Re: [PATCH 4/4] xfs: report the correct read/write dio alignment for reflinked inodes

[John Garry <john.g.garry@oracle.com>]

On 06/01/2025 15:15, Christoph Hellwig wrote:
> For I/O to reflinked blocks we always need to write an entire new
> file system block, and the code enforces the file system block alignment
> for the entire file if it has any reflinked blocks.
> 
> Use the new STATX_DIO_READ_ALIGN flag to report the asymmetric read
> vs write alignments for reflinked files.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   fs/xfs/xfs_iops.c | 24 +++++++++++++++++++++---
>   1 file changed, 21 insertions(+), 3 deletions(-)
> 
> diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
> index 6b0228a21617..053d05f5567d 100644
> --- a/fs/xfs/xfs_iops.c
> +++ b/fs/xfs/xfs_iops.c
> @@ -580,9 +580,27 @@ xfs_report_dioalign(
>   	struct xfs_buftarg	*target = xfs_inode_buftarg(ip);
>   	struct block_device	*bdev = target->bt_bdev;
>   
> -	stat->result_mask |= STATX_DIOALIGN;
> +	stat->result_mask |= STATX_DIOALIGN | STATX_DIO_READ_ALIGN;
>   	stat->dio_mem_align = bdev_dma_alignment(bdev) + 1;
> -	stat->dio_offset_align = bdev_logical_block_size(bdev);
> +	stat->dio_read_offset_align = bdev_logical_block_size(bdev);
> +
> +	/*
> +	 * On COW inodes we are forced to always rewrite an entire file system
> +	 * block or RT extent.
> +	 *
> +	 * Because applications assume they can do sector sized direct writes
> +	 * on XFS we fall back to buffered I/O for sub-block direct I/O in that
> +	 * case.  Because that needs to copy the entire block into the buffer
> +	 * cache it is highly inefficient and can easily lead to page cache
> +	 * invalidation races.
> +	 *
> +	 * Tell applications to avoid this case by reporting the natively
> +	 * supported direct I/O read alignment.

Maybe I mis-read the complete comment, but did you really mean "natively 
supported direct I/O write alignment"? You have been talking about 
writes only, but then finally mention read alignment.

> +	 */
> +	if (xfs_is_cow_inode(ip))
> +		stat->dio_offset_align = xfs_inode_alloc_unitsize(ip);
> +	else
> +		stat->dio_offset_align = stat->dio_read_offset_align;
>   }
>   
>   static void
> @@ -658,7 +676,7 @@ xfs_vn_getattr(
>   		stat->rdev = inode->i_rdev;
>   		break;
>   	case S_IFREG:
> -		if (request_mask & STATX_DIOALIGN)
> +		if (request_mask & (STATX_DIOALIGN | STATX_DIO_READ_ALIGN))
>   			xfs_report_dioalign(ip, stat);
>   		if (request_mask & STATX_WRITE_ATOMIC)
>   			xfs_report_atomic_write(ip, stat);