Re: [PATCH v4 6/6] xfs: introduce software write streams

[Christoph Hellwig <hch@lst.de>]

On Fri, Jul 17, 2026 at 06:25:38PM +0530, Kanchan Joshi wrote:
> Even when the underlying block device does not advertise write streams,
> XFS can choose do so, as that enables logical spatial isolation and
> dynamic AG-set based concurrency for the standard storage, excluding
> rtvolume.

Ok, this gets back to my previous point.  Pleadse just ignore the
hardware write streams in the initial patch, and then add the aligning
to hardware write streams later.

> +	nr_streams = bdev_max_write_streams(bdev);
> +	if (nr_streams > 0)
> +		return nr_streams;
> +	/*
> +	 * Enable software-only streams if hardware streams are not available.
> +	 * This helps to
> +	 * - improve isolation; reduce allocation interleaving.
> +	 * - improve concurrency using AG-set based steering within and across streams.

Overly long line.

> +	 */
> +	nr_ags = mp->m_sb.sb_agcount;
> +	if (nr_ags >= 16)
> +		ag_set_size = 4;
> +	else if (nr_ags >= 8)
> +		ag_set_size = 2;
> +	else
> +		ag_set_size = 1;
> +	nr_streams = nr_ags / ag_set_size;
> +	return min_t(uint16_t, nr_streams, XFS_SW_WRITE_STREAMS_MAX);

.. and this value should be cached somewhere.

>  	nr_streams = bdev_max_write_streams(mp->m_ddev_targp->bt_bdev);
> -	if (nr_streams) {
> -		mp->m_streams_in_use = bitmap_zalloc(nr_streams, GFP_KERNEL);
> -		if (!mp->m_streams_in_use) {
> -			error = -ENOMEM;
> -			goto out_shutdown_devices;
> -		}
> +	if (!nr_streams)
> +		nr_streams = XFS_SW_WRITE_STREAMS_MAX;
> +	mp->m_streams_in_use = bitmap_zalloc(nr_streams, GFP_KERNEL);
> +	if (!mp->m_streams_in_use) {
> +		error = -ENOMEM;
> +		goto out_shutdown_devices;
>  	}

and also be used here?

While we're at it, the stream allocator should probably go into
the buftarg to prepare for also adding it for the RT volume.