Re: [PATCH] xfs: add FALLOC_FL_WRITE_ZEROES to XFS code base

[Christoph Hellwig <hch@infradead.org>]

On Tue, Oct 21, 2025 at 04:17:44PM +0200, Lukas Herbolt wrote:
> Add support for FALLOC_FL_WRITE_ZEROES if the underlying device enable
> -		error = xfs_bmapi_write(tp, ip, startoffset_fsb,
> -				allocatesize_fsb, XFS_BMAPI_PREALLOC, 0, imapp,
> -				&nimaps);
> +		error = xfs_bmapi_write(tp, ip, startoffset_fsb, allocatesize_fsb,
> +				flags, 0, imapp, &nimaps);

Please drop the reformatting that introduces an overly long line.

> -int	xfs_alloc_file_space(struct xfs_inode *ip, xfs_off_t offset,
> -		xfs_off_t len);
> +int	xfs_alloc_file_space(struct xfs_inode *ip, uint32_t flags,
> +		xfs_off_t offset, xfs_off_t len);

Also normal argument order in XFS would keep the flags last, I think
it's best to stick to that.

> -	int			mode,
> +	int				mode,

Spurious whitespace changes here.

>  	len = round_up(offset + len, blksize) - round_down(offset, blksize);
>  	offset = round_down(offset, blksize);
> -	error = xfs_alloc_file_space(XFS_I(inode), offset, len);
> +	if (mode & FALLOC_FL_WRITE_ZEROES) {
> +		if (!bdev_write_zeroes_unmap_sectors(xfs_inode_buftarg(ip)->bt_bdev))
> +			return -EOPNOTSUPP;

Overly long line.

> +		xfs_alloc_file_space(ip, XFS_BMAPI_ZERO, offset, len);

As already mentioned, missing error return.


Also how is the interaction of FALLOC_FL_WRITE_ZEROES and
FALLOC_FL_KEEP_SIZE defined?