Re: [PATCH 04/11] xfs: add xfs_ilock_attr_map_shared

[Dave Chinner <david@fromorbit.com>]

On Fri, Dec 06, 2013 at 12:30:10PM -0800, Christoph Hellwig wrote:
> Equivalent to xfs_ilock_data_map_shared, except for the attribute fork.
> 
> Make xfs_getbmap use it if called for the attribute fork instead of
> xfs_ilock_data_map_shared.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> 
> Index: xfs/fs/xfs/xfs_bmap_util.c
> ===================================================================
> --- xfs.orig/fs/xfs/xfs_bmap_util.c	2013-12-06 19:58:15.759137296 +0100
> +++ xfs/fs/xfs/xfs_bmap_util.c	2013-12-06 20:02:23.291132216 +0100
> @@ -617,22 +617,27 @@ xfs_getbmap(
>  		return XFS_ERROR(ENOMEM);
>  
>  	xfs_ilock(ip, XFS_IOLOCK_SHARED);
> -	if (whichfork == XFS_DATA_FORK && !(iflags & BMV_IF_DELALLOC)) {
> -		if (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size) {
> +	if (whichfork == XFS_DATA_FORK) {
> +		if (!(iflags & BMV_IF_DELALLOC) &&
> +		    (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size)) {
>  			error = -filemap_write_and_wait(VFS_I(ip)->i_mapping);
>  			if (error)
>  				goto out_unlock_iolock;
> +
> +			/*
> +			 * Even after flushing the inode, there can still be
> +			 * delalloc blocks on the inode beyond EOF due to
> +			 * speculative reallocation.  These are not removed

"speculative preallocation"

> Index: xfs/fs/xfs/xfs_inode.c
> ===================================================================
> --- xfs.orig/fs/xfs/xfs_inode.c	2013-12-06 19:58:41.667136764 +0100
> +++ xfs/fs/xfs/xfs_inode.c	2013-12-06 20:00:44.535134243 +0100
> @@ -77,17 +77,18 @@ xfs_get_extsz_hint(
>  }
>  
>  /*
> - * This is a wrapper routine around the xfs_ilock() routine used to centralize
> - * some grungy code.  It is used in places that wish to lock the inode solely
> - * for reading the extents.  The reason these places can't just call
> - * xfs_ilock(SHARED) is that the inode lock also guards to bringing in of the
> - * extents from disk for a file in b-tree format.  If the inode is in b-tree
> - * format, then we need to lock the inode exclusively until the extents are read
> - * in.  Locking it exclusively all the time would limit our parallelism
> - * unnecessarily, though.  What we do instead is check to see if the extents
> - * have been read in yet, and only lock the inode exclusively if they have not.
> + * These two are wrapper routines around the xfs_ilock() routine used to
> + * centralize some grungy code.  They are used in places that wish to lock the
> + * inode solely for reading the extents.  The reason these places can't just
> + * call xfs_ilock(ip, XFS_ILOCK_SHARED) is that the inode lock also guards to
> + * bringing in of the extents from disk for a file in b-tree format.  If the
> + * inode is in b-tree format, then we need to lock the inode exclusively until
> + * the extents are read in.  Locking it exclusively all the time would limit
> + * our parallelism unnecessarily, though.  What we do instead is check to see
> + * if the extents have been read in yet, and only lock the inode exclusively
> + * if they have not.
>   *
> - * The function returns a value which should be given to the corresponding
> + * The functions return a value which should be given to the corresponding
>   * xfs_iunlock() call.
>   */
>  uint
> @@ -101,6 +102,19 @@ xfs_ilock_data_map_shared(
>  		lock_mode = XFS_ILOCK_EXCL;
>  	xfs_ilock(ip, lock_mode);
>  	return lock_mode;
> +}
> +
> +uint
> +xfs_ilock_attr_map_shared(
> +	struct xfs_inode	*ip)
> +{
> +	uint			lock_mode = XFS_ILOCK_SHARED;
> +
> +	if (ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE &&
> +	    (ip->i_afp->if_flags & XFS_IFEXTENTS) == 0)
> +		lock_mode = XFS_ILOCK_EXCL;
> +	xfs_ilock(ip, lock_mode);
> +	return lock_mode;
>  }

Again, I think a static inline is appropriate here...

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs