Re: [PATCH 3/3] xfs: track AGs with reclaimable inodes in per-ag radix tree

[Christoph Hellwig <hch@xxxxxxxxxxxxx>]

> + */
> +static struct xfs_perag *
> +xfs_inode_ag_iter_next_pag(
> +     struct xfs_mount        *mp,
> +     xfs_agnumber_t          *first,
> +     int                     tag)
> +{
> +     struct xfs_perag        *pag = NULL;
> +
> +     if (tag == XFS_ICI_RECLAIM_TAG) {
> +             int found;
> +             int ref;
> +
> +             spin_lock(&mp->m_perag_lock);
> +             found = radix_tree_gang_lookup_tag(&mp->m_perag_tree,
> +                             (void **)&pag, *first, 1, tag);
> +             if (found <= 0) {
> +                     spin_unlock(&mp->m_perag_lock);
> +                     return NULL;
> +             }
> +             *first = pag->pag_agno + 1;
> +             /* open coded pag reference increment */
> +             ref = atomic_inc_return(&pag->pag_ref);
> +             spin_unlock(&mp->m_perag_lock);
> +             trace_xfs_perag_get_reclaim(mp, pag->pag_agno, ref, _RET_IP_);
> +     } else {
> +             pag = xfs_perag_get(mp, *first);
> +             (*first)++;
> +     }

I wonder if we should just split the AG iterator for inode reclaim vs
the rest.  We now have this difference in addition to taking the per-AG
lock exclusive instead of shared.

Anyway, the patch looks good for now,


Reviewed-by: Christoph Hellwig <hch@xxxxxx>