On Wed, Jul 13, 2016 at 12:16:33PM -0400, Brian Foster wrote:
> xfs_buf_read_uncached() allocates an uncached buffer and performs a read
> in one go. As part of the upcoming buftarg I/O accounting mechanism,
> some sites may need to set flags on a buffer before I/O submission.
>
> Create a new helper to support the ability to set flags on a buffer
> before it is submitted for I/O. This use case is the exception, so
> create a wrapper for the original xfs_buf_read_uncached().
>
> Signed-off-by: Brian Foster <bfoster@xxxxxxxxxx>
> ---
> fs/xfs/xfs_buf.c | 20 +++++++++++++++++---
> fs/xfs/xfs_buf.h | 4 ++++
> 2 files changed, 21 insertions(+), 3 deletions(-)
>
> diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
> index 4665ff6..f007713 100644
> --- a/fs/xfs/xfs_buf.c
> +++ b/fs/xfs/xfs_buf.c
> @@ -693,13 +693,14 @@ xfs_buf_readahead_map(
> * buffer containing the disk contents or nothing.
> */
> int
> -xfs_buf_read_uncached(
> +xfs_buf_read_uncached_flags(
> struct xfs_buftarg *target,
> xfs_daddr_t daddr,
> size_t numblks,
> int flags,
> struct xfs_buf **bpp,
> - const struct xfs_buf_ops *ops)
> + const struct xfs_buf_ops *ops,
> + int bflags)
> {
We already have a flags field being passed in. Why can't that be
used to pass the XBF_NO_IOACCT flag? i.e:
> ASSERT(bp->b_map_count == 1);
> bp->b_bn = XFS_BUF_DADDR_NULL; /* always null for uncached buffers */
> bp->b_maps[0].bm_bn = daddr;
> - bp->b_flags |= XBF_READ;
> + bp->b_flags |= XBF_READ | bflags;
bp->b_flags |= XBF_READ;
bp->b_flags |= (flags & XBF_NO_IOACCT);
Cheers,
Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
