *bp,
> + int level)
> {
> - xfs_da_blkinfo_t *info;
> - xfs_da_intnode_t *node;
> - xfs_dablk_t child_fsb;
> - xfs_daddr_t parent_blkno, child_blkno;
> - int error, i;
> - struct xfs_buf *child_bp;
> - struct xfs_da_node_entry *btree;
> + struct xfs_da_blkinfo *info;
> + xfs_dablk_t child_fsb;
> + xfs_daddr_t parent_blkno, child_blkno;
> + struct xfs_buf *child_bp;
> struct xfs_da3_icnode_hdr ichdr;
> + int error, i;
>
> /*
> * Since this code is recursive (gasp!) we must protect ourselves.
> @@ -212,15 +210,13 @@ xfs_attr3_node_inactive(
> return -EFSCORRUPTED;
> }
>
> - node = bp->b_addr;
> - xfs_da3_node_hdr_from_disk(dp->i_mount, &ichdr, node);
> + xfs_da3_node_hdr_from_disk(dp->i_mount, &ichdr, bp->b_addr);
> parent_blkno = bp->b_bn;
> if (!ichdr.count) {
> xfs_trans_brelse(*trans, bp);
> return 0;
> }
> - btree = dp->d_ops->node_tree_p(node);
> - child_fsb = be32_to_cpu(btree[0].before);
> + child_fsb = be32_to_cpu(ichdr.btree[0].before);
> xfs_trans_brelse(*trans, bp); /* no locks for later trans */
>
> /*
> @@ -279,13 +275,15 @@ xfs_attr3_node_inactive(
> * child block number.
> */
> if (i + 1 < ichdr.count) {
> + struct xfs_da3_icnode_hdr phdr;
> +
> error = xfs_da3_node_read(*trans, dp, 0, parent_blkno,
> &bp, XFS_ATTR_FORK);
> if (error)
> return error;
> - node = bp->b_addr;
> - btree = dp->d_ops->node_tree_p(node);
> - child_fsb = be32_to_cpu(btree[i + 1].before);
> + xfs_da3_node_hdr_from_disk(dp->i_mount, &phdr,
> + bp->b_addr);
> + child_fsb = be32_to_cpu(phdr.btree[i + 1].before);
> xfs_trans_brelse(*trans, bp);
> }
> /*
> diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c
> index 5b5a3772ed24..032920952aa2 100644
> --- a/fs/xfs/xfs_attr_list.c
> +++ b/fs/xfs/xfs_attr_list.c
> @@ -254,7 +254,7 @@ xfs_attr_node_list_lookup(
> else
> expected_level--;
>
> - btree = dp->d_ops->node_tree_p(node);
> + btree = nodehdr.btree;
> for (i = 0; i < nodehdr.count; btree++, i++) {
> if (cursor->hashval <= be32_to_cpu(btree->hashval)) {
> cursor->blkno = be32_to_cpu(btree->before);
> --
> 2.20.1
>
