Re: [PATCH] Add a page cache-backed balloon device driver.

["Michael S. Tsirkin" <mst@redhat.com>]

On Tue, Jun 26, 2012 at 01:32:58PM -0700, Frank Swiderski wrote:
> +static void fill_balloon(struct virtio_balloon *vb, size_t num)
> +{
> +	int err;
> +
> +	/* We can only do one array worth at a time. */
> +	num = min(num, ARRAY_SIZE(vb->pfns));
> +
> +	for (vb->num_pfns = 0; vb->num_pfns < num; vb->num_pfns++) {
> +		struct page *page;
> +		unsigned long inode_pfn = find_available_inode_page(vb);
> +		/* Should always be able to find a page. */
> +		BUG_ON(!inode_pfn);
> +		page = read_mapping_page(the_inode.inode.i_mapping, inode_pfn,
> +					 NULL);
> +		if (IS_ERR(page)) {
> +			if (printk_ratelimit())
> +				dev_printk(KERN_INFO, &vb->vdev->dev,
> +					   "Out of puff! Can't get %zu pages\n",
> +					   num);
> +			break;
> +		}
> +
> +		/* Set the page to be dirty */
> +		set_page_dirty(page);
> +
> +		vb->pfns[vb->num_pfns] = page_to_pfn(page);
> +	}
> +
> +	/* Didn't get any?  Oh well. */
> +	if (vb->num_pfns == 0)
> +		return;

Went to look at this driver, and noticed caller will re-invoke
this immediately if this triggers, so we busy-wait
re-trying this. When does read_mapping_page fail?
Is there any condition we could wait on?

-- 
MST